The fences are unusually easy to see in this industry, because federal law stripped away everything that normally hides them.

In 1966 a man named Bob McKee bought five hundred acres of logged-over ridge at Whitethorn, in southern Humboldt County, for fifty thousand dollars. He put two hundred and fifty dollars down and paid the seller off over three years, no interest. A hundred dollars an acre. A century of logging and cattle had left the country with what one local account calls “unstable, eroded hillsides and damaged watersheds,” and the land was going cheap to anyone who would take it. Over the next two decades McKee and others subdivided tens of thousands of such acres, often conveying parcels described against old federal land patents rather than through the Subdivision Map Act, and the buyers were the back-to-the-land settlers of the Emerald Triangle (Scott-Goforth 2013). The land was steep, remote, unwatered, and cheap.

On January 1, 2026, four hundred and seventy-six of Mendocino County’s five hundred and thirty-five cannabis cultivation licenses were still provisional, could no longer be renewed, and could not be converted to annual permits. Eighty-nine percent of the county’s licensed farms lost their legal standing on that date. County cannabis tax revenue had fallen from $3.7 million in fiscal 2018–19 to about $1 million projected for 2025–26, delinquency on the cannabis business tax had been running near 88 percent, and a local realtor put rural land values down at least a fifth (Coryell 2026).

The same ground, cleared twice, sixty years apart.

Land that capital has no present use for becomes available to people capital has no present use for. When a use is found, the land is repriced and the people are removed. The timber companies ran the first pass and a licensing statute ran the second, and the second was more thorough, because the timber companies at least had to buy the land back.

Active US cannabis licenses stood at 36,665 at the end of 2025, down fourteen percent in two years. Cultivation permits fell twenty-four percent from the third quarter of 2023, roughly five thousand farms. Retail licenses fell by three hundred and thirty (CRB Monitor 2026a).

The farms were destroyed and the dispensaries were left standing.

That asymmetry is not what a market correction looks like. A glut punishes everyone holding inventory. What happened instead was that one stage of the chain was emptied out while the stage that sells to the public went on almost untouched, which is the signature of value being pulled off the land and captured somewhere downstream. Everything that follows is an attempt to say which instrument did the pulling, in what order, and to whom.

The fence goes up before anyone calls it a fence

Enclosure is the process by which things held in common are fenced, titled, and handed to private owners: the village green hedged off and sold, the shared field converted into a rent. Between 1604 and 1914 Parliament passed somewhere between 5,200 and 5,400 acts of enclosure covering roughly 6.8 million acres, a little over a fifth of England (UK Parliament n.d.; Tate and Turner 1978). Marx called the statutory form of it “decrees by which the landlords grant themselves the people’s land as private property, decrees of expropriation of the people” (Marx 1867, vol. 1, ch. 27). Polanyi took up an older description of enclosure as a revolution of the rich against the poor, and endorsed it (Polanyi 1944, ch. 3).

Applying that to a plant requires saying what the commons consisted of, because there was no village green.

Three things were held in common in American cannabis. The genetics came first: landraces carried out of Afghanistan and Mexico and Thailand, then selected over four decades by growers working in isolation, trading seed through the mail and at the edges of harvest season. That library had no owner. It circulated precisely because it could not be sold through any lawful channel, which is the condition under which things stay common. The cultivation knowledge came second and by the same route. Light cycles, feeding schedules, harvest timing, the length of the cure, the management of mold in a wet coastal autumn: all of it developed with no extension agents, no agricultural colleges, no published trials, and no patents, and it passed hand to hand because prohibition had made it legally unownable.

The third one nobody counts, and it is the one that is most important.

What the movement produced across fifty years, at a cost measured in arrests, criminal records, forfeited land, and prison time, was legality itself. Every petition drive, every ballot initiative, every patient who testified, every grower who took a felony so the argument could be made, was labor invested in a good that would exist for everyone or for no one. Legality was a commons under construction, and it was built by the people who paid for it in advance.

At the moment it finally arrived, it was surveyed, subdivided, and issued as title to whoever could afford the survey. Which is what a license is. A fence with a toll booth on it.

The operation runs in three steps and it has run this way against every commons that has ever been taken. First the commons is identified, mapped, and inventoried, and so are the people who will be excluded from it: seed-to-sale tracking, canopy measured to the square foot, potency assayed to the decimal, a registry naming every participant and every plant, and behind all of it forty years of arrest records naming everyone who would not qualify. Almost nothing about cannabis was measured under prohibition, and measurement was the first act of legalization, because a thing cannot be fenced until it has been surveyed. Second comes the enclosure itself, executed through legal-political force alongside unilateral private action: the licensing statute, the vertical rules, the residency requirements, the compliance costs, and running beside them the sale-leaseback and the credit bid. Legacy growers were not in the room when Colorado, Washington, and California wrote the rules, and the rules were drafted so that they could not meet them. Third comes ratification, after the fact and at leisure: the abatement proceedings, the provisional license deadline of January 1, 2026, the recriminalization of hemp, and a Ninth Circuit holding this January that the constitutional doctrine of the national market has nothing to say about any of it.

England did all of this first, and it left the documents that codified it.

Section 31 of the Inclosure Act 1845 empowered the Commissioners to require, as a condition of enclosure, an allotment for the labouring poor. The provision was permissive rather than mandatory. It set no minimum acreage, left the size to whatever the Commissioners thought necessary “with reference to the Circumstances of each particular Case,” and attached a rent-charge to whatever was granted. In July 1875 the House of Commons was told that of 368,000 acres enclosed under the Act, 2,223 had gone to the poor (Hansard 1875). Six-tenths of one percent.

Then section 142, which is the sentence worth carrying through the rest of this essay. It provided for the sale of an allottee’s land, by the valuer, to cover the expenses of enclosing it. Section 126 supplied remedies for non-payment. Parliament wrote the foreclosure into the redistribution.

Parliament understood exactly what it was doing. A grant of land to someone with no capital is a grant of an obligation. Enclosing cost a lower bound near £12 an acre, which for a twenty-acre holder came to roughly five times a year’s income (Heldring, Robinson, and Vollmer 2022, citing Turner and Mingay), and two-thirds of Northamptonshire commoners sold some or all of their land or gave up their tenancies (Neeson 1977). The small owners in enclosed parishes fell by twenty-one percent while the number in unenclosed parishes did not move at all (Neeson 1993).

There is a qualifier to this analysis. Leigh Shaw-Taylor’s study of ten Midlands settlements found that most agricultural labourers held no common rights before enclosure and therefore cannot have been proletarianized by losing them (Shaw-Taylor 2001). His unit is the labourer. The mechanism described here concerns the smallholder and the cottager, people who did hold rights and did receive allotments. Trimmers and budtenders lost no property in 2016, because they held none in 2015. This essay is about the people who had something.

An experiment nobody meant to run

Cannabis is the one significant American industry operating without the institutions ordinary capitalism takes for granted. Federal prohibition removed four of them at once: commercial banking, bankruptcy protection, interstate commerce, and the deduction of ordinary business expenses. Those four smooth a transaction, and in smoothing it they hide who is claiming what from whom. A mortgage reads as a service. A wholesale margin reads as a price. A bank rate reads as something like the weather.

In the third volume of Capital, Marx worked out what becomes of a surplus once production ends. It leaves the producer and is parcelled among fractions of capital, each holding a different kind of claim, each enforced through a different legal instrument. Industrial capital keeps the profit of enterprise. Merchant capital takes its share for moving the commodity. Interest-bearing capital takes interest. Landed property takes rent. The state takes tax. In a mature industry those five arrive quietly, as a lease renewal, a distributor’s markup, a revolving facility, a property assessment, and the question of who is taking what from whom stays buried in the accounting. In cannabis they arrive as a sale-leaseback priced against 2021 wholesale assumptions, a note at 12.75 percent, an invoice nobody intends to pay, and a tax levied on gross receipts.

A real estate investment trust that converted a grower’s building into an escalating claim on that grower’s future output, then repossessed the building when the output stopped paying, has done nothing unusual. It occupied the slot of landed property in a market where the absence of banks let landed property set its own price. Bondholders taking a company’s assets by credit recovery are exercising the only remedy the law left them, in the only venue it left open. Neither required malice. Neither would change if every executive involved were replaced tomorrow with someone of better character, and the distribution would come out the same to the dollar.

The state works the same way, according to a structuralist narrative. Nicos Poulantzas argued that the capitalist state is best understood as the condensation of a relation of forces, holding real autonomy from any particular capitalist and answering to none of them individually. Claus Offe put the operative point more usefully for a case like this one: the state contains selective mechanisms, institutional filters that admit some strategies and exclude others regardless of who happens to staff it. A licensing statute applied identically to every applicant is formally universal and substantively selective, because the attribute it selects on is access to capital and the ability to wait.

No corruption is required. Corruption is a weaker hypothesis: it predicts that honest regulators would produce a different distribution of ownership, and across two dozen states with wildly varying administrative quality the distribution keeps coming out the same shape.

Marx distinguished the formal subsumption of labor under capital, where capital takes its cut from a labor process it has not reorganized, from real subsumption, where capital remakes the process itself. Prohibition held cannabis close to the first and often short of it. The grower owned the ground and the genetics, performed or directly supervised the work, and sold into a market that skimmed him through risk premiums and middlemen without ever touching how he farmed. Legalization completed the second. Cultivation became wage labor inside a building owned by a REIT, financed at thirteen percent by a specialty lender, tracked by mandated software, tested by a laboratory the distributor selected, and sold through a store owned by the same parent company as the farm across the state line.

Every one of those relations is a claim, and each claim is prior to the grower’s. The five claims are the toll collectors, and the license is the gate they stand at.

Set against that description, the industry’s account of itself starts to look evasive. Operators describe their condition as oversupply, or a difficult market, or bad luck with timing. Those words name a quantity and put the cause in nature, where nobody drafted it. The five claims name a distribution and put the cause in legal instruments with authors, dates, and beneficiaries.

How the growers were put on bad ground

Policy toward this plant has always tracked what capital wanted from it. Parliament paid bounties on colonial hemp from 1705 under 3 & 4 Anne c. 10 because Britain bought its cordage from Russia and wanted an imperial substitute, and George Washington drew nearly eleven pounds of Virginia bounty money in November 1767 (Smith 2016; Butler 2020). The American industry then died of labor costs. Lyster Dewey, the USDA’s botanist in charge of fiber-plant investigations, wrote in 1913 that the causes were “the increasing difficulty in securing sufficient labor to take care of the crop,” better returns from stock and tobacco, and substitution by jute, abaca, and wire rope (Dewey 1913). Emancipation and the closing of the cheap-labor frontier finished what the wire rigging began.

The prohibition that followed was assembled out of pharmacy regulation, bureaucratic self-preservation, and racial fear, in roughly that order. California’s 1913 ban was an obscure amendment to the state Poison Act, drafted by the Board of Pharmacy, and the anxiety animating its author concerned South Asian immigrants (Gieringer 2024). The Federal Bureau of Narcotics, created in 1930, did not begin as a marijuana crusade. Richard Bonnie and Charles Whitebread found that the Bureau’s priority through 1934 was persuading states to adopt the Uniform State Narcotic Drug Act, and that it turned to marijuana in late 1934 because marijuana worked as publicity. Adoption then moved: thirty-five states had the Uniform Act by 1937, every one of them with a marijuana provision (Bonnie and Whitebread 1970).

The story that Hearst, DuPont, and Mellon engineered prohibition to protect nylon and wood pulp originates with Jack Herer in 1985 but it is not supported by the historical record. The Tax Act took effect in October 1937; the Popular Mechanics piece about the billion-dollar hemp crop ran in February 1938; nylon competed with silk. Bureaucratic survival is the duller explanation, the better one, and the more damning, because it means nobody had to want this.

The Marihuana Tax Act of 1937 taxed transfers to a registered buyer at one dollar an ounce and to an unregistered buyer at one hundred dollars, roughly a hundred times the retail price, and required a Treasury order form only registrants could obtain (50 Stat. 551, §§ 2, 6, 7). Prohibition was executed as a licensing chokepoint with a revenue price tag on it, which is the same instrument the states would reach for eighty years later. The American Medical Association’s legislative counsel, William Woodward, told the House Ways and Means Committee on May 4, 1937 that “no one has been produced from the Bureau of Prisons to show the number of prisoners who have been found addicted,” and the committee answered by telling him he should have come with constructive proposals instead of obstruction (Woodward 1937). Thirty-two years later the Supreme Court voided the transfer provisions in Leary v. United States because compliance required a citizen to identify himself as part of a “selective group inherently suspect of criminal activities.”

Congress replaced the tax fiction with the Controlled Substances Act in 1970 and placed cannabis in Schedule I as an admitted holding position, pending studies it wrote into the same statute at § 601 (see NORML v. Bell, 488 F. Supp. 123 (D.D.C. 1980)). Those studies were the Shafer Commission. Nixon rejected its findings at a news conference on May 1, 1971, eleven months before it reported: “Even if the Commission does recommend that it be legalized, I will not follow that recommendation” (Nixon 1971). The report, when it came, recommended decriminalizing personal possession and non-commercial transfer while leaving cultivation and sale for profit felonious (National Commission on Marihuana and Drug Abuse 1972).

Tolerated consumption and criminal supply. That configuration governed the next forty years, and it is what built the Emerald Triangle.

Enforcement then specified the form production would take. Operation Intercept sealed the Mexican border for twenty days in 1969 and touched off a domestic harvesting boom (Brecher 1972); paraquat spraying in Mexico after 1975 made imported cannabis chemically suspect; domestic production rose from a tenth of national supply in 1980 to a quarter by 1992 (Drug Science n.d.). Two statutes then shaped what that production looked like. The Anti-Drug Abuse Act of 1986 set mandatory minimums by plant count “regardless of weight,” so one hundred seedlings carried the same five-year floor as one hundred kilograms of finished product (21 U.S.C. § 841(b)(1)(B)). The Comprehensive Crime Control Act of 1984 created the Assets Forfeiture Fund and equitable sharing, letting local departments keep up to ninety percent of what they seized in federal partnership, in a country where the median state share had been zero. Forfeiture receipts went from $27 million in fiscal 1984 to $556 million in fiscal 1993 (Kantor, Kitchens, and Pawlowski 2017).

Growers responded rationally. Plant counts fell and value per plant rose, which is the economic definition of sinsemilla. Median plot size dropped to around one hundred plants by 1985, sitting just under the federal threshold. Production moved onto public land, where there was no title to forfeit, and then indoors, where there was nothing to see from a helicopter. And the helicopters came. The Campaign Against Marijuana Planting destroyed 158,493 plants in its 1984 season across twenty-four counties, seized 524 firearms and 47 vehicles, and fielded ninety-one participating agencies (CAMP 1984). In the summer of 1990, Operation Green Sweep put roughly 230 personnel onto BLM land in Humboldt County, some sixty of them regular Army from the 7th Infantry Division.

All of this was productive, and that is the part the drug-war literature keeps missing. Prohibition was accumulation by dispossession in David Harvey’s sense, and what it manufactured was a class: small producers holding land, skill, and no legal standing whatsoever, parked on ground nobody else wanted (Harvey 2003). Joseph Mello’s reading of legalization as an extension rather than a dismantling of the carceral order is the correct sequel (Mello 2024). The punishment apparatus was reconfigured, not retired, and even now adult-use laws reduce arrests for Black and white populations alike while “narrowing but not eliminating disparities,” with illegal-drug incarcerations falling only among white populations (Meinhofer, Rubli, and Cunningham 2026).

Forty years of enforcement produced exactly the producer the licensing statutes would later find unqualified. It also produced the paperwork proving he was unqualified.

The survey was conducted by police

On April 16, 1968, a Houston grand jury indicted Lee Otis Johnson, an organizer for the Student Nonviolent Coordinating Committee, for giving one marijuana cigarette to an undercover officer. He was tried that August and sentenced to thirty years in the Texas Department of Corrections. When a federal court finally granted habeas relief in January 1972, Judge Carl Bue found that Johnson’s name had appeared in Houston’s two major newspapers forty-four times in the eighteen months before trial, that he “had the reputation in the community as a dangerous black militant,” and that “any black militant connected with the TSU riot could not get a fair trial in Houston at that time” (Johnson v. Beto, 337 F. Supp. 1371 (S.D. Tex. 1972), aff’d 469 F.2d 1396 (5th Cir. 1972)). TIME reported that the district attorney had used his peremptory challenges to strike every Black member of the venire (TIME 1972). Johnson served roughly four years for the cigarette.

On December 22, 1966, John Sinclair handed two marijuana cigarettes to an undercover Detroit officer. He was convicted of possessing them, sentenced on July 28, 1969 to nine and a half to ten years, and was still in prison in December 1971 when fifteen thousand people filled Crisler Arena in Ann Arbor for a rally at which John Lennon performed. The Michigan Supreme Court ordered his release three days later, and in March 1972 struck down the statute itself, holding that marijuana had been “erroneously classified with the opiates,” that the arrest was entrapment, and that the sentence was cruel and unusual (People v. Sinclair, 387 Mich. 91 (1972)).

One cigarette, thirty years. Two cigarettes, ten. The quantity was never what the sentence was measuring.

What it was measuring is on the record, in a memorandum. On July 5, 1968, J. Edgar Hoover circulated to every FBI field office a list of twelve counterintelligence techniques for use against the New Left. Item five reads: “Since the use of marijuana and other narcotics is widespread among members of the New Left, you should be alert to opportunities to have them arrested by local authorities on drug charges.” Item four instructed agents to place articles about New Left drug use with university officials and parents. Item ten directed them to tip local police about coffeehouses near military bases.

A federal police agency instructed its field offices, in writing, to procure drug arrests of political dissidents.

The same instrument was aimed at individuals. An FBI memorandum of July 27, 1972, three weeks before the Republican National Convention, recorded the Bureau’s intent to have Lennon arrested “if at all possible on narcotics charges,” on the reasoning that a narcotics conviction would render him more immediately deportable; Senator Strom Thurmond had written to the Nixon White House the previous year warning that Lennon might “combine rock music with politics and organize young people to vote against Nixon in the 1972 election” (Wiener 2000). The White Panther Party’s prosecution, meanwhile, produced United States v. U.S. District Court, 407 U.S. 297 (1972), in which the government argued to the Supreme Court that the President held an inherent power to wiretap domestic radicals without a warrant, and lost eight to nothing.

The White House tapes say the rest. On May 26, 1971, Nixon told H. R. Haldeman that “every one of the bastards that are out for legalizing marijuana is Jewish,” and that “we are going to hit the marijuana thing.” On May 18 he had described the demonstrators as “all on drugs. Oh yeah, virtually all.” On September 9, 1971, he instructed Raymond Shafer, chairman of the ostensibly independent commission he had appointed, to “keep your Commission in line” and not appear soft on marijuana. On March 21, 1972, the day the Commission reported that private possession should be decriminalized, he told Haldeman: “Enforce the law, you’ve got to scare them.”

The scaring was the policy. The commission was a formality he had already overruled in a private conversation seven months earlier.

The racial content of all this is so thoroughly documented that its most famous piece of evidence turns out to be a forgery. The line attributed to Harry Anslinger about jazz, “Satanic music,” and the hundred thousand marijuana smokers who are mostly “Negroes, Hispanics, Filipinos and entertainers” is quoted everywhere, and Penn State University Libraries, which holds the Anslinger Papers, searched the folder it is always cited to and reported that no document there contains it. The case against prohibition never needed the quotation and is stronger without it.

The arrest data reflects the enforcement strategy. Between 2001 and 2010 there were 8,244,943 marijuana arrests in the United States, 88 percent of them for possession. In 2010, 14 percent of Black Americans and 12 percent of white Americans reported using marijuana in the previous year, and a Black person was 3.73 times more likely to be arrested for possessing it (ACLU 2013). Marijuana arrests peaked at roughly 872,000 in 2007 and stayed above 850,000 through 2010.

Then the reform decade, which is where the argument stops being about the past. Between 2010 and 2018 there were another 6.1 million marijuana arrests, and as states legalized the racial disparity widened, from 3.31 to 3.64 (ACLU 2020). Between 2002 and 2019, cannabis possession arrests rose 28.1 percent for Black Americans while falling 22.7 percent for white Americans (National Academies 2024).

Legalization arrived and the arrests became more unequal, not less.

Kathleen Frydl’s account of why an obviously failing policy persisted for eight decades is the one that survives contact with this record: drug control endured because suppressing drugs was never the operative objective, and enforcement “advanced certain state agendas, such as policing inner cities or exercising power abroad” (Frydl 2013). Elizabeth Hinton shows the same apparatus being built by Democratic administrations as enthusiastically as Republican ones, out of the Great Society itself, which is what makes this structural rather than partisan (Hinton 2016). And Stuart Hall and his colleagues described the mechanism in 1978: a moral panic is how a state that can no longer rule by consent manufactures consent for ruling by force, with police, courts, press, and politicians converging on a single account of a threat whose measured scale never justified the response (Hall et al. 1978).

All of this would be history, and belongs in a different essay, except for what the record did next.

Michelle Alexander’s central point is that the felony record is the instrument: race-neutral on its face, race-specific in operation, and durable long after the law that produced it has changed (Alexander 2010). In cannabis that durability has a specific and documentable form. Colorado’s adult-use regulations, written in 2013 in the first year of legal sales, disqualified from licensure anyone with a felony controlled-substance conviction in the preceding decade, and applied the bar to managers and employees as well as owners (1 CCR 212-2). Washington scored a felony at twelve points against a threshold of eight, so a single conviction disqualified automatically (WAC 314-55). Nevada and Alaska ran ten-year and five-year bars. Oregon, California, and Massachusetts wrote discretionary versions of the same thing.

The decade of convictions those rules excluded is the decade the ACLU measured at 3.73 to one.

Before a commons can be enclosed, the people who will have no standing in it must be identified and marked. In England that was done by commissioners with chains and field books. In America it was done by police, over forty years, at a documented racial ratio, and the resulting records were then written into the licensing statutes as a disqualification.

The state manufactured the disqualifying characteristic and then applied it as a filter. Nobody had to intend the sequence for it to work, and nobody has ever had to answer for it.

The rules did move. New York’s conditional retail program made a prior marijuana conviction an eligibility criterion rather than a bar, Illinois built arrest history into its social equity definition, and Connecticut’s current disqualification list covers fraud and money laundering while omitting drug offenses. Twelve states now mandate license allocation to equity applicants.

Every one of those reforms arrived after the licenses, the capital, and the market positions had been distributed. The National Academies reports that 75 to 80 percent of cannabis retail outlets are owned by white people and that fewer than 6 percent of owners or employees are Black (National Academies 2024). Roughly 2.3 million marijuana records have been cleared nationally, against more than 14 million arrests in the two study windows above, and where relief requires a petition rather than arriving automatically, about 6.5 percent of eligible people obtain it within five years.

The gate was opened after the room had been filled.

The collective was a fiction everyone found useful

California’s Proposition 215 passed on November 5, 1996 with 55.58 percent of the vote and created a criminal defense. It licensed nothing, taxed nothing, and specified no lawful channel of supply. Every institution of the medical era was improvised in that gap. SB 420 added the collective in 2003, exempting patients and caregivers who “associate within the State of California in order collectively or cooperatively to cultivate marijuana for medical purposes” (Cal. Health & Safety Code § 11362.775). Attorney General Jerry Brown’s 2008 guidelines closed the circle: nothing in state law “authorizes collectives, cooperatives, or individuals to profit from the sale or distribution of marijuana,” and collectives were to maintain “a closed-circuit of marijuana cultivation and consumption with no purchases or sales to or from non-members” (California Attorney General 2008).

A membership was signed at the counter in half a minute and reimbursement for overhead was a price. Everyone involved understood this.

What the fiction produced was structural, and for fifteen years it protected the people it pretended to regulate. A business that legally cannot earn a return cannot issue equity against future returns, so growth was financed out of cash flow. Banking exclusion made the sector all-cash. Section 280E denied ordinary deductions. Legal ambiguity was itself the barrier to entry, and it selected for operators willing to absorb criminal risk, which meant the people already in the trade. The medical era held the Emerald Triangle grower in place by making the industry unattractive to anyone with an accountant. For fifteen years the commons stayed open because nobody with money wanted to be seen inside it.

The federal government then demonstrated what would end that.

The Ogden memorandum of October 19, 2009 told US Attorneys not to focus resources on people in “clear and unambiguous compliance with existing state laws,” while stating in the same document that “prosecution of commercial enterprises that unlawfully market and sell marijuana for profit continues to be an enforcement priority” (Ogden 2009). No storefront could be unambiguously compliant under a non-profit statute, so the safe harbor was empty, and the investment boom of 2010 and 2011 rested on a misreading of it. James Cole withdrew the inference in June 2011. On October 7, 2011 California’s four US Attorneys announced coordinated action, and the instrument was the landlord letter. Warning property owners that “the operation of a marijuana store may result in criminal prosecution, imprisonment, fines, and forfeiture of assets” transferred enforcement to private parties, who evicted without any court ever reaching the question of state-law compliance (US Attorney, C.D. Cal. 2011).

Colorado and Washington voted to legalize thirteen months later, on November 6, 2012, by 55.32 and 55.70 percent. Both measures were drafted and campaigned during the crackdown, and their design reads as an answer to it. Seed-to-sale tracking, license caps, tax structure, residency, vertical rules: these were compliance signals addressed to a Department of Justice that had said it wanted “strong and effective regulatory and enforcement systems.”

The people capable of producing those signals were, almost by construction, not the people who had supplied the market for the preceding four decades. That is step one of the operation, and it was performed in public, by ballot initiative, with the movement campaigning for it.

The license is the fence

The state entered cannabis in 2012 on a scale it had never before attempted with an agricultural commodity. Government now determines who may cultivate, at what canopy, in which jurisdictions, with which tracking software, in what package, at what potency, subject to which tests, taxed at which rates, and sold through whose stores. Every one of those determinations allocates income.

Calling a license permission misdescribes it. A license is a state-created title to a stream of revenue that others are legally barred from earning, which is monopoly rent in the exact sense: a claim on surplus arising from exclusive control of access to a market rather than from anything the holder produces. Capped licenses acquire speculative value on the day they issue and trade accordingly. Uncapped licenses with expensive compliance leave entry formally free and move the barrier into the cost structure.

The sixth part of Capital’s third volume turns out to be the working manual for this industry, and all three of its rent forms are running at once. Monopoly rent is the license. Absolute rent is what a jurisdiction collects simply by permitting retail when two-thirds of its neighbors refuse to, a barrier made of nothing but municipal politics. Differential rent is the gap between a hillside parcel in Mendocino with no recorded water right and a flat, permitted, road-served acre in Salinas (Marx 1894, vol. 3, pt. VI).

Land in a legalized cannabis market stops being where the plant grows and becomes an instrument for capturing other people’s output.

Colorado wrote the barrier into its retail regime, carrying its medical “70/30 rule” into adult use so a store had to grow at least seventy percent of what it sold. An entrant had to finance a farm and a storefront at once, without bank credit, under § 280E. Jacob Sullum reported at the time that incumbent medical operators were the rule’s constituency, because it barred newcomers who would otherwise have competed on a single function (Sullum 2013). Washington did the opposite, barring producers and processors from any financial interest in a retailer (RCW 69.50.328), which lowered the capital threshold for each tier and left cultivators with no downstream margin to retreat into when prices fell. Both states imposed residency requirements and both relaxed them later.

Local operators bore the highest-risk, lowest-capital phase. Outside capital was admitted once the legal risk had been retired. That is a transfer of value executed on a timetable, and it was written into the statutes in advance.

Compliance cost is the fence bill, and it works now exactly as it worked in 1845. Testing is a real public good and the labs catch real hazards. Pablo Valdes-Donoso, Daniel Sumner, and Robin Goldstein put California’s mandatory testing at roughly $136 per pound of dried flower, about a tenth of average wholesale price at the time (Valdes-Donoso, Sumner, and Goldstein 2020). A firm running thousands of pounds absorbs that as a rounding error. A farm running eighty pounds absorbs it as its margin. Michael Polson and Hekia Bodwitch, surveying more than seven hundred California cultivators, found licensing cost and complexity among the decisive barriers to entering the legal market (Polson and Bodwitch 2021).

Uniform in law, regressive in effect. The state that imposed it built no shared laboratories and no graduated standard, and no commissioner had to intend the result.

Then the geography. More than two-thirds of California jurisdictions ban commercial cultivation, and 16.4 million Californians live where licensed retail is prohibited, 85 percent of them unaware of it (California DCC n.d.; Getz, Petersen-Rockney, and Polson 2024). Local prohibition converts the surviving retail licenses into pure locational rent and guarantees the illicit market a permanent customer base, which the state then polices at public expense.

Michael Polson and his colleagues at Berkeley have described the pattern: counties with higher incomes and more traditional agriculture issue fewer cannabis licenses, and the crop is pushed onto marginal land whose marginality is “socially and geographically produced” (Dillis, Petersen-Rockney, and Polson 2024). Which returns the argument to Whitethorn. Those parcels were cheap because they were steep, roadless, waterless, and often not legal parcels at all. Forty years later the same attributes were the reasons a cultivator could not demonstrate a legal parcel, show an appropriative water right, or meet grading and access standards.

Cheap land in 1970 and unlicensable land in 2018 are the same land. Both regimes assigned the crop to ground nobody else wanted, and then the second regime disqualified the crop for being there.

A market with the doors welded shut

Cannabis is the only significant American commodity that is legally produced in most states and legally transportable in none. Oregon’s SB 582 and California’s SB 1326 authorize interstate agreements contingent on a change in federal policy that has not come. The constitutional route closed in January 2026, when the Ninth Circuit held unanimously in Peridot Tree WA Inc. v. Washington State Liquor and Cannabis Control Board that the dormant Commerce Clause does not apply, reasoning that the doctrine’s objective of preserving a national market “surely wanes when the national marketplace is prohibited under federal law.” Jefferson Packing House withdrew its Oregon challenge on June 17, 2026, its counsel citing that decision and the federal hemp redefinition as having “directly impacted the key arguments in our lawsuit” (Cannabis Business Times 2026).

The First and Second Circuits used it to strike down Maine’s and New York’s residency and equity preferences (Northeast Patients Group, 45 F.4th 542 (1st Cir. 2022); Variscite NY Four (2d Cir. 2025)). The Ninth then held it unavailable to anyone trying to open the market itself. Constitutional law was made to dissolve the protections and made unavailable for the fences.

The consequence is the least discussed fact about this industry. Capital cannot equalize the rate of profit across space. There are two dozen sealed pools of cannabis capital, each with its own supply and price, and none can export a surplus or import a shortage. A glut in Oregon cannot relieve scarcity in Minnesota. Retail flower at roughly two dollars a gram in Michigan and ten to eighteen in Minnesota cannot arbitrage. Overproduction inside a sealed market has nowhere to go except into price, and price collapse inside a sealed market resolves only through the destruction of local producers.

Capital ordinarily migrates out of low-return regions and into high-return ones, and that migration is what pulls returns toward a general rate. Federal law has switched the mechanism off inside a single national industry, and left the wreckage where it fell.

It also explains why the multi-state operator takes the peculiar form it does. An MSO cannot run one large farm and ship, so it replicates cultivation, processing, and retail separately inside every state it enters. In engineering terms this is waste. In competitive terms it is the whole point, because financing twenty duplicated build-outs, absorbing years of losses across them, and buying distressed licenses in each is precisely the capacity small producers lack. Joel Garrod’s study of the Canadian sector reaches the same conclusion from the other direction: the major cannabis firms are transnational and asset-light by design, their ownership financial rather than national, which is why regional economic salvation through cannabis keeps failing to arrive (Garrod 2025).

Federal prohibition, which the industry treats as its principal grievance, selected for the firms best able to pay the cost of working around it.

Abundance arrives as catastrophe

MJBizDaily reported in July 2026 that regulated US cannabis revenue fell from about $32 billion in 2024 to $29.94 billion in 2025, the first annual decline in its series, while consumers bought more product for less money. Michigan retailers moved 348 million units against 291 million the year before and took in $3.17 billion against $3.38 billion (Jackson 2026). Vangst and Whitney Economics, on a different methodology, put 2025 at $29.1 billion and reach the same conclusion (Vangst 2026). Michigan’s average ounce went from $95.08 in December 2023 to $58.20 in December 2025.

More cannabis, sold to more people, for less money, and the industry calls this a crisis.

California’s collapse is deeper and better documented. ERA Economics, reporting to the Department of Cannabis Control, found inflation-adjusted wholesale prices in the fourth quarter of 2024 down fifty-seven percent against the same quarter of 2020, outdoor flower down seventy-four percent, indoor down forty-six, while licensed production kept growing (ERA Economics 2025). Mendocino’s county-average farmgate price fell from $1,377 a pound in 2020 to $721 in 2024, against $3,000 in 2009 (Coryell 2026). The Berkeley team counted roughly 4,900 licensed cultivators in April 2024, down forty-one percent in two years against tens of thousands before legalization, and recorded trimmer wages in the unlicensed sector falling by as much as seventy-five percent while room and board disappeared (Polson, Laudati, and Sayre 2024).

This is a crisis of overproduction in the exact sense. The surplus exists relative to what can be sold profitably inside the licensed channel, and that channel has been narrowed deliberately: by taxes, by local retail bans, by an unlicensed market supplying an estimated sixty-two percent of California consumption, and until recently by a hemp sector that outsold it. Society has more cannabis than the market can absorb at a price that services the debt taken on to grow it.

Those are two different problems and only one of them is being treated. Abundance is what a commons produces. Scarcity is what a fence produces, and the entire licensed structure was built to produce it.

Three claims, arriving in order

Each of the three claims below is a gate, and each was installed on ground the grower had already paid for. The commonest way for a cultivator to raise money without a bank was to sell the building and lease it back. Innovative Industrial Properties, a real estate investment trust, built a $2.5 billion portfolio on that trade: triple-net leases of fifteen to twenty years with annual escalators, priced off cannabis cap rates far above conventional industrial ones. An operator sold a facility at a low multiple and took on rent at a high implied cost of capital, permanent, escalating, and senior to everything else. The industry’s own advisory literature has a name for the result, over-renting, and its advice was to test debt service against conservative wholesale-price scenarios (420 Property n.d.).

Wholesale prices then fell by half. The rent claim did not adjust, because rent claims never do.

IIPR’s fiscal 2025 filings show rental revenue of $265.5 million, down fourteen percent, and net income to common shareholders down twenty-eight percent. The dividend was raised one percent, to $7.60. PharmaCann accounted for 9.9 percent of annualized contractual rent and owed $29.2 million; 4Front, Gold Flora, and TILT together accounted for another 11.6 percent and $31.3 million, all in receivership or bankruptcy. Roughly a fifth of the landlord’s contractual rent was owed by insolvent tenants, some sixty million dollars unpaid, and in December 2025 the landlord recovered possession of PharmaCann’s Illinois cultivation facility (IIPR 2026). IIPR then placed $50 million into preferred shares of a life-sciences REIT at a fifteen percent yield and committed $120 million more.

The landlord converted its tenants’ fixed capital into a claim on their surplus, repossessed the fixed capital when the surplus failed to appear, raised the dividend, and began moving its money into another industry. Every step of that was lawful and none of it was hidden.

Interest-bearing capital takes its cut at similar rates. Ascend Wellness carries $300 million of senior secured notes at 12.75 percent, Cresco Labs a $325 million term loan at 12.5 percent, Ayr Wellness an exit facility at 13.00 percent. Ascend’s interest expense reached $20.3 million in a single quarter, nearly double the year before, against total liabilities of $947.9 million and total assets of $872.1 million, a stockholders’ deficit of $75.8 million.

Ten and a half to thirteen percent is what finance charges to stand in the space where a bank would be.

The third claim is the one nobody designed and everybody uses, and it is the single most under-reported fact in American cannabis. Whitney Economics found $3.8 billion of delinquent receivables against $28.8 billion in legal sales: thirteen percent of all industry revenue, the equivalent of 1.6 months of national turnover, more than half of it over forty-five days past due. Multi-state operators and large corporations accounted for $1.4 billion of that, thirty-six percent, averaging $34.8 million each. Cultivation bears the heaviest delinquency and retail the lightest. Forty-four percent of operators said unpaid receivables impaired their ability to service debt, and 57.3 percent said the problem was worse than § 280E. Only 27.3 percent of US cannabis operators were profitable at all (Whitney Economics 2024a, 2024b).

Without the euphemism, that arrangement is a credit system. Cultivators lend to retailers and to multi-state operators at zero percent, involuntarily, with no bankruptcy court to appeal to and no commercial paper to enforce. It is the largest source of working capital in the industry and it is extracted from the poorest participants in it.

The farm finances the store.

New York now requires payment within thirty days and mandatory reporting of delinquency within seven, while stating that the Office of Cannabis Management “is not a collections agency or payment enforcer” (NY OCM n.d.). California cultivators maintain a user-generated blacklist of the state’s worst-paying buyers, which is a privatized reconstruction of the commercial law the state declines to supply.

Who eats the adjustment

Marx distinguished the concentration of capital, firms growing by reinvesting their own profits, from its centralization, existing firms and assets brought under fewer centers of control through acquisition, credit, and failure (Marx 1867, vol. 1, ch. 25). Centralization normally proceeds through negotiated restructurings and orderly courts, which soften what is happening and slow it down. Neither is available here, and the result is the process running with its clothes off.

The Cannabist Company, formerly Columbia Care, defaulted on roughly $179 million of senior secured notes in January 2026, filed under Canada’s Companies’ Creditors Arrangement Act on March 24, and obtained Chapter 15 recognition from the US Bankruptcy Court in Delaware on May 9, the first time a US bankruptcy court has recognized a foreign insolvency proceeding involving a cannabis enterprise. The debtors were holding companies and the plant-touching operations sat in non-debtor subsidiaries, which is precisely why recognition was available (Morrison Foerster 2026). Its licenses were then sold off by jurisdiction: Virginia for $130 million, Ohio for $47 million, Delaware for $16.5 million. Ayr Wellness, with more than ninety stores across seven states, entered wind-down while its senior secured noteholders formed a vehicle called Arboretum Bidco and began taking the state assets by credit bid, receiving equity in the acquiring parent in satisfaction of their notes.

Equity extinguished. Assets intact. New owners, same greenhouses, same plants in them.

A credit recovery is an enclosure performed on an enclosure. The fence stays exactly where it was and the gate changes hands, and the grower who built what is inside it has no standing in the proceeding, which is the one continuity running from 1845 to Delaware.

Because Chapter 11 is closed to plant-touching firms, restructuring runs through Article 9 foreclosure, state receiverships, and Canadian insolvency proceedings, venues with far weaker debtor protections and no automatic stay of the American kind. Federal prohibition, which was supposed to inconvenience capital, has delivered the industry’s productive assets to creditors on terms no ordinary American manufacturer would face. The AdvisorShares MSOS ETF fell 87.1 percent in five years, from $36.50 at the end of 2020 to $4.72 at the end of 2025. Curaleaf, adjusted for its 2026 reverse split, has gone from the equivalent of $54.00 in February 2021 to $9.21 in July 2026.

Consolidation here is an overbuilt, debt-laden sector reorganizing through failure, and the survivors are the firms that can lose money longest.

Labor eats the adjustment first and hardest. Vangst counted 412,500 US cannabis jobs in 2026, down 2.7 percent, the first year employment and revenue fell together (Vangst 2026). The modal worker is a budtender earning between seventeen and twenty-six dollars an hour. In Ventura and Santa Barbara counties, farmworkers earned $717 million in 2025 against $4.65 billion of agricultural value produced (Santa Barbara Independent 2026).

A wage share of 15.4 percent, in the sector that was going to be different.

On July 10, 2025, federal immigration agents raided two facilities of Glass House Brands, California’s largest licensed cannabis cultivator, in the largest workplace immigration operation of the administration. Three hundred and sixty-one workers were detained, fourteen of them minors. Jaime Alanis Garcia, fifty-six, fell thirty feet from a greenhouse roof while hiding from agents and died. A year on, local advocates report that lost income and fear have pushed more children into agricultural work. The state sent armed federal officers into the licensed sector, the one it built, while the unlicensed growers the enforcement rhetoric names went untouched.

In November 2024, fifty-nine workers at PharmaCann’s LivWell facility in Warren, Michigan voted to join Teamsters Local 337. The facility shut down within weeks, eliminating all two hundred and twenty-two jobs, the company citing low prices and competition. The following month PharmaCann began defaulting on its IIPR leases; a firm collapsing under rent and price compression closed the plant that organized. Workers at Exclusive Brands in Ann Arbor struck on August 28, 2025 in the longest strike in the history of American legal cannabis; four pro-union workers lost their jobs, the company hired replacements, and the union vote failed. Units have been decertified in Massachusetts and New Jersey. Labor peace agreements, the one structural union foothold in this industry, are under federal attack on National Labor Relations Act preemption grounds, with Oregon’s requirement already struck down.

The rent claim reached the union drive through the balance sheet, without anyone in the boardroom needing to mention the union.

The state chose twice in twelve months

Neither of the two federal actions below required a single corrupt official. Both were defensible on their stated grounds. Both distributed in the same direction, which is the property a structural account predicts and an account built on individual capture cannot explain.

On November 12, 2025, the President signed the appropriations act that ended the government shutdown. Division B redefined hemp. The 2018 Farm Bill had set the threshold at 0.3 percent delta-9 THC while affirmatively legalizing “all derivatives, extracts, cannabinoids, isomers,” which produced the intoxicating hemp sector: delta-8, delta-10, THCA flower, everything convertible from CBD isolate. The new definition measures total THC, excludes synthesized cannabinoids and anything with “similar effects (or marketed as having similar effects) to THC,” and caps finished products at 0.4 milligrams per container. Typical products contain between five and a thousand milligrams. One prominent manufacturer estimates fewer than ten percent would remain compliant (Yakowicz 2025). It takes effect on November 12, 2026.

Whitney Economics put hemp-derived cannabinoid sales at $28.4 billion in 2022, larger that year than all legal medical and adult-use cannabis combined, supporting 328,000 workers and $13 billion in wages. In Texas alone the sector ran 8,500 businesses and 53,300 jobs at a reported seventy-two percent profitability, against 27.3 percent in licensed cannabis. A low-barrier, high-margin, largely non-union competitor operating outside § 280E, outside seed-to-sale tracking, and outside licensure was beating the licensed industry on price.

Price competition was available and went untapped. The answer came as a rider on a shutdown-ending bill, no hearing, written by the senator who legalized hemp in 2018.

Licensed capital used the state to delete a competitor and, in doing so, restored the criminal apparatus to an entire product class that legalization was supposed to retire. That is step three of the operation, ratification, performed against a commons that had escaped the first enclosure by finding a gap in the statute.

The second choice came five months later. On April 23, 2026, the Justice Department placed into Schedule III both FDA-approved cannabis drug products and marijuana “subject to a qualifying state-issued medical license,” using treaty authority under 21 U.S.C. § 811(d) and bypassing notice and comment (US Department of Justice 2026). Everything else, meaning the entire adult-use industry, stayed in Schedule I pending a separate proceeding whose hearing ran from June 29 to July 15, 2026 and whose post-hearing briefs are not due until August 17. Because § 280E applies only to Schedules I and II, the order lifted the punitive tax off medical operators and left it on recreational ones. Trulieve’s effective rate fell from 258 percent to 87 percent. Aggregate § 280E exposure across the major operators had reached $1.6 billion, and the IRS is now litigating to claw back refunds already paid, having sued TerrAscend on May 18, 2026 (MJBizDaily 2026; CRB Monitor 2026b).

Two acts, five months apart, distributing the same way. The state extinguished the cheap competitor and cut the tax burden of capital organized as medicine, which means the incumbents holding vertically integrated medical infrastructure.

Both arrived after the shakeout had already destroyed the small and mid-sized tier. Relief timed like that is a subsidy to survivors.

Intoxicating hemp arguably needed regulating and § 280E was always indefensible. Neither point explains why the correction was scoped and timed to land exactly where it did, and a defender of these decisions has to explain the coincidence without invoking anybody’s bad faith, because there is no bad faith available.

The allotment clause

Twenty-two of twenty-four adult-use states had social equity provisions by 2023, and the National Academies recommended automatic rather than petition-based record relief and systematic evaluation of whether the programs meet their stated goals (National Academies 2024).

The evaluation exists. It is written in the licensing tables, and it has been available the whole time.

Illinois issued eighty-seven craft grower licenses and every one went to a social equity applicant. Twenty-one are operational. Of fifty-five infuser licenses, fifty-four went to equity applicants and sixteen are operational. Twenty-one cultivation center licenses, the high-margin chokepoint of the entire supply chain, sit with incumbent, vertically integrated, largely non-equity operators (IDFPR 2025). The Parabola Center found seventeen incumbents holding twenty of those twenty-one large-scale licenses; multi-state operators moving forty-two percent of units in the fourth quarter of 2025 while capturing sixty-nine percent of revenue, against craft growers moving twenty-seven percent of units for 8.1 percent of revenue; and active brands nearly tripling to 264 while the parent companies behind them fell from ninety-one to seventy-nine (Parabola Center 2026).

The consumer sees 264 brands. There are 79 owners, and the number is falling.

What Illinois distributed to equity applicants was the right to invest capital they did not have, in the exact segment of the chain where prices then collapsed. Three-quarters of the craft growers never opened.

New York made the financing explicit. Its social equity fund, announced at $150 million and administered through a public authority in partnership with Chicago Atlantic, a commercial mortgage REIT, built stores through state-approved vendors and handed justice-involved licensees the resulting multi-million-dollar loan obligations, into a market whose prices then fell below projection while enforcement against unlicensed retail lagged. The Office of Cannabis Management reports that eighty-eight percent of conditional licensees are operational and that equity businesses hold fifty-five percent of all adult-use licenses, above the statutory goal. Both accounts are accurate and they measure different things. The stores opened. Who owns the cash flows is a separate question, and the answer runs through the mortgage REIT.

Massachusetts illustrates the arithmetic in one line. The state and its municipalities have extracted roughly $1.9 billion from cannabis since 2018 and returned $28.9 million through the Cannabis Social Equity Fund. One and a half percent.

Set that beside the 1875 figure. Six-tenths of one percent of the enclosed acreage went to the labouring poor. One and a half percent of the extracted revenue went to the communities the drug war was run against. Two statutes, a hundred and eighty years apart, converging on the same order of magnitude.

And the second half of the 1845 mechanism has been reproduced with unusual fidelity. Section 142 provided for selling an allottee’s land to pay for his fence. New York provided for a mortgage REIT to build the allottee’s store and hold the paper on it.

Parliament wrote the foreclosure into the redistribution. Albany outsourced it and called it a fund.

What the structuralist reading predicts

An interpretation that only explains what has already happened is a label. This one commits to things that have not happened yet, and can be wrong about them.

It predicts that the next wave of failures lands on cultivation and on the most recently licensed cohorts, in New York and Illinois above all, while retail counts stay roughly flat. The filter selects on access to capital and on the ability to survive an unprofitable interval, and the newest entrants have the least of both.

It predicts that full rescheduling, whenever the second proceeding concludes, will be received as relief for the industry and will function as a transfer to survivors. Removing a tax from a sector that has already shed its small and mid-sized tier raises the asset values of whoever is still holding licenses. Watch multi-state operator share prices in the week after, and independent cultivator counts in the year after.

It predicts that equity programs will keep producing ownership churn in place of ownership, on the 1845 schedule. Licenses issued to equity applicants, financed by outside capital on terms the applicant cannot refuse, will pass to that capital within five to seven years through management agreements, foreclosure, and negotiated buyout. Northamptonshire took a generation because land changes hands slowly. A licensed storefront carrying REIT paper will take less. The number to track is beneficial ownership, which almost no state publishes, rather than licenses issued, which every state publishes.

It predicts that if interstate commerce opens without protections in place first, differential rent will reassign cultivation to the cheapest suitable ground within about two seasons, and the appellation regions will not survive it. Salinas and the Central Valley have flat land, water rights, road access, an existing agricultural labor force, and no legacy of hillside parcels conveyed against federal patents.

And it predicts that the current consolidation will end in margin recovery for the firms that survive it, which will be reported as the industry returning to health. Restored margins in a concentrated sector are the completion of centralization.

The account is falsifiable, and the test is already running. If concentration here is driven by particular firms capturing particular regulators, a state with a well-designed equity statute and a clean, competent administration should produce a materially different distribution of ownership from a state with a captured one. If the distribution instead converges toward the same shape across states with very different administrative quality, the structural explanation holds and the corruption explanation has nothing to do with it.

Two dozen states have run that experiment. The results are in the licensing tables, and they are not ambiguous.

Where the industry is right, and where it stops

The industry’s own remedy for price compression is supply management: tie license numbers to demand, pause issuance when capacity runs ahead. There is something to it. Uncontrolled licensing in a market that cannot export produces exactly the boom and bust now underway, caps can prevent wasteful overinvestment and give environmental and labor standards something to stand on, and agricultural supply management has a long and respectable history that the left has generally supported.

The question a Marxist has to press is who receives the scarcity a cap creates.

A capped license becomes property, and property appreciates. Caps protect incumbents as readily as small farms, and they invite the lobbying that produced the Illinois cultivation-center structure in the first place. Unrestricted licensing offers formal equality of entry and lets competition destroy everyone who arrived undercapitalized. Restricted licensing prevents that destruction and hands the resulting rent to whoever already holds a permit. Neither arrangement is socialist. Both decide who collects.

What separates planning from incumbent protection is whether the scarcity is publicly owned: production targets set and published; licenses held as conditional public privileges rather than tradable assets; capacity reserved in advance across large, small, cooperative, tribal, and equity producers; inventory data public so that no participant holds an informational advantage. Supply management with those features is planning.

Without them it is a cartel with a state seal, and a cap is only a fence with fewer gates in it.

What would change the ownership

Every item below is a way of taking something back out of the fences. The alternative has to do more than restore prohibition, and more than celebrate small business while leaving workers and patients subordinate to whoever owns the farm.

Begin with the carceral debt, which is still outstanding: automatic expungement and resentencing, and reinvestment measured against what was extracted. Protect the channels that let people obtain cannabis without transacting with a corporation at all, which means home cultivation, nonprofit social clubs, farmgate sales, farmers’ markets, direct delivery, cooperative provision. Build the credit institutions the federal government has refused to allow, so that an equity licensee has an alternative to surrendering control to a mortgage REIT in exchange for a build-out: public and cooperative banks, patient working capital, shared-equity financing. Graduate the compliance burden and socialize its fixed costs through shared public testing laboratories and common processing infrastructure, keeping the safety standard intact while removing its function as a barrier. Require beneficial-ownership disclosure and count parent companies rather than brands when applying market-share and canopy limits, since 264 brands owned by 79 companies is precisely what disclosure exists to reveal. Separate production from retail, or impose fair-shelf-access rules on integrated firms. Establish sectoral wage standards, union neutrality, enforceable safety rules, and worker representation on the regulatory boards. Set taxes by public-health and distributive goals instead of by an assumption that prices stay high. Protect regional appellations before interstate commerce opens, so that place-based knowledge in Humboldt and Mendocino survives contact with national-scale cultivation. And treat a public retail option as a serious candidate, on the condition that it is governed democratically instead of run as a revenue maximizer on the state liquor model.

That is the same principle worked out in the medical sphere in the Public Domain Whole-Plant Cannabis Medicines Program published in these pages on June 3: the plant’s value can be developed without the developer owning the plant.

The industrial question and the pharmaceutical one are one question, asked at different points in the chain.

The plant was never the point

The sequence runs: mercantile subsidy, agricultural collapse, pharmacy regulation, racialized criminal prohibition, interdiction and the domestic production it created, aerial eradication, the medical gray market, licensed boutique capitalism, speculative overbuilding, price collapse, and consolidation under landlords and bondholders. At no point in four hundred years did the people who grew this plant control the conditions under which it was grown.

Cannabis is a weed. It grows nearly anywhere, cheaply, in enormous quantity, and that abundance could have produced affordable medicine, safe adult use, stable rural livelihoods, and a diverse regional agriculture. Under the present arrangement it registers as catastrophe, because the firms and the lenders and the landlords and the treasuries all need scarcity to make their claims good. Farms close while the harvest grows. Workers are laid off while unit sales rise. Quality declines because the market rewards the cheapest measurable proxy for it, and independent testing has repeatedly found retail flower running well below its labeled potency (Schwabe et al. 2023). Competitors are eliminated by legislation instead of by price. And the survivors wait for concentration to restore the margins competition took away.

Cheap cannabis was never the crisis. The crisis is that cheapness is being administered through foreclosure, layoff, dispossession, and the transfer of productive assets to creditors, rather than through any decision anybody voted on.

The commons here was a library of genetics, a body of practice, and a legal status that fifty years of arrests had paid for in advance. All three have been surveyed and issued as title, and the title is held by people who bore none of the cost of producing what it covers.

Parliament at least had the honesty to call its instrument an act of enclosure. The American version is called a licensing statute, it is administered by agencies with public-health mandates and equity officers, and it works the same.

Prohibition decided who went to prison. Commercial legalization decides who collects the rent. Both were choices, and the second one is still being made.

Who owns the license is the question. Who owns the land is the question. Who owns the harvest is the question.

Disclosure. Del Potter is a working pharmaceutical scientist and the Founder and Chief Scientific Officer of Prism Sciences LLC and BioUnbound Inc. His commercial work intersects directly with subjects this essay addresses, including cannabinoid formulation, delivery-platform intellectual property, regulatory pathways, and the political economy of pharmaceutical access. He holds no interest in any licensed cannabis cultivator, processor, distributor, or retailer, and none in any hemp-derived cannabinoid business, real estate investment trust, or specialty lender named or described here. The Public Domain Whole-Plant Cannabis Medicines Program published in this space on June 3, 2026 is the explicit political-economic alternative to the ownership structure this essay criticizes. Readers should know the intersection exists and weigh the writing accordingly.

A note on figures. Two national revenue estimates for 2025 are in circulation, $29.94 billion from MJBizDaily and $29.1 billion from Vangst and Whitney Economics. They rest on different methodologies and are not interchangeable; both record the same first-ever annual decline. The national wholesale spot index has partially recovered from its 2023 trough because it is reweighted toward newly opened high-price markets, and it should not be read as evidence about state-level prices, which have fallen much further. Whitney Economics is retained by hemp-industry clients, and its estimates of that sector’s size should be read with that in mind; the 2022 figure of $28.4 billion is the best-documented of them.

Companion essays. “How the Cannabis Movement Just Lost the Argument It Won,” June 1, 2026. “Public Domain Whole-Plant Cannabis Medicines Program,” June 3, 2026. “The Plant Has a Brand New Name,” June 17, 2026.

Bibliography

420 Property. n.d. “Sale-Leasebacks in Cannabis: Valuations, Yields, Risks.” Accessed July 27, 2026. https://www.420property.com/sale-leasebacks-in-cannabis-valuations-yields-risks/.

Alexander, Michelle. 2010. The New Jim Crow: Mass Incarceration in the Age of Colorblindness. New York: The New Press.

American Civil Liberties Union. 2013. The War on Marijuana in Black and White: Billions of Dollars Wasted on Racially Biased Arrests. New York: ACLU, June.

American Civil Liberties Union. 2020. A Tale of Two Countries: Racially Targeted Arrests in the Era of Marijuana Reform. New York: ACLU.

Bonnie, Richard J., and Charles H. Whitebread II. 1970. “The Forbidden Fruit and the Tree of Knowledge: An Inquiry into the Legal History of American Marijuana Prohibition.” Virginia Law Review 56 (6): 971–1203.

Brecher, Edward M., and the Editors of Consumer Reports. 1972. Licit and Illicit Drugs. Mount Vernon, NY: Consumers Union.

Butler, Nic. 2020. “Hemp Cultivation in Early South Carolina.” Charleston Time Machine, Charleston County Public Library, April 24. https://www.ccpl.org/charleston-time-machine/hemp-cultivation-early-south-carolina.

California Attorney General. 2008. Guidelines for the Security and Non-Diversion of Marijuana Grown for Medical Use. Sacramento: Office of the Attorney General, August.

California Department of Cannabis Control. n.d. “When Legal Cannabis Is Unavailable, What Fills the Gap?” Accessed July 27, 2026. https://www.cannabis.ca.gov/cannabis-laws/where-cannabis-businesses-are-allowed/when-legal-cannabis-is-unavailable-what-fills-the-gap.

Campaign Against Marijuana Planting. 1984. CAMP 1984 Final Report. Sacramento: California Department of Justice. Cal Poly Humboldt Special Collections.

Cannabis Business Times. 2026. “Oregon Cannabis Wholesaler Withdraws Challenge to Interstate Commerce.” June 17.

Coryell, Roger. 2026. “The Lights Are Flickering Out Across Mendocino County.” The Mendocino Voice, April 25. https://mendovoice.com/2026/04/the-lights-are-flickering-out-across-mendocino-county/.

CRB Monitor. 2026a. “2025 U.S. and Canadian Cannabis Business Licensing Activity.” February.

CRB Monitor. 2026b. “IRS Pushes Back on Cannabis Tax Refunds as Rescheduling Reshapes 280E Fight.” July.

Dewey, Lyster H. 1913. “Hemp.” In Yearbook of the United States Department of Agriculture, 1913. Washington, DC: Government Printing Office.

Dillis, Chris, Margiana Petersen-Rockney, and Michael Polson. 2024. “A Theory of Geo-Social Marginalization: A Case Study of the Licensed Cannabis Industry in California.” Journal of Environmental Management 355.

Drug Science. n.d. “Marijuana Research: Effects of Cannabis Eradication on Cultivation Trends and Availability.” Accessed July 27, 2026. https://www.drugscience.org/Petition/C5H.html.

ERA Economics LLC. 2025. California Cannabis Market Outlook, 2024 Report. Prepared for the California Department of Cannabis Control, January 2025; released March 3, 2026.

Federal Bureau of Investigation. 1968. “Counterintelligence Program, Internal Security, Disruption of the New Left.” Memorandum, Director to SAC Albany and all field offices, July 5.

Frydl, Kathleen J. 2013. The Drug Wars in America, 1940–1973. New York: Cambridge University Press.

Garrod, Joel Z. 2025. “Canadian Political Economy and the Transnational Character of Canada’s Cannabis Industry.” Science & Society 89 (3): 348–71.

Getz, Christy, Margiana Petersen-Rockney, and Michael Polson. 2024. Cannabis Bans, Local Control and the Effects and Efficacy of Proposition 64. Report to the California Department of Cannabis Control, April.

Gieringer, Dale H. 2024. “The Origins of Cannabis Prohibition in California.” Revised edition. Originally published as “The Forgotten Origins of Cannabis Prohibition in California,” Contemporary Drug Problems 26 (2), 1999.

Hall, Stuart, Chas Critcher, Tony Jefferson, John Clarke, and Brian Roberts. 1978. Policing the Crisis: Mugging, the State, and Law and Order. London: Macmillan.

Hansard. 1875. “Enclosure of Lands.” HC Deb, July 23, 1875. Parliamentary Debates, Commons.

Harvey, David. 2003. The New Imperialism. Oxford: Oxford University Press.

Heldring, Leander, James A. Robinson, and Sebastian Vollmer. 2022. The Economic Effects of the English Parliamentary Enclosures. NBER Working Paper 29772.

Hinton, Elizabeth. 2016. From the War on Poverty to the War on Crime: The Making of Mass Incarceration in America. Cambridge, MA: Harvard University Press.

Illinois Department of Financial and Professional Regulation. 2025. Annual Cannabis Report 2025. Springfield, September 30.

Inclosure Act 1845, 8 & 9 Vict. c. 118, §§ 31, 126, 142.

Innovative Industrial Properties, Inc. 2026. Annual Report on Form 10-K for the Fiscal Year Ended December 31, 2025; and First Quarter 2026 Results, May 4.

Jackson, Margaret. 2026. “The Cannabis Industry Is Having a Price Compression Crisis. What’s the Solution?” MJBizDaily, July 27. https://mjbizdaily.com/news/the-cannabis-industry-is-having-a-price-compression-crisis-whats-the-solution/617103/.

Kantor, Shawn, Carl Kitchens, and Steven Pawlowski. 2017. “Civil Asset Forfeiture, Crime, and Police Incentives: Evidence from the Comprehensive Crime Control Act of 1984.” NBER Working Paper 23873, September.

Marx, Karl. (1866) 1976. “Results of the Immediate Process of Production.” Appendix to Capital, vol. 1, translated by Ben Fowkes. London: Penguin.

Marx, Karl. (1867) 1976. Capital: A Critique of Political Economy. Vol. 1. Translated by Ben Fowkes. London: Penguin.

Marx, Karl. (1894) 1981. Capital: A Critique of Political Economy. Vol. 3. Translated by David Fernbach. London: Penguin.

Meinhofer, Angélica, Adrian Rubli, and Jamein P. Cunningham. 2026. “State Recreational Cannabis Laws and Racial Disparities in the Criminal Legal System.” American Economic Journal: Economic Policy 18 (2): 346–74.

Mello, Joseph. 2024. Pot for Profit: Cannabis Legalization, Racial Capitalism, and the Expansion of the Carceral State. Stanford: Stanford University Press.

MJBizDaily. 2026. “Marijuana MSOs’ 280E Bill Grows to $1.6 Billion as IRS Fights Tax Gambit.” March 12.

Morrison Foerster. 2026. “The Cannabist Company’s Chapter 15 Recognition.” May 26.

National Academies of Sciences, Engineering, and Medicine. 2024. Cannabis Policy Impacts Public Health and Health Equity. Washington, DC: The National Academies Press.

National Commission on Marihuana and Drug Abuse. 1972. Marihuana: A Signal of Misunderstanding. Washington, DC: Government Printing Office, March 22.

Neeson, J. M. 1977. “Common Right and Enclosure in Eighteenth-Century Northamptonshire.” PhD diss., University of Warwick.

Neeson, J. M. 1993. Commoners: Common Right, Enclosure and Social Change in England, 1700–1820. Cambridge: Cambridge University Press.

New York Office of Cannabis Management. 2025. Chief Equity Officer Annual Report. Albany, December (data as of November 30, 2025).

New York Office of Cannabis Management. n.d. “Delinquent Payment Reporting.” Accessed July 27, 2026. https://cannabis.ny.gov/delinquent-payment-reporting.

Nixon, Richard. 1971. “The President’s News Conference.” May 1. The American Presidency Project, University of California, Santa Barbara.

Offe, Claus. 1974. “Structural Problems of the Capitalist State: Class Rule and the Political System.” In German Political Studies, vol. 1, edited by Klaus von Beyme, 31–57. London: Sage.

Ogden, David W. 2009. Memorandum for Selected United States Attorneys: Investigations and Prosecutions in States Authorizing the Medical Use of Marijuana. Washington, DC: US Department of Justice, October 19.

Parabola Center for Law and Policy. 2026. Illinois Cannabis Market Concentration: A Report on Incumbent Control, Market Entry, and Structural Inequality. May. https://www.parabolacenter.com/illinoismarket.

Polanyi, Karl. 1944. The Great Transformation: The Political and Economic Origins of Our Time. New York: Farrar and Rinehart.

Polson, Michael, and Hekia Bodwitch. 2021. “Why Comply? Farmer Motivations and Barriers in Cannabis Agriculture.” Journal of Rural Studies 86: 489–504.

Polson, Michael, Ann Laudati, and Nathan F. Sayre. 2024. Transformation of Unregulated Cannabis Cultivation Under Proposition 64. Final report to the California Department of Cannabis Control, Agreement #065306. Berkeley: University of California, April 1.

Poulantzas, Nicos. (1978) 2000. State, Power, Socialism. Translated by Patrick Camiller. London: Verso.

Santa Barbara Independent. 2026. “One Year After the Glass House Raid, California Still Fails Farmworkers.” July 15.

Schwabe, Anna L., Vanessa Johnson, Joshua Harrelson, and Mitchell E. McGlaughlin. 2023. “Uncomfortably High: Testing Reveals Inflated THC Potency on Retail Cannabis Labels.” PLoS ONE 18 (4): e0282396.

Scott-Goforth, Grant. 2013. “Subdividing Humboldt.” North Coast Journal, July 11.

Shaw-Taylor, Leigh. 2001. “Parliamentary Enclosure and the Emergence of an English Agricultural Proletariat.” Journal of Economic History 61 (3): 640–62.

Smith, John L., Jr. 2016. “The Truth About George Washington and Hemp.” Journal of the American Revolution, October 6.

Sullum, Jacob. 2013. “Pot Protectionism in Colorado.” Reason, February 22.

Tate, W. E., and Michael Turner. 1978. A Domesday of English Enclosure Acts and Awards. Reading: University of Reading.

TIME. 1972. “The Law: The Agitator.” March 13.

UK Parliament. n.d. “Enclosing the Land.” Living Heritage. Accessed July 29, 2026.

US Attorney’s Office, Central District of California. 2011. Press release on coordinated marijuana enforcement action, October 7.

US Department of Justice. 2026. “Justice Department Places FDA-Approved Marijuana Products and Products Containing Marijuana Subject to a Qualifying State-Issued License in Schedule III.” Office of Public Affairs, April 23.

Valdes-Donoso, Pablo, Daniel A. Sumner, and Robin Goldstein. 2020. “Costs of Cannabis Testing Compliance: Assessing Mandatory Testing in the California Cannabis Market.” PLoS ONE 15 (4): e0232041.

Vangst. 2026. The 2026 U.S. Cannabis Jobs Report. June 23. https://vangst.com/cannabis-jobs-report-2026/.

Whitney Economics. 2024a. Cannabis Delinquencies: An Existential Threat to the U.S. Cannabis Industry. March 29.

Whitney Economics. 2024b. Cannabis Business Conditions Survey. July.

Wiener, Jon. 2000. Gimme Some Truth: The John Lennon FBI Files. Berkeley: University of California Press.

Woodward, William C. 1937. Statement before the House Committee on Ways and Means on H.R. 6385, May 4. 75th Congress, 1st Session.

Yakowicz, Will. 2025. “$28 Billion Hemp Industry Faces Extinction with Government Re-Opening.” Forbes, November 13.

Cases and statutes. Marihuana Tax Act of 1937, Pub. L. No. 75-238, 50 Stat. 551. Controlled Substances Act of 1970, Pub. L. No. 91-513. Comprehensive Crime Control Act of 1984, Pub. L. No. 98-473. Anti-Drug Abuse Act of 1986, Pub. L. No. 99-570; 21 U.S.C. § 841(b)(1). Agriculture Improvement Act of 2018, § 10113; 7 U.S.C. § 1639o. Continuing Appropriations and Extensions Act, 2026, Pub. L. No. 119-37, div. B. Cal. Health & Safety Code §§ 11362.5, 11362.775. RCW 69.50.328. Leary v. United States, 395 U.S. 6 (1969). NORML v. Bell, 488 F. Supp. 123 (D.D.C. 1980). Northeast Patients Group v. United Cannabis Patients and Caregivers of Maine, 45 F.4th 542 (1st Cir. 2022). Variscite NY Four, LLC v. New York State Cannabis Control Board (2d Cir. 2025). Peridot Tree WA Inc. v. Washington State Liquor and Cannabis Control Board, No. 24-3481 (9th Cir. 2026). Johnson v. Beto, 337 F. Supp. 1371 (S.D. Tex. 1972), aff’d 469 F.2d 1396 (5th Cir. 1972). People v. Sinclair, 387 Mich. 91 (1972). United States v. U.S. District Court, 407 U.S. 297 (1972). Colorado Retail Marijuana Code, 1 CCR 212-2. Wash. Admin. Code 314-55.

There is a woman in Fort Bragg, Jeanine Storm, 81 years old, with Stage 4 cancer and her life packed in boxes by the door of a trailer she has been told to leave. Her landlord, asked to explain, said: “It’s my two houses, my two acres.”

I keep turning that sentence over, because it is so confident. It arrives with the full weight of four centuries behind it. My two acres. The woman who said it did not invent the idea that a deed outweighs a dying tenant; she inherited it, the way we all did, and the inheritance has a history.

The history is enclosure. I want to trace it here at some length, because I have come to believe it is the master key to what is happening in rural places like mine, and because the word has almost vanished from the American vocabulary at precisely the moment we need it most.

The hedge

Enclosure was, before it was anything else, a hedge: hawthorn, usually, planted in a line across land where no line had been. On one side of the hedge, what had been. On the other, what was coming.

What had been was the commons. The commons has been slandered for so long that even sympathetic people picture it wrong. The commons was not a free-for-all, not an unowned waste, not the tragedy of the economics textbooks. It was the most intricately governed land in England. The commoner’s cow could graze, but only so many cows, only in season, only by rules argued out in manor courts for generations. The right to glean after harvest, the right to gather fallen wood but not to cut, the right to cut turf, to run pigs in the oak woods in acorn season: each right had its name, its limits, its custodians. E.P. Thompson spent the last decades of his life documenting this lattice of customary right, and his conclusion was: what the poor held, they held not as charity and not as property, but as custom, which is to say as law that lived in memory and practice rather than parchment.

Custom’s weakness was that it could not be sold. That was also its function.

Between 1750 and 1850 Parliament passed roughly four thousand acts of enclosure, converting nearly seven million acres of commons into private, hedged, alienable property. The commoners were compensated, where they were compensated at all, in proportion to what they could document, and custom by its nature documents nothing. The cottager who had kept a cow, gathered fuel, and gleaned enough to bridge the winter received a strip of land too small to live on and too burdened with legal fees to keep, then sold it, then went to work for the man who bought it. This was the sequence, repeated a million times, that created the English working class: not machinery first, but dispossession first. The factory did not lure the peasant off the land. The hedge drove her to the factory.

The people watching understood exactly what it meant. John Clare, the field laborer who became the finest poet of the English countryside, wrote it plainly: “Inclosure came and trampled on the grave / Of labour’s rights and left the poor a slave.” Clare ended his life in an asylum, and it is not sentimental to connect the two facts. His subject, his home ground, the whole web of paths and commons that made his world legible, was fenced while he watched.

And Marx, reading the record in the British Museum a generation later, gave the process its analytic name: primitive accumulation, the original sin of the system, “written in the annals of mankind in letters of blood and fire.” His account of the Duchess of Sutherland, who between 1814 and 1820 cleared fifteen thousand people from nearly eight hundred thousand acres of the Scottish Highlands and burned the villages behind them to run sheep, reads today like reportage. The Duchess did not think of herself as a monster. She thought of herself as an improver. The land was hers; the arithmetic favored sheep; the people could find another place to live.

My two houses. My two acres.

What was actually taken

Here is where the philosophy comes in, but the standard account, even the standard left account, undersells the loss.

The usual framing is economic: the commons were an income supplement, enclosure removed it, poverty followed. True, and yet the commoners themselves, in their petitions and riots and in the testimony Thompson recovered, grieved something the economic framing cannot hold. They had lost a relationship to place that was constitutive of who they were. The commons was where you knew and were known. Its calendar was your calendar. Rights in it attached to you not as an owner but as a member, and membership, unlike title, cannot be foreclosed, only extinguished.

Karl Polanyi, writing in exile while the world the enclosures built was burning in the 1940s, got closest to the heart of it. Land, he argued, is a fictitious commodity. Labor, land, and money are all fictitious commodities: things treated by the market as if they had been produced for sale, when in fact they are simply human beings, the earth, and the tokens of trust between us. The fiction is workable in small doses and catastrophic in large ones, because when you subject land fully to the market you are not pricing a product, you are pricing the substrate of life itself, and the people standing on it become, in the market’s eyes, an encumbrance. Every eviction notice is the fiction asserting itself: the premise that shelter was never shelter at all, only an asset temporarily obstructed by a human being.

What enclosure abolished, in other words, was an entire competing theory of ownership: that use confers standing, that long presence ripens into right, that the earth, in Gerrard Winstanley’s phrase from 1649, was made a common treasury for all. The Diggers were crushed in a season. Their idea has been crushed in every season since, and it keeps coming back, because it corresponds to something people stubbornly feel: that fifty years of living and working in a place must count for something. Jeannie Storm’s whole predicament, the shock of it, the wrongness she feels in her body, is the collision between that felt law and the recorded kind.

Letters of blood and fire, California edition

Americans tend to hear “enclosure” as English history, quaint and hedged. But the American continent is the largest enclosure ever carried out, and Mendocino County is not a bystander to that history. It is a roped off crime scene of it.

The land my county sits on was Pomo, Yuki, Cahto, Coast Yuki land, held for millennia under exactly the kind of tenure the enclosers could not see: managed intensively, burned deliberately, harvested on calendars older than English itself, and owned by no one in the sense a deed office would recognize. So it was declared empty. What followed in the 1850s and 60s was not settlement but clearance: militia raids subsidized by the state of California, massacre, indenture under the grotesquely named Act for the Government and Protection of Indians, and finally concentration. The Mendocino Indian Reservation was established at Noyo in 1856 and dissolved a decade later, its land passed to timber interests; the survivors were marched inland to Round Valley, on a trail their descendants still walk in memorial each year. Then the reservations themselves were enclosed: the Dawes Act of 1887 broke communal tribal land into individual allotments, precisely so that it could be sold, and two-thirds of all Native land in the country, some ninety million acres, passed out of Native hands within two generations. Parliament would have recognized the technique immediately. Survey, allot, individuate, acquire.

And on the cleared land, the second enclosure: the timber companies. For a century the Mendocino Coast was effectively a proprietary landscape, company towns on company land, workers housed in company houses that came with the job and went with it. The foreman’s house where Jeannie Storm lived for twenty years, until her partner died and the ranch gave her ninety days, is a fossil of that system, still functioning exactly as designed. Housing tied to labor is housing that disciplines labor, and it evicts the widow as efficiently in 2024 as in 1924.

Now the third enclosure, the one we are living through, which is the strangest of the three because it happens without a single act of visible violence. No militia, no hedge, no burned village. Just a price.

The enclosure of shelter

The Midnight Notes Collective, in 1990, called what was coming “the new enclosures,” and the phrase has proven more durable than the pamphlet. They meant the whole planetary wave: structural adjustment stripping communal land in Africa, privatization devouring the socialist bloc, the fencing of everything from seed genetics to the human genome. The commons, it turns out, is not a stage of history that ends. It is a standing temptation. Wherever people hold anything in common, by custom, membership, or mere inertia, capital experiences that holding as an inefficiency, a resource pinned beneath people who are not paying rent on it.

In rural America the commons under enclosure is housing itself. Not the deeds; the deeds were always private. What is being enclosed is the informal, customary layer that actually functioned as a commons of shelter: the trailer rented on a handshake, the cabin that never quite appeared on the assessor’s map, the granny unit, the room over the shop, the foreman’s house, the whole shadow stock where the rural poor and the rural old have always actually lived. That layer persisted for decades because the land beneath it wasn’t worth optimizing. The amenity economy changed the arithmetic. A structure that shelters a caregiver for $300 a month grosses that in a weekend on Airbnb. A house is worth more empty, appreciating in the fog eleven months a year, than occupied by anyone who works here. On my coast nearly a fifth of the housing stands vacant, most of it seasonal, while the schools shrink for want of families and the senior housing waitlists stretch to six years, which for an 81-year-old is not a waitlist but a wager against her own survival.

Notice that no law was broken and no one conspired. That is what makes this enclosure so hard to see and so hard to fight. The eighteenth-century commoner could at least riot against the hedge; the hedge had a location. This hedge is a spread between two revenue streams. It runs through every parcel simultaneously. Its enforcement mechanism is not the sheriff, though the sheriff is available, but the simple, apparently neutral proposition that property must be free to seek its highest use. Polanyi’s fiction, fully automated.

And the ideological work is done by that little possessive pronoun. My two acres. The genius of enclosure was never the fence; it was the conversion of a relationship into a commodity. Once land is a thing, the questions that custom asked, who has lived here, who has worked here, what do we owe each other, become category errors, sentimental noise interfering with a title search. The landlady in Fort Bragg is not an aberration of the system. She is its fluent native speaker.

Re-commoning

I do not want to end in the elegiac register, because elegy is one more thing the enclosers are happy to grant us. Grieve the commons all you like, so long as you concede it is gone.

It is not gone. The commons was never a place; it was a practice, the practice of holding something so that membership rather than money governs access, and the practice keeps reasserting itself wherever the fiction of the land market produces results too obscene to live with. A community land trust is a commons: land taken off the market permanently, housing sold or rented atop it under rules the community writes. A limited-equity co-op is a commons. A mobile home park bought by its residents is a commons, and there are hundreds now. Social housing, the kind Vienna has run for a century, is a commons administered at municipal scale. Even the humble things, the food bank garden, the volunteer fire department, the water association, are commons, fragments of the old lattice of customary right, maintained by people who mostly do not know they are practicing a rival theory of property.

The philosophical point is as important as the policy one. Every generation is taught that the current distribution of the earth is a fact of nature, and every generation the teaching requires more effort, because the arrangement keeps producing scenes that refuse to be naturalized. An old woman with cancer and her son’s photographs in a box, told by the person enclosing her shelter that she can find another place to live, when both of them know there is no other place: that scene is not complicated. Fifty years on this coast, caring for its bodies. Set against it: a deed, and the confidence of four centuries.

The commoners of Northamptonshire, watching the surveyors come through, did not have the vocabulary of ground rent and fictitious commodities. They had a simpler formulation, which survives in a rhyme from the seventeenth century: The law locks up the man or woman who steals the goose from off the common, but leaves the greater villain loose who steals the common from off the goose.

We are the goose. It is late, but not too late, to steal the common back. It begins, as it always has, locally: an ordinance, a trust, a tax on emptiness, a refusal to accept that a waitlist is a housing policy or that an answering machine message is a fate. The hedge was planted by hand. It comes out the same way.

Companion piece to my op-ed on Jeannie Storm’s eviction and county-level remedies. Sources and further reading: Elise Cox’s reporting at MendoLocal.News (July 2026); E.P. Thompson, “Customs in Common”; Karl Marx, “Capital” Vol. I, Part VIII; Karl Polanyi, “The Great Transformation”; Midnight Notes Collective, “The New Enclosures” (1990); Cal Winslow, “Mendocino Coast: The Crisis in Affordable Housing” (Mendocino Institute, 2002); Benjamin Madley, “An American Genocide: The United States and the California Indian Catastrophe”; John Clare, “The Mores.”

The purpose of a system is what it does. What chemical agriculture does is enclose.

For a decade the argument over glyphosate has been fought on a field Big Agriculture chose. Is it carcinogenic, or is it not? Courts, regulators, and the press have litigated that single question until it became the whole controversy, and while they argued the more important fact went unexamined: that the cancer question, whatever its answer, is a diversion from the purpose of the enterprise. The purpose of chemical-based agriculture is not yield, and it is emphatically not feeding the world. Its purpose is enclosure, the conversion of a self-reproducing commons, the seed and the soil and the accumulated agronomic knowledge of ten thousand harvests, into a stream of private, rent-bearing property. Glyphosate is not, in the first instance, a herbicide. It is the chemical latch on a proprietary seed, the device by which the last biological commons was fenced. To argue about whether the latch causes lymphoma while ignoring what it locks is to accept the terms of the people who built the fence.

This essay makes the case for caution, on glyphosate specifically and on the whole edifice of chemical herbicides and insecticides, but it refuses the narrow toxicological frame in which that case is usually confined. The harm to human health is real and contested; the harm to soil, water, and non-target life is real and larger; the resistance treadmill is a certainty; and the promise that all of this is required to feed humanity is a fiction. But these are not separate findings. They are facets of a single process, and the name of the process is enclosure. Read that way, the evidence stops being a scattered pile of grievances and becomes what it is: the anatomy of a dispossession.

The Old Technique

Enclosure is not a metaphor borrowed for effect. It is a genealogy with a documented history. Between the fifteenth and nineteenth centuries the open fields and common lands of England were fenced, consolidated, and made private by custom, by force, and finally by some four thousand Acts of Parliament. Marx called the whole movement the “so-called primitive accumulation” (Capital, Volume I): the violent separation of the agricultural population from the land, which produced at a single stroke both the consolidated private estate and the propertyless labourer who, owning nothing that could feed him, had no choice but to sell his hours. The common lands were fenced because they were ungovernable by capital. A household that could graze a cow, gather fuel, and glean a field could not be fully compelled into the wage relation; it retained a margin of independence, and independence is precisely what accumulation cannot tolerate. Enclosure closed that margin.

The justification offered then is the justification offered now. It was called improvement, the doctrine, elaborated in a whole eighteenth-century literature, that land held in common was wasted and that private, rationalized management would extract more from it for the benefit of all. The improvers were not lying about the yields; enclosed and capitalized farms often did produce more per acre. They were lying about the beneficiary. The abundance flowed to the landlord and the improving tenant; the commoner received a cottage economy destroyed and a one-way ticket to the manufacturing towns. The improvement pamphlet of 1770 is the direct literary ancestor of the “feed the world” brochure of 2010. Both dress expropriation in the vocabulary of productivity, and both count the harvest while refusing to count the dispossessed.

One commons resisted longest, because it resisted by its nature: the seed. A seed is an anomaly among commodities, for it is at once the product sold and the means of producing next year’s product. The farmer who harvests wheat harvests next season’s seed inside this season’s grain. So long as farmers saved seed, the market for seed collapsed to nearly nothing each year, a permanent, biological barrier to accumulation, a commons that reproduced itself for free. Jack Kloppenburg’s First the Seed(1988) is the definitive account of how capital spent the twentieth century trying to breach that barrier. The first breach was hybridization: hybrid seed does not breed true, so its saved offspring are agronomically worthless, and the farmer must return to the dealer every spring. That was a biological lock on the commons. The second breach, when biotechnology arrived, was legal and absolute: the patent. A genetically engineered seed could be claimed as an invention, its saving and replanting forbidden by law.

It is here that glyphosate enters, and its real function becomes visible. The engineered trait that conquered the market was herbicide tolerance, Roundup Ready, a gene that lets the crop survive a drenching of glyphosate that kills everything green around it. Consider what that couples. The farmer who buys the patented, glyphosate-tolerant seed is bound to buy the branded glyphosate that the same corporation sells, and forbidden by patent law to save the seed for next year. Seed and chemical, sold by one firm, locking the farmer into an annual repurchase of both. Roundup Ready was never an agronomic breakthrough. It was an enclosure device: the fence, the lock, and the toll booth in a single package, and its brilliance was to make the seed, the one commons that had always reproduced itself, depend on a proprietary chemical to be worth planting at all.

The Instrument and Its Alibi

Against this background the famous cancer fight looks less like the heart of the matter and more like its most visible distraction. In 2015 the International Agency for Research on Cancer classified glyphosate as “probably carcinogenic to humans,” Group 2A. IARC conducts a hazard assessment, whether a substance is capable, under realistic conditions, of causing cancer. The European Food Safety Authority and the U.S. Environmental Protection Agency conduct risk assessments, whether expected exposures are likely to. These are genuinely different questions. A hazard can be real while the risk at typical exposure remains modest; conceding this is the price of being believed about everything else.

Within the hazard frame, the strongest human evidence is the meta-analysis by Zhang and colleagues (Mutation Research, 2019), which pooled the updated Agricultural Health Study cohort with five case-control studies and reported a 41% increase in the relative risk of non-Hodgkin lymphoma among the most highly exposed (meta-RR ≈ 1.41; 95% CI 1.13–1.75). That is a signal warranting concern, not a settled causal dose-response: it concerns the highest-exposure stratum and depends on defensible but contestable analytic choices. Two further facts sharpen it. First, toxicity is a property of the formulation, not the pure active ingredient: commercial Roundup contains surfactants such as polyethoxylated tallowamine (POEA) that are, in many assays, markedly more toxic to cells and non-target organisms than glyphosate alone, yet much regulatory testing was conducted on the naked molecule, systematically understating what is actually sprayed. Second, and decisive for the argument of this essay, the safety science itself was enclosed.

The “Monsanto Papers,” disgorged in U.S. litigation, documented the corporation ghostwriting ostensibly independent reviews and coordinating with a senior EPA official, Jess Rowland, who was positioned to shape, and to slow, the agency’s cancer assessment. The influential review by Williams, Kroes and Munro (2000), cited for two decades as authoritative evidence that glyphosate was “not likely” carcinogenic, was formally retracted in 2025 after disclosures that Monsanto personnel had substantially authored it and that it leaned on unpublished, company-held studies. This is the same operation applied to knowledge that the patent applied to the seed: the privatization of a commons. Independent science, the public’s capacity to know what it is being exposed to. was converted into a proprietary asset, manufactured to specification and released for public consumption under a borrowed name. When regulators then lean on confidential, industry-generated dossiers, the “scientific consensus” they cite is partly a product they were sold.

The regulatory record reflects the contest rather than resolving it. In June 2022 the U.S. Court of Appeals for the Ninth Circuit vacated the human-health portion of EPA’s interim decision, holding that the agency’s “not likely to be carcinogenic” finding was not supported by substantial evidence and contradicted its own admission that it could not rule out a lymphoma association; EPA withdrew the decision. In Europe, EFSA reported in 2023 that it found “no critical areas of concern,” and the Commission renewed glyphosate for ten years, but only after member states twice failed to reach a qualified majority in either direction, and only alongside EFSA’s own acknowledgment of unresolved data gaps, including an inability to conclude on biodiversity. The honest summary is not “science says safe.” It is “contested, captured, and hedged,” which is the epistemic fingerprint of enclosure: manufacture the consensus, privatize the evidence, and then dare the public to prove harm to a standard the privatization has made unreachable.

The Enclosure Reaches the Soil

Enclosure never respects the property line. Having fenced the seed and the science, the chemical model externalizes its costs onto commons that cannot be fenced because no one holds title to them, the soil beneath the field, the water leaving it, the insects crossing it. Glyphosate kills by inhibiting the enzyme EPSPS in the shikimate pathway, which builds aromatic amino acids. The industry’s original alibi was that this pathway is absent in animals, true, and that the compound is therefore selectively harmless. But the shikimate pathway is emphatically present in bacteria and fungi. A herbicide marketed as inert to animal life acts directly on the biochemistry of the soil microbiome, the living commons on which the fertility of every field depends. This is the single most consequential and least discussed fact in the debate.

The consequences are accumulating. Studies report shifts in soil microbial communities after repeated exposure, with suppression of beneficial taxa such as Pseudomonas and Bacillus, reduced root colonization by arbuscular mycorrhizal fungi after even a few applications, and disruption of nitrogen-fixing and phosphate-solubilizing organisms. The direction of harm is consistent across studies; the magnitude and durability remain genuinely debated. On soil fauna the evidence is compelling: Gaupp-Berghausen and colleagues (Scientific Reports, 2015) found that the surface-casting activity of deep-burrowing earthworms nearly ceased three weeks after a glyphosate-based herbicide was applied, that reproduction of soil-dwelling Aporrectodea caliginosa fell by 56% within three months, and that soil nitrate rose by roughly 1,600%, a signature of a nutrient cycle thrown open to leaching.

Nor does the compound stay put. U.S. Geological Survey sampling finds glyphosate and its persistent degradate AMPA to be frequent and widespread across soils, streams, groundwater, and even rainfall, with AMPA often outlasting the parent molecule. Carried into water, the toxicity meets organisms far more sensitive than mammals. Relyea (Ecological Applications, 2005) found that a Roundup formulation produced a 70% decline in amphibian species richness and an 86% decline in tadpole biomass in experimental communities, an effect driven substantially by the POEA surfactant, which is exactly why formulation-blind assessment is inadequate. And because the shikimate pathway lives in the bee gut as well as the soil, glyphosate is not inert to pollinators: Motta, Raymann and Moran (PNAS, 2018) showed that field-relevant exposure perturbs the honeybee gut microbiome and raises mortality under subsequent pathogen challenge, with other work reporting sublethal impairment of foraging and navigation, the quiet damage that acute-lethality testing is built to miss.

The monarch butterfly is the emblem of the landscape-scale version of this harm. Milkweed, the sole host plant of monarch larvae, has been all but eliminated from Corn Belt crop fields. Iowa studies estimate declines near 98% between 1999 and 2012 as glyphosate-tolerant corn and soybeans enabled total in-field weed clearance, and monarch numbers fell in close parallel. The correlation is strong and biologically plausible. It is also contested: a 2019 analysis argued that monarch and milkweed declines predate the adoption of glyphosate-tolerant crops, implicating climate and migratory-corridor mortality as well. The defensible claim is the careful one, herbicide-driven habitat simplification is a major contributor, not the sole cause, and it is more than enough to make the point. The biotic commons is the sink into which the costs of private profit are poured. No one owns the soil’s microbes or the roadside milkweed, so no one is billed for their destruction. The gain is privatized; the loss is socialized across the living world. That is not a flaw in the technology. It is the logic of enclosure, working exactly as designed.

The Treadmill Is the Business Model

Blanket a continent with a single chemistry, season after season, and resistance is not a risk but a mathematical certainty. Any weed that survives glyphosate enjoys an overwhelming reproductive advantage and bequeaths its survival to the next generation; resistance is not the failure of the system but its inevitable product. More than forty weed species have now evolved resistance to glyphosate worldwide, by the reckoning of Ian Heap’s International Survey of Herbicide-Resistant Weeds, and in the United States the infested acreage climbed from roughly 32 million acres in 2010 to more than 60 million by 2012. Palmer amaranth and waterhemp, each capable of hundreds of thousands of seeds per plant, became agronomic emergencies across the South and Midwest.

The industry’s answer to the failure of its chemical was more chemistry, and older, nastier chemistry at that. As glyphosate lost its grip, the seed-and-chemical firms rolled out crops stacked with tolerance to dicamba and 2,4-D, the latter a component of the Vietnam-era defoliant Agent Orange, the former so prone to volatilize and drift that it has set neighbour against neighbour across the farm belt and damaged millions of acres of non-target crops. Weeds resistant to these modes of action already exist. This is not a malfunction to be corrected; it is the engine. Each cycle of resistance manufactures the demand for the next proprietary seed-and-chemical package, and the treadmill turns the biological certainty of resistance into a reliable schedule of repurchase. Planned obsolescence, the oldest trick in consumer capitalism, here operates on the genome of the food supply. The same dynamic drives insecticide dependence, where heavy reliance on neonicotinoid seed treatments and Bt crops has selected for resistant pests while imposing broad non-target costs.

And the enclosure is enforced, in the end, by law. In Bowman v. Monsanto (2013) the U.S. Supreme Court held unanimously that the doctrine of patent exhaustion does not permit a farmer to save and replant patented seed. Vernon Bowman, an Indiana farmer, had bought commodity soybeans from a grain elevator, beans meant for the feed mill, planted them, and saved the descendants; because those beans happened to carry the Roundup Ready trait, the Court found he had unlawfully “made” Monsanto’s patented invention by the act of growing it. In Monsanto Canada v. Schmeiser (2004) the Supreme Court of Canada held, five to four, that a canola farmer’s saving and planting of seed containing the patented gene constituted infringing “use,” even amid questions of how the gene arrived in his fields. Read plainly, these rulings criminalize the founding practice of agriculture. The saving of seed, the ten-thousand-year mechanism by which the agricultural commons reproduced itself for free, was converted into a tort. The commoner who once gleaned the field is now the infringer who planted it.

The fence, moreover, has very few owners. After Bayer swallowed Monsanto in 2018, Dow and DuPont merged and spun out Corteva, and ChemChina absorbed Syngenta, four firms: Bayer, Corteva, Syngenta, and BASF, came to control on the order of half to sixty percent of the global commercial seed market and sixty to seventy percent of the agrochemical market (USDA; Public Eye; Heinrich Böll Foundation). A generation ago the biggest four seed firms held roughly a fifth of the market; the ratchet has only ever turned one way. The seed of the species, the literal germ of the human food supply, the inheritance of every farming culture that ever lived, now has landlords, and they are the same handful of chemical companies that sell the sprays the seed is engineered to require.

The Hunger Alibi

Which returns us to the alibi, the moral cover under which the whole operation proceeds: that the fence is a form of charity, and that without proprietary seed and its attendant chemistry the world would starve. This is the major claim of the industry’s self-defense, and it is the weakest thing it says. Begin with yield. The most authoritative review available, the U.S. National Academies of Sciences’ Genetically Engineered Crops: Experiences and Prospects (2016), found no evidence that genetic engineering increased the rate of yield gain in U.S. crops. Yields have risen steadily since long before biotechnology and continued at much the same pace after it; the herbicide-tolerant trait added management convenience and, by the committee’s own finding, increased overall herbicide use. Glyphosate tolerance is a logistics technology. It protects yield potential that weeds would otherwise take, as any effective weed control would, but it does not raise the ceiling, and the claim that it feeds a growing world by making plants more productive is simply not supported by the evidence.

Then confront the premise beneath the premise: that hunger is caused by insufficient food. It is not, and has not been for a very long time. The world already produces more than 2,300 kilocalories per person per day, enough, in aggregate, for well over ten billion people (FAO). People go hungry because of poverty, distribution, conflict, and access, not because the earth grows too little; roughly a third of all food is lost or wasted, and of the cereals actually used, only about 42% feeds people directly while the rest goes to livestock and industry. Producing still more patented commodity grain does not reach the hungry. It flows into the channels that already absorb the surplus. Follow the crop and the pretense collapses: glyphosate is sprayed overwhelmingly on corn, soybeans, cotton, and canola, and in the United States roughly 35% of the corn goes to animal feed and about a third to ethanol, while the majority of soybeans are exported or crushed for feed and fuel. Meat, fuel, and traded commodities, not the plates of the food-insecure poor.

And the cost of stepping back is modest, which is the fact the industry most needs to obscure. The organic-to-conventional yield gap, a deliberately punishing benchmark, since organic forgoes every synthetic input at once, averages about 19% (Ponisio et al., Proc. R. Soc. B, 2015) to 25% (Seufert et al., Nature, 2012), and narrows to roughly 8–9% under diversified management such as multi-cropping and rotation. Integrated pest management, which keeps herbicides as a targeted last resort rather than a calendar ritual, recovers most of that yield while shedding most of the externalities catalogued above. So “feed the world” is to glyphosate exactly what “improvement” was to the enclosure acts: the moral vocabulary that obscures expropriation as productivity. It was a lie in the eighteenth century. the enclosures produced not abundance for the commoner but a landless proletariat driven into the mills, and it is a lie now, told by the firms that collect the rent, about crops that are mostly not food, to justify a fence that feeds their balance sheets rather than the hungry.

Against the Chemical Enclosure

No single thread in this indictment is decisive on its own, and the industry’s strategy has always been to force its critics to argue each in isolation, to demand of environmental science a standard of individual certainty it rarely delivers, and to treat every unproven particular as an acquittal of the whole. But the argument was never in any single thread. It is in their convergence, and in the structure of the risk they jointly describe: an exposure that is planetary and effectively irreversible; harms that are systemic and slow: microbiome disruption, biodiversity loss, resistance evolution, cheap to cause and enormously dear to undo; and a burden of proof distributed backwards, so that the public and the biosphere must prove injury while the compound remains in universal use. That inversion is not a neutral feature of regulatory caution. It is itself an enclosure strategy, the privatization of the benefit of the doubt for capital, and the socialization of the risk onto everyone and everything that cannot hire a lobbyist.

The precautionary principle, developed in European scholarship and embedded in EU environmental law, was built for exactly this configuration: where an activity threatens serious or irreversible harm, scientific uncertainty about the precise magnitude of that harm is a reason for protective action, not a license for delay. Applied here it does not require the melodramatic claim that glyphosate is a proven poison. It requires only the sober conclusion that a compound acting on the biochemistry of soil, water, and pollinator life, sprayed at continental scale, breeding resistance as a matter of evolutionary law, defended by a literature of engineered independence, and justified by a humanitarian claim the evidence refutes, should not enjoy the benefit of the doubt.

But precaution is only the defensive half of the answer. The offensive half is to refuse the enclosure outright. That means seed sovereignty in the sense Vandana Shiva has fought for, the farmer’s recovered right to save, share, and breed seed, against the patents that made those acts a crime. It means agroecology and integrated pest management that rebuild the biological commons the chemical model depletes, rather than renting it back to farmers one proprietary trait at a time. It means public, unpatented plant breeding that treats the seed as what it is: not the asset of four firms but the inheritance of the species. None of this is nostalgia for the moldboard plow, and none of it denies that a just transition carries real costs. Labor, knowledge, and land that decades of chemical dependence have allowed to atrophy. The point is that those costs are transitional and bounded, while the costs of the enclosure are structural and compounding.

So the case for caution on glyphosate turns out to be far larger than glyphosate. It is a stand on the oldest question in political economy: whether the ground of life is a commons we hold or a rent we pay. The cancer headline was never the strongest argument against this chemical. The strongest argument is that we have handed over the seed, first the seed, and with it the food, the soil, and the science that was supposed to guard them, to the same technique that once took the fields, and we have let it call the taking generosity. The enclosures were reversed once, in law and in struggle, by people who refused to accept that the commons had to become property. The fence around the seed is newer, and its owners are richer, but it is a fence like any other. It can be taken down.

Sources and Further Reading

Marx, K. Capital, Volume I, Part VIII, “So-Called Primitive Accumulation” — the enclosure of the commons and the expropriation of the agricultural population.

Kloppenburg, J.R. First the Seed: The Political Economy of Plant Biotechnology (1988; 2nd ed. 2004) — hybridization and patents as strategies to commodify the self-reproducing seed.

Shiva, V. Seed Sovereignty, Food Security (and related works) — seeds as commons, patents as “biopiracy,” the farmer’s right to save and breed seed.

IARC (2015). Monographs Vol. 112 — glyphosate classified Group 2A, “probably carcinogenic to humans” (hazard assessment).

Zhang, L., et al. (2019). Exposure to glyphosate-based herbicides and risk for non-Hodgkin lymphoma. Mutation Research 781: 186–206 (meta-RR 1.41; 95% CI 1.13–1.75, highest-exposure).

The “Monsanto Papers” (U.S. litigation discovery); U.S. Right to Know documentation; retraction (2025) of Williams, Kroes & Munro (2000), Regulatory Toxicology and Pharmacology; EPA official Jess Rowland correspondence.

National Family Farm Coalition v. EPA, 9th Cir. (June 17, 2022) — vacated the human-health portion of EPA’s glyphosate interim decision; EPA subsequently withdrew it.

EFSA (2023). Peer review of the pesticide risk assessment of glyphosate, EFSA Journal 21(7):8164 — “no critical areas of concern,” with unresolved data gaps incl. biodiversity; EU 10-year renewal (Nov 2023) after two inconclusive member-state votes.

Gaupp-Berghausen, M., et al. (2015). Glyphosate-based herbicides reduce the activity and reproduction of earthworms… Scientific Reports 5: 12886.

Relyea, R.A. (2005). The lethal impact of Roundup on aquatic and terrestrial amphibians. Ecological Applications 15(4): 1118–1124.

Motta, E.V.S., Raymann, K., Moran, N.A. (2018). Glyphosate perturbs the gut microbiota of honey bees. PNAS 115(41): 10305–10310.

U.S. Geological Survey (Battaglin et al.). Glyphosate and AMPA occurrence in U.S. soils, surface water, groundwater, and precipitation.

Pleasants, J.M. & Oberhauser, K.S. (2013). Milkweed loss and the monarch butterfly. Insect Conservation and Diversity 6: 135–144. Counterpoint: Boyle et al. (2019), PNAS — declines predate GM-crop adoption.

Heap, I. The International Survey of Herbicide-Resistant Weeds (weedscience.org) — >40 glyphosate-resistant weed species.

Bowman v. Monsanto Co., 569 U.S. 278 (2013) — patent exhaustion does not permit replanting saved patented seed (9–0). Monsanto Canada Inc. v. Schmeiser, 2004 SCC 34 — saving/planting patented-gene seed is infringing “use” (5–4).

National Academies of Sciences (2016). Genetically Engineered Crops: Experiences and Prospects — no evidence GE raised the rate of yield gain; herbicide-tolerant crops increased herbicide use.

FAO — global food production exceeds need (>2,300 kcal/person/day); hunger driven by poverty, distribution, waste. USDA / American Farm Bureau — U.S. corn and soybean end-uses (feed, ethanol, export).

Ponisio, L.C., et al. (2015), Proc. R. Soc. B 282:20141396; Seufert, V., et al. (2012), Nature 485:229–232 — organic-conventional yield gap ~19–25%, narrowing to ~8–9% under diversification.

On industry consolidation: USDA ERS; Public Eye; Heinrich Böll Foundation — four firms (Bayer, Corteva, Syngenta, BASF) control ~50–60% of seed and ~60–70% of agrochemical markets.

In May of 2023 a crate of root bark left the forests of Gabon under paperwork, legally, for the first time. It was bound for a clinical-stage company in Vancouver, and it traveled with the consent of the state and a set of documents asserting that the people from whose land and tradition the root had come would receive something in return. The root was iboga, the Tabernanthe iboga whose bark the Bwiti have eaten as a sacrament for longer than anyone can document, the plant that holds, in the Bwiti understanding, the door to the country of the ancestors. The crate was the first export of iboga ever certified under the Nagoya Protocol, the international instrument that is supposed to ensure that when the genetic and traditional resources of the poor are turned into the medicines of the rich, the poor are paid. Gabon was the first country on earth to sign that protocol, and it signed it, in large part, to protect this root.

This crate represents one of the two responses that have emerged from the communities and the advocates who understand that the molecules at the center of this series are not the property of the companies racing to own them. The two responses look, from a distance, like the same response. They are not. One of them sells the commons back to capital as a royalty. The other denies that the commons was ever capital’s to enclose. The distance between them is the subject of this essay, because the distance between them is the difference between improving the terms of a defeat and refusing the defeat.

II.

The structural fact underneath both responses is the one the series has returned to from the beginning. The foundational psychedelic molecules cannot be patented, because they are ancient and because they entered human use through indigenous traditions that the patent system does not recognize as inventors. What makes those molecules into medicine, the container, the cosmology, the relational and ceremonial knowledge of how to hold a person through the experience and back into a life, is held by communities the same system does not recognize as producers of knowledge. The Bwiti nganga, the Mazatec curandera, the Wixárika who walk to Wirikuta: these are the inventors and the knowledge-holders, in any honest accounting, and they hold inheritances that the law has no category for and the market has every intention of mining.

Two answers to this injustice are now in the field. The first answer is reciprocity: bring the communities inside the transaction, give them consent and a contract and a share of the revenue, and let the medicine go to market on terms that pay its origin. The second answer is refusal: deny that the knowledge can be enclosed at all, keep it in the open where no patent can fence it, and starve the enclosure of the legal monopoly it requires. Reciprocity works inside the market. Refusal works on the question of whether the market gets to exist here in the first place. They are two different politics, and only the second is a politics in the full sense, because only the second contests the legitimacy of the enclosure rather than negotiating its price.

III.

Take reciprocity first, on its strongest terms. The organization Blessings of the Forest was the first to apply the Nagoya Protocol to a psychedelic substance, and the arrangement it built with Gabonese stakeholders is, by every measure of the thing it is trying to be, an improvement on the alternative, which is theft. Through 2026 Gabon has moved toward a legal, regulated export of sustainable iboga, and the advocates have begun to speak of the next and harder step, which is to compel, by law if persuasion fails, the ibogaine clinics and the pharmaceutical companies that have built businesses on this root to enter into some form of reciprocity with the people whose root it is. National Geographic surveyed the scene and reported the obvious truth in its headline: the drug could save lives, and everyone wants to cash in. Reciprocity is the attempt to make sure that when the cashing-in happens, the forest is not robbed in silence.

And yet. Vandana Shiva, who has spent a working lifetime on exactly this terrain, taught us to see what the benefit-sharing contract concedes in the act of signing. To accept a royalty for a sacred inheritance is to accept that the inheritance is the kind of thing a royalty can be paid on, which is to say a commodity, a fungible asset with a price, a unit of property that can be licensed. Polanyi called land, labor, and money the fictitious commodities, the things that were never produced for sale and that the market nonetheless insists on treating as though they were, and the cosmology of the Bwiti is a fictitious commodity of the purest kind. Reciprocity prices it. Reciprocity says: here is what your ancestors’ door is worth per kilogram, here is your share, sign here. It can fund a chapter house and pay for the conservation of a threatened plant and put money into hands that have had too little of it, and all of that is real and none of it is nothing. But it does not refuse the enclosure. It is the commons agreeing to be enclosed on better terms, and the better terms are still enclosure. The check clears, and the logic that the sacred is property survives the clearing, intact, and ready for the next root.

IV.

Refusal has two modes, and the difference between them is as significant as the difference between refusal and reciprocity.

The first mode is defensive, and it works through prior art. A patent requires novelty; a thing already known to the public cannot be newly invented and fenced. Therefore, if the knowledge a company seeks to patent can be shown to predate the company, the patent fails. Shiva and her colleagues demonstrated the power of this mode a generation ago, when the patents granted in the global North on the neem tree and on turmeric, both of them medicines that Indian tradition had used for millennia, were challenged and revoked on the strength of documented prior use, and when India then built the Traditional Knowledge Digital Library, a vast inscription of its traditional medicine, precisely so that the patent offices of the world would have the prior art in front of them and would have to say no. Prior art turns the antiquity of the commons into a legal weapon. The knowledge is too old to own, and its age is the defense.

In psychedelics this mode has a standard-bearer. The nonprofit Freedom to Operate, led by Carey Turnbull, challenged the patents that Compass Pathways holds on its crystalline psilocybin, filing for post-grant review and arguing that the claimed invention was not novel. The Patent Trial and Appeal Board declined to undo the patents, and on the surface Compass won. But read the result the way Turnbull read it. The board construed Compass’s claims so narrowly, holding that they cover only psilocybin exhibiting the exact diffraction signature specified, that, in Turnbull’s account, generic manufacturers are left with wide latitude to produce and sell psilocybin without infringing. The patent stood and was hollowed in the same ruling. This is what defensive refusal achieves at its most realistic: not the abolition of the monopoly, but its confinement, the narrowing of the fence until most of the field lies outside it. It is lawyer’s work, partial and unglamorous, and it is one of the few things that has actually constrained the enclosure rather than merely lamenting it.

The second mode of refusal is affirmative, and it does not wait for the patent to be filed. It places the knowledge in the public domain in advance, deliberately, so that the fence can never be built, the way a programmer releases code under a license designed to keep it free. The Usona Institute is the working example. Usona is a nonprofit, it conducts its psilocybin research as open science, and when its chemists developed an efficient method for synthesizing pharmaceutical-grade psilocybin at scale, they published the method and placed it in the public domain, where it cannot be patented, where any researcher may use it, where it functions as a permanent obstacle to anyone who would claim that ground as their own. Usona does not patent the psilocybin it brings to trial. It dedicates. And here the essay arrives at the fact that made me want to write it: in April of 2026, the same federal apparatus I described in the companion essay handed Usona one of its three priority vouchers, the same accelerant it gave to Compass. The commons has been admitted to the corridor. A public-domain psilocybin program is now racing a patent-thicket psilocybin program toward the same market, on the same fast track, and the question of which model wins, the dedication or the thicket, is going to be answered in public over the next two years. I do not know the answer. I know that it is the most interesting unresolved question in the field, and that almost no one on the left is watching it.

V.

But there is a trap inside the defensive mode of refusal, and the trap connects this essay to the one before it in the series, the essay about the algorithmic container, and I would be concealing the hardest part of the argument if I did not name it.

To defend a commons through prior art, you must make the commons legible. You must write the knowledge down, inscribe it, document it, render it into a form a patent examiner can read and a tribunal can cite. The Traditional Knowledge Digital Library is exactly such an inscription, and it works, and it is also a complete, searchable, machine-readable corpus of traditional medicine, sitting in a database, available to be read by more than examiners. James C. Scott spent two great books on the proposition that legibility is the precondition of both protection and capture, that the same act of making a population or a forest or a body of knowledge readable to authority is what allows authority to both defend it and seize it, and that communities have often survived precisely by remaining illegible, by keeping their knowledge in forms the state and the market could not parse. The defensive refusal of enclosure requires the legibility that the deeper enclosure feeds on. The Bwiti category written into a Stanford grounded-theory paper so that the world may know the tradition is old, and may use that age against a future patent, is the same Bwiti category, in the same machine-readable form, that trains the integration chatbot I described last time. To inscribe the commons in order to defend it from the patent is to deliver it to the model. The shield and the extraction run on the same documentation.

VI.

Which is why the deepest refusal is the one that does not inscribe at all, and why it cannot, by its nature, be a complete strategy. There is a register of traditional knowledge that has survived for centuries by being withheld, transmitted only through initiation, spoken only in the ceremony, never written because writing it would end it. Fred Moten and Stefano Harney describe this register and the social life that protects it: the undercommons, the fugitive knowledge that persists by refusing the terms on which it would be recognized, that declines the invitation to become legible, fundable, citable, scalable, ownable. Scott called the broader phenomenon the art of not being governed. The clandestine guide, the uninitiated who is never told the last thing, the curandera who will show the anthropologist the song but not its meaning: these are practitioners of a refusal the apparatus cannot ingest, because the apparatus can only ingest what has been written, and they have refused to write it.

Legible refusal, the prior art and the public-domain dedication, can be extracted even as it defends, because it lives in documents. Illegible refusal, the undercommons, cannot be extracted, but it also cannot defend itself in a tribunal, cannot fund a clinic, cannot scale to meet the suffering the molecules are supposed to address, and cannot stop a single patent, because it will not enter the room where patents are decided. The traditions that are safest from the model are the ones that are least able to fight it on the field where it is winning. This is not a paradox to be resolved with a clever phrase. It is the actual shape of the terrain, and any politics that pretends otherwise is selling something.

VII.

So the two faces of the commons are really, on inspection, three faces, and they array along a single axis. At one end is reciprocity, which accepts the commodity form and bargains over the price, and which is materially better than theft and politically a surrender. In the middle is legible refusal, in its defensive and its affirmative modes, prior art and dedication, which contest the enclosure on the enclosure’s own ground, win real and partial victories, and pay for those victories in the coin of legibility, which the deeper enclosure collects. At the far end is illegible refusal, the undercommons, which cannot be enclosed and cannot be deployed, which keeps the sacred safe by keeping it dark, and which can hold a tradition together for centuries while doing nothing whatever to stop the corporation across the ocean.

Peter Linebaugh, the great historian of the commons, insisted that commoning is a verb, not a noun, an activity and a set of relations rather than a stock of assets, and that the enclosures of history succeeded precisely by converting the verb into the noun, by turning a living practice of shared use into a thing that could be parceled, titled, and sold. Reciprocity is the conversion of the verb into the noun with a royalty attached. Refusal, in both its registers, is the attempt to keep the verb a verb. That is the real distinction the field has been struggling to articulate, and it is why the benefit-sharing contract, for all its decency, cannot be the resolution. A check is a noun. The Bwiti hold a verb.

VIII.

What, then, would a politics adequate to this look like, given that no single one of the three faces is sufficient and the easiest one is a trap dressed as a solution? It would not choose among them. It would sequence them, and it would keep the order straight.

It would dedicate before it defends: place the syntheses and the protocols and the methods in the public domain in advance, the Usona move, so that the fence cannot be built, because a fence prevented is worth ten fences litigated. It would defend where dedication came too late: back the prior-art challenges, fund the Freedom to Operate of every molecule, narrow every thicket that has already been raised. It would treat reciprocity as a floor and never as a ceiling: pay the Bwiti, pay them more than the market would, pay them as restitution rather than as license, but never permit the payment to be mistaken for justice or to launder the commodity form it travels inside. And it would do the thing none of the legal mechanisms can do, the thing that is finally the whole point, which is to resource the living traditions and the relational workforce directly and unconditionally, so that the curandera and the nganga and the integration worker and the underground guide do not have to sell their legibility to survive, do not have to choose between starving in the undercommons and being inscribed into the training corpus, because a society that wanted these medicines to heal rather than to pay rent would simply fund the people who know how to hold them.

That last clause is the alliance the previous essay ended on, restated from the side of the commons rather than the side of the workforce, and they are the same alliance seen from two windows. The crate of root bark that left Gabon under paperwork in 2023 was the best version of the answer the existing order can produce, and the existing order’s best version is a priced and licensed and consented enclosure of a thing that should never have had a price. The forest is still there. The root still opens the door it has always opened. What is owed to the people who keep that door is not a percentage. It is the recognition that the door was never merchandise, and the resources to go on keeping it without having to sell the keeping. Reciprocity offers them a royalty. Refusal offers them a verb. We should be on the side of the verb.

♦

Bibliography

I. The Commons and Its Enclosure

Linebaugh, Peter. The Magna Carta Manifesto: Liberties and Commons for All. Berkeley: University of California Press, 2008. (On commoning as activity and the conversion of the verb into the noun.)

Federici, Silvia. Re-enchanting the World: Feminism and the Politics of the Commons. Oakland: PM Press, 2019.

Hyde, Lewis. Common as Air: Revolution, Art, and Ownership. New York: Farrar, Straus and Giroux, 2010. (On the cultural commons and the defense of the public domain.)

Polanyi, Karl. The Great Transformation. [1944.] Boston: Beacon Press, 2001. (On the fictitious commodities.)

II. Biopiracy, Prior Art, and Traditional Knowledge

Shiva, Vandana. Biopiracy: The Plunder of Nature and Knowledge. Boston: South End Press, 1997. (On the neem and turmeric patent challenges and the prior-art defense of traditional knowledge.)

Convention on Biological Diversity. Nagoya Protocol on Access to Genetic Resources and the Fair and Equitable Sharing of Benefits Arising from their Utilization. 2010.

III. Refusal and Illegibility

Scott, James C. Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed.New Haven: Yale University Press, 1998.

Scott, James C. The Art of Not Being Governed: An Anarchist History of Upland Southeast Asia. New Haven: Yale University Press, 2009.

Harney, Stefano, and Fred Moten. The Undercommons: Fugitive Planning and Black Study. Wivenhoe / New York / Port Watson: Minor Compositions, 2013.

IV. Ethnography of the Bwiti and Iboga

Fernandez, James W. Bwiti: An Ethnography of the Religious Imagination in Africa. Princeton: Princeton University Press, 1982.

V. Reporting and Primary Sources

“Gabon Takes First Step Toward Legal Export of Sustainable Iboga.” Lucid News, 2026. (Source for the 2026 export developments and the move toward compelled reciprocity.)

“The psychoactive drug ibogaine could save lives, and everyone wants to cash in.” National Geographic.(Source for the Gabonese fair-trade scene and the 2023 Nagoya-compliant export to Filament Health.)

“The Promise of Nagoya: Indigenous Reciprocity in the Psychedelic Renaissance.” Georgetown Journal of International Law blog. (On Blessings of the Forest and the application of the Nagoya Protocol to iboga.)

“Psychedelics Could Become Extractive Capitalism, Unless We Hold Stakeholders Accountable.” DoubleBlind.

“Usona Puts New Psilocybin Synthesis Method in the Public Domain.” Lucid News. (Source for Usona’s affirmative dedication.)

“As COMPASS Pathways continues its attempts to build a psychedelic monopoly, Usona Institute keeps publishing open research.” Psymposia.

Freedom to Operate, Inc. Statements regarding the Patent Trial and Appeal Board’s response to the petitions for post-grant review of Compass’s psilocybin patents, 2022 (PR Newswire). See also “Appeal Board Upholds Compass Pathways’ Patent on Synthetic Psilocybin Formulation,” and coverage of the narrowed Polymorph A construction. (Source for the Turnbull quotation on generic latitude.)

Hardman, Josh. “Breaking: FDA Awards Priority Review Vouchers to Otsuka, Compass, and Usona.” Psychedelic Alpha, April 24, 2026. (Source for Usona’s receipt of a CNPV alongside Compass.)

I.

On the morning of Saturday, April 18, 2026, in a ceremony staged with the particular informality this administration reserves for its largest interventions, the President signed an executive order on the acceleration of medical treatments for serious mental illness, and his FDA Commissioner, standing nearby, explained to the press that the agency would shortly issue a small number of vouchers to psychedelic drug programs, and that three of them would be awarded right away. Six days later the three arrived. Compass Pathways received one for its synthetic psilocybin in treatment-resistant depression. The nonprofit Usona Institute received one for its psilocybin in major depression. The third, the one that surprised the people who follow this field for a living, went to Otsuka, the Japanese pharmaceutical company, for a molecule called methylone, an MDMA-like compound that Otsuka had acquired along with the small company Transcend Therapeutics in a deal valued at one and a quarter billion dollars.

I want to hold the third voucher in view for most of this essay, because the third voucher tells you what the first two are for. But I want to begin with the object itself, because the object is new, and the series I have been writing has not yet had occasion to describe it. What is a Commissioner’s National Priority Voucher, and why is its arrival the event I am claiming it to be?

II.

The series has, to this point, located the enclosure of psychedelic medicine at two sites. The first site is the molecule. The apparatus moved into psychedelics intending to patent the compound, run the trials, secure the approvals, and sell the dose, and it discovered the difficulty that the foundational molecules, psilocybin and mescaline and ibogaine and DMT, entered human use before the patent office existed and cannot be owned. The second site is the container, the relational and communal labor that makes the molecule therapeutic rather than merely intoxicating, and I argued in the preceding essay that the apparatus, unable to patent slow communal labor, has built a container of its own in software, an algorithmic stand-in trained on the extracted traces of the human relation it replaces. The molecule it could not own outright. The container it could not buy, so it manufactured a substitute. These were the first and the second moves.

The voucher is the third part of the strategy, and it operates at a third site, which is neither the molecule nor the container but the pathway: the regulatory route by which a drug crosses from candidate to product, from the laboratory to the prescription pad and the reimbursement code. The pathway has always been understood, by the people who defend it and the people who attack it alike, as a rule-bound corridor, slow and expensive and frustrating, but governed by published criteria of safety and efficacy that apply, in principle, to all comers equally. What the voucher does is convert a stretch of that corridor into something allocated by decision rather than by rule. This is the enclosure of the pathway, and it is, by my count, the seventh enclosure.

III.

The Commissioner’s National Priority Voucher pilot was launched in June of 2025, and it compresses the review of a new drug application from the conventional ten to twelve months down to one or two. It is awarded at the discretion of the Commissioner. It is tied, in the program’s own language, to United States based manufacturing, to excellence in chemistry and controls, and to what the agency calls credible affordability commitments. And it is, in a detail that distinguishes it sharply from the priority review vouchers that preceded it, non-transferable and non-tradable. It cannot be sold. There is no secondary market.

The obvious reading of a government voucher is that it is a financial instrument, a thing with a price, and the older priority review vouchers were exactly that: a company that earned one for developing a drug against a neglected tropical disease could turn around and sell it to a different company for a sum that ran, in the active years of that market, past a hundred million dollars. Those vouchers were a commodity. The Commissioner’s voucher is not. It cannot be detached from the company that receives it and converted into cash. Its value is locked to the discretion that grants it.

And that is why it is the more dangerous object. A voucher that can be sold is, for all its faults, a voucher governed by a market, and a market, whatever else it is, is a mechanism that operates without asking the seller’s politics. A voucher that cannot be sold, that can only be granted, is a voucher governed by the grantor’s judgment, and the grantor here is a Commissioner who serves at the pleasure of a President. The first new molecule ever to receive one of these vouchers was not a psychedelic at all. It was Eli Lilly’s orforglipron, a daily pill for weight management, approved fifty days after its application was submitted. The national priority, in other words, is whatever the Commissioner decides the national priority is. The criterion is the decision.

IV.

Karl Polanyi, writing in 1944 from the wreckage the self-regulating market had made of the European century, established the proposition that the contemporary defenders of the free market have spent eighty years trying to bury, which is that there is nothing natural or spontaneous about a market, and that the great market society of the nineteenth century was not a thing that grew but a thing that was built, deliberately, by states, through legislation, against enormous resistance, at the cost of immense human suffering. Laissez-faire, in Polanyi’s flat and devastating phrase, was planned. The market did not arise and then get regulated; the market was constituted by regulation, called into being by the same state power that the market’s ideologues insisted should stand aside.

The psychedelic medicine market is being constituted in front of us, in real time, by exactly this kind of state action, and the voucher is the instrument of the construction. There was no market for licensed psychedelic therapeutics in the United States until the apparatus of the state began assembling one: the breakthrough therapy designations, the executive order, the real-world evidence infrastructure I described in the previous essay, and now the vouchers that compress the final approval. Each of these is a public act. None of them is the market discovering an efficiency. The drugs do not approach approval because an invisible hand has weighed their merits. They approach approval because the state has decided, as a matter of policy set at the highest level, that they should, and has built the corridor through which they will pass. The fiction that the FDA merely certifies the safe and the effective, neutrally, at arm’s length, dissolves the moment you watch the Commissioner hand three companies a shortcut the others do not get. The agency is not certifying a market. The agency is making one, and choosing its first occupants.

V.

There is a German jurist whose name the left invokes with the proper revulsion and then, having invoked it, cannot stop using, because he described the thing too well. Carl Schmitt wrote that the sovereign is the one who decides on the exception, the one who stands, in the moment that matters, above the rule and determines when the rule applies and to whom. The voucher is a Schmittian object. It is the codified power to make an exception to the ordinary corridor, and the way it has been administered in its first months has displayed the exception in its pure form.

In the autumn of 2025, Compass Pathways appeared on a draft list of voucher recipients, and the White House, at the last moment, vetoed its inclusion, reportedly out of concern for how the President’s political base would receive the spectacle of the administration blessing a psychedelic corporation. Then, in the spring of 2026, the White House reversed itself, and Compass received its voucher after all. The reversal has been traced, in the reporting, to a text message exchange between the President and a podcaster. FDA staff and members of Congress have said openly that the White House was too directly involved in decisions that are supposed to turn on safety and efficacy and nothing else, and one critic described it: there is a taint here of political involvement, and none of it has to do with the criteria that are supposed to govern. The voucher to Otsuka’s methylone, the least advanced of the three programs, a drug only now entering its final trials, was awarded over the actual MDMA program of the company once called Lykos and now renamed Resilient, the program the field had expected to win. The corridor was opened for one molecule and kept closed for another, and the difference was not maturity, and was not evidence.

This is what the enclosure of the pathway looks like in practice. It looks like the apparatus that Giorgio Agamben taught us to call by that name, the dispositif, the arrangement of laws and institutions and discourses through which power administers life, operating now in the register of the exception, granting and withholding the corridor by a judgment that answers to the sovereign and not to the rule. The pathway has been captured, and the capture is invisible to the population because the corridor is still there, still publishing its criteria, still pretending to be neutral procedure, while the decisions that matter are made above it.

VI.

Returning now to the third voucher, Otsuka acquired Transcend Therapeutics, and with it the methylone program, and the methylone program is the one the agency chose to fast-track for post-traumatic stress disorder, in preference to the MDMA program that had spent a decade and a fortune trying to bring the actual entactogen, wedded to actual psychotherapy, through the actual trials. In August of 2024 that MDMA program was rejected. The agency’s letter, made public a year later, cited a failure to collect favorable adverse events, a lack of durability data, and the high proportion of trial participants who had used MDMA before. But underneath the procedural findings was a structural fact that the field understood immediately and that the subsequent behavior of capital has confirmed. The MDMA program failed in part because MDMA is old, unpatentable, and inextricable from the therapy that surrounds it, and the therapy that surrounds it is the relation, the container, the thing that cannot be standardized, cannot be scaled, cannot be owned, and cannot be controlled for in the way a regulator trained on small molecules expects.

The lesson the industry drew from that rejection was: do not bring the regulator a generic molecule married to a messy human relation. Bring a novel analog, patentable, manufactured to specification, with the relation stripped out and replaced by something the trial design can hold constant. Methylone is that analog. It is to MDMA what the algorithmically generated molecules of Mindstate and Psylo are to psilocybin: a compound engineered, among its other purposes, to be ownable, and to need as little of the unpatentable human container as the molecule can be made to need. The market selected against MDMA because MDMA could not be enclosed. It selected for methylone because methylone can. And the voucher, the state’s instrument, ratified the selection, fast-tracking the analog and snubbing the original. The molecule-side enclosure that the series described at the level of corporate strategy has now been confirmed at the level of state policy. The two sites are welded. The analog wins the corridor.

VII.

The executive order that produced the vouchers carried a second provision that belongs in this essay, because it shows the same logic operating on the body of the patient rather than on the body of the corporation. The order commits fifty million federal dollars to ibogaine research and instructs the agencies to build a pathway, through Right to Try and through the DEA, by which eligible patients may obtain ibogaine in what the order calls advanced, licensed, hospital-grade settings. Ibogaine is the molecule of the Bwiti, the root the series has returned to before, and it is also a molecule with a known and documented capacity to stop the human heart. The legal analysts have flagged the cardiac shadow. The order accelerates access to it anyway.

The state can manufacture access to a molecule in a matter of months, by executive order, with fifty million dollars and a stroke of the pen. The container that would make that molecule safe, the screening and the preparation and the trained human presence and the longitudinal community that holds the patient before and during and after, cannot be manufactured by executive order, cannot be conjured with fifty million dollars, and is not, in the order, even attempted. The access is fast because access is what the pathway is built to produce. The container is slow because the container is the thing the apparatus has spent this entire series failing to industrialize. So the molecule arrives at the kitchen table, or the hospital-grade setting, ahead of the relation that would make it medicine, and the gap between them is absorbed, as such gaps are always absorbed, by the body of the person at the table.

We have already seen what the absorption looks like, because the ketamine telehealth sector ran the experiment first. Mindbloom, the largest of the at-home ketamine providers, has facilitated something on the order of six hundred thousand unsupervised dosing sessions, and Mindbloom is now defending a wrongful death suit brought by the father of a man who overdosed at home, a man with a history of substance use and a racing heart, who had missed appointments and fallen behind on his subscription payments, and who was permitted to dose himself anyway. The pathway delivered the molecule. The relation that would have caught him was the part the model had disposed of. The executive order’s ibogaine provision is the same architecture, scaled up, with a more dangerous molecule and the full weight of the state behind the acceleration.

VIII.

There remains the matter of the affordability commitment, the criterion the voucher program names as one of its conditions, and I raise it last because it is the alibi the whole arrangement requires. The voucher is granted, in part, on the strength of a company’s credible commitment to make the resulting drug affordable, and this commitment is what allows the program to present itself as an instrument of access, a public-spirited acceleration of treatment for the suffering and the underserved and, above all, for the veteran, who appears in every official document as the human face of the policy.

Set the commitment against the only real-world data we have on who actually receives these treatments when they reach the market. In Oregon, where regulated psilocybin services have operated since 2023, the clients who disclose their income are disproportionately people earning more than ninety-five thousand dollars a year, clustered in the cities, while the service centers close under the weight of ten-thousand-dollar license fees and the rural population goes without. This is the distribution the actually existing market produces, and there is no reason to expect the federally accelerated market to produce a different one, because the affordability commitment is a promise made to the grantor, enforceable at the grantor’s discretion, by companies whose entire strategy is the capture of a patentable, reimbursable, premium-priced product. David Harvey gave the process its name. Accumulation by dispossession is the conversion of a common inheritance, in this case the ancient molecules and the relational knowledge of the traditions that carried them, into a stream of private rent, and the affordability commitment is the rhetorical cushion that lets the conversion proceed while wearing the language of the public good. The veteran is the warrant. The rent is the point.

IX.

I began with a signing ceremony and I will end with the figure the ceremony was staged to invoke, the veteran, who is real, whose suffering is real, and whose abandonment by the existing mental health system is the genuine fact on which the entire architecture is erected and which the architecture exploits. The veteran will, in the months ahead, receive faster access to molecules than he has ever had. He will not receive the container. The pathway has been built to deliver the first and was never designed to deliver the second, and the affordability commitment will not change this, and the executive order will not change this, and the voucher, which can accelerate a weight-loss pill to approval in fifty days, will not be spent on the slow communal labor that would actually hold him, because that labor cannot be patented, cannot be billed per session at scale, and cannot be made into a product that returns rent to a shareholder.

The political content of this essay is contained in a single observation, and Polanyi supplies it. If the market for psychedelic medicine is being constituted by state action, deliberately, from above, by the same discretionary power that fast-tracked methylone and orforglipron, then that power could constitute the market otherwise. It could fund the container instead of accelerating the analog. It could pour fifty million dollars into the training and the fair compensation of the relational workforce instead of into a molecule that stops the heart. It could build a pathway to public provision rather than a corridor to private rent. The state is already doing the building. The question the voucher forces, and the reason the voucher is the enclosure, is the question of who the building is for. The corridor has been opened. We are entitled to ask why it was opened for them, and to demand that it be opened, instead, for us.

♦

Bibliography

I. Theoretical and Critical Sources

Agamben, Giorgio. What Is an Apparatus? and Other Essays. Translated by David Kishik and Stefan Pedatella. Stanford: Stanford University Press, 2009.

Harvey, David. The New Imperialism. Oxford: Oxford University Press, 2003. (See the chapter on accumulation by dispossession.)

Marx, Karl. Capital: A Critique of Political Economy, Volume One. [1867.] Translated by Ben Fowkes. London: Penguin, 1976. (Part Eight, “So-Called Primitive Accumulation,” on the enclosure of the commons.)

Polanyi, Karl. The Great Transformation: The Political and Economic Origins of Our Time. [1944.] Boston: Beacon Press, 2001. (On the planned character of laissez-faire and the state’s constitution of the market.)

Schmitt, Carl. Political Theology: Four Chapters on the Concept of Sovereignty. [1922.] Translated by George Schwab. Chicago: University of Chicago Press, 2005. (On the sovereign decision and the exception.)

II. Policy Documents and Legislation

Executive Order No. 14401, “Accelerating Medical Treatments for Serious Mental Illness.” Signed April 18, 2026. Federal Register 91 (April 22, 2026): 21,709. (Citation as given in the preceding essay; verify against the final Register text.)

U.S. Food and Drug Administration. “Commissioner’s National Priority Voucher (CNPV) Pilot Program.” Program launched June 17, 2025. FDA industry and FAQ pages, fda.gov.

U.S. Food and Drug Administration. “FDA Accelerates Action on Treatments for Serious Mental Illness Following Executive Order.” Press announcement, April 2026.

III. Reporting on the Vouchers and the Field

Hardman, Josh. “Breaking: FDA Awards Priority Review Vouchers to Otsuka, Compass, and Usona.” Psychedelic Alpha, April 24, 2026. (Source for the three CNPV awards, the Otsuka/Transcend acquisition, the Compass veto and reversal, and the Resilient/Lykos context.)

“FDA moves to fast-track review of psilocybin and methylone for mental health.” CNN, April 24, 2026.

“FDA rejected MDMA-assisted PTSD therapy. Other psychedelics firms intend to avoid that fate.” Science / AAAS. (Source for the industry’s strategic turn toward patentable analogs.)

“At first public hearing for FDA’s CNPV program, industry support meets ‘deep concern’ from experts.” Fierce Biotech. (Source for expert criticism of the voucher program.)

Frier Levitt. “The Psychedelics Executive Order: Priority Review Vouchers, Ibogaine’s Cardiac Shadow, and the Road Ahead.” Firm publication, 2026.

IV. The Ketamine Precedent

“Telehealth ketamine companies put users at risk.” STAT News, March 2, 2026.

“Telehealth ketamine service sued after patient dies of overdose.” HealthExec, 2026. (The Mindbloom wrongful death suit; the Phillip Ward case.)

V. Access and Distribution Data

“Oregon psilocybin therapy clients tend to be wealthier, new data suggests.” OPB, July 3, 2025.

“The Oregon Psilocybin Services Tracker.” Psychedelic Alpha, 2026.

There is an old technique, older than Taibbi, older than the Substack it now pays for. The Stalinist prosecutors of the 1930s perfected it and described it in the dissident literature: the amalgam. You take the accused, the engineer who missed a production quota, and you bind him by rhetorical association to the saboteur, the foreign agent, or the assassin. The engineer never met the assassin. No matter. They are now one defendant. The crime of the worst is transferred, by proximity and innuendo, onto the merely inconvenient.

Taibbi’s essay on the New York primaries is an amalgam from its first paragraph to its last, because the piece is built to be enjoyed rather than examined. But we should examine it.

What the candidates actually ran on
Start with the boring part, the part Taibbi just races past, because it is fatal to his thesis. What did the people who won this week actually campaign on?

Darializa Avila Chevalier, who unseated five-term incumbent Adriano Espaillat in NY-13, ran under the slogan “Babies, Not Bombs.” Her platform was a rent-stabilization fight, a Green New Deal for public housing to repair the rotting NYCHA stock her constituents actually live in, immigrant protection, and an expanded safety net (Democracy Now!; Norwood News). Brad Lander, whom Taibbi files under “DSA candidates” though he is not a DSA member, ran on the affordability agenda of the Mamdani mayoralty he had endorsed (NBC News). And the movement’s avatar, Mamdani himself, governs on a program of a rent freeze for two million stabilized tenants, fare-free buses, universal childcare from six weeks to five years, a thirty-dollar minimum wage, and a pilot of five city-owned grocery stores (CNN; Time Out).

This is the “whole-hog, fuck-it, real-thing version of Marxian socialism, instantly recognizable to anyone who’s ever lived in a socialist country.” Free buses. Daycare. Cheaper milk. I guess the tanks will come later.

Taibbi knows this platform is unremarkable, which is why he cannot quote it. Instead he reaches past the candidates to a deleted student Instagram post, a six-year-old Twitch clip from a man who is on no ballot, and a programmatic resolution authored by a caucus most DSA members have never heard of. This is the amalgam: bind the daycare advocate to the streamer, transfer the crime, and hope the reader does not notice the substitution.

The “elite fixation” claim, against the returns
The main assertion of the essay is that this is not a popular movement at all but “an obvious elite fixation,” a plaything of “America’s nitwit rich,” beloved of Mark Ruffalo and Sunny Hostin and the “Democratic spa set.” It is a satisfying line. It is also refuted by the precinct data Taibbi declined to consult.

Mamdani won the low-income neighborhoods outright. In areas with median household income below $45,000, he took 52 percent to Cuomo’s 42, and in twelve such neighborhoods across Brooklyn, the Bronx, and Queens he gained more than thirty points between the primary and the general (THE CITY). He carried 61 percent of majority-Black precincts and 57 percent of majority-Hispanic ones (Bloomberg). He won the Bronx, the poorest urban county in the United States, by eleven points after losing it by eighteen in the primary. The turnout that powered him was the highest in a New York mayoral race since 1993 (City & State).
A movement that wins working-class Black and Latino neighborhoods by double digits, in the outer boroughs, on the strength of bus fares and rent, is not an “elite fixation.” Ruffalo and Spike Lee are not the base; they are the camp followers who turn up later. Taibbi has mistaken the celebrities photographing the parade for the people marching in it. It is a strange error for a man who spent fifteen years insisting that the political establishment was congenitally blind to exactly these voters.

Guilt by the longest possible association
The emotional core of the piece, the charge that these people “hate this place,” that they regard 9/11 as deserved, rests almost entirely on Hasan Piker, a Twitch streamer who holds no office, ran in no race, and was endorsed by none of the winners. The “deserved 9/11” line is from 2019, was disavowed by Piker as imprecise and inappropriate, and was, in its actual context, the wholly conventional foreign-policy argument that decades of covert intervention produce blowback, a word, incidentally, coined not by the DSA but by the CIA (TheWrap; Wikipedia).

Set aside whether the line was defensible. Ask the real structural question: what is a streamer’s six-year-old remark doing in an essay about three congressional primaries? It is carrying the entire argument. Strip out Piker and the deleted CUAD post, and Taibbi has no evidence whatsoever that the winning candidates hold the views he assigns to the movement. He has imported the most extreme sentence he could find from the widest possible perimeter and pinned it to people who never said it. That is the amalgam.

Connolly, the constitution, and the bait-and-switch
The most intellectually dishonest move is the smallest. Taibbi quotes National Political Committee member Cliff Connolly on abolishing the presidency and the Supreme Court, and a DSA convention resolution on founding a “democratic socialist republic,” and presents these as the operative governing program of everyone who won on Tuesday.

But that resolution was the product of the Marxist Unity Group, a specific internal caucus, and its maximalist constitutional language is precisely the kind of programmatic document that organizations of the left have produced and shelved for a century and a half (2025 DSA Convention proposals; Marxist Unity Group). It is not what Avila Chevalier ran on. It is not what Lander ran on. It is not what Mamdani is doing with the buses. Taibbi takes the most distant caucus document in the building and hangs it around the neck of the daycare platform, and then, having manufactured the threat, congratulates the Founders for the checks and balances that will contain it.

The Senate, the presidency, the Court, Taibbi praises them, as Connolly bluntly noted, precisely because they are “levers the ruling class can pull” to stop popular legislation that has already passed the House. Taibbi treats that as a feature. He is, in the end, defending the counter-majoritarian machinery of the American state because it works. That is an honest position. It is not a populist one. The man who built a career on Griftopia and the “vampire squid” now finds himself explaining why it is good that the people cannot get what majorities of them vote for.

The European point Taibbi needs you to forget
The essay’s rhetorical engine is the slippery slope: rent freeze today, “concentrate power absolutely” tomorrow, the gulag by Thursday. “Anyone who’s ever lived in a socialist country” will recognize it.

This is the oldest bad-faith move in the anti-socialist repertoire, but Europe disposes of it without any effort. Vienna has housed nearly half its population in municipal and limited-profit social housing for a century. Gemeindebau, begun under Red Vienna in the 1920s, and Austria remains a parliamentary democracy with private property, contested elections, and an independent judiciary. Tallinn has run fare-free public transit since 2013. The Nordic states deliver near-universal childcare without abolishing markets or the rule of law. Decommodifying a slice of housing, transit, and care is not the first step to totalitarianism; it is ordinary social-democratic governance, practiced across the wealthiest and freest societies on earth. Taibbi needs the reader to believe there is nothing between Dan Goldman and Pol Pot, because the moment you concede the European middle ground, “city-owned grocery store” stops sounding like Kronstadt and starts sounding like a co-op.

And note who actually concentrates power in his account. Taibbi concedes, almost in passing, that “a handful of billionaires own somewhere between 17 and 20 percent of the world’s wealth.” He treats this as the neutral baseline, the natural order against which redistribution is the aggression. But a politics in which two million tenants get a rent freeze and every toddler gets a daycare slot is not the concentration of power. The concentration already happened. He is defending it.

Where Taibbi is right, which is the point
Honesty is the better polemic, so I’ll concede what is true. Avila Chevalier beat Espaillat by less than four points, not the “obliteration” Taibbi opens with, and his own prose can’t decide whether this was a “tidal wave” or a fluke. Some of the online ephemera he quotes is genuinely silly; “the pyromania associated with anarchism is very intriguing to me” is not a serious sentence, and pretending otherwise would insult the reader. The celebrity bandwagon is real and frequently insufferable.

But the existence of a foolish tweet does not refute a housing program, any more than a senator’s foolish tweet refutes the existence of the Senate. The whole apparatus of Taibbi’s essay exists to make you judge a movement that won the poorest borough in America on bus fares and rent by the worst thing said by the most peripheral person within a hundred yards of it. That is the amalgam. Stepan Verkhovensky, whom Taibbi name-checks to flatter himself, would have recognized the technique immediately, not in the young radicals of Demons, but in the prosecutors who came for them.

The economists, for what it’s worth, ran the numbers on the actual program and signed their names to it (The Nation). The nitwits, it turns out, did the math. Taibbi just didn’t read it.

Sources consulted: NPR, NBC News, CNN, Bloomberg, THE CITY, City & State New York, Time Out, The Nation, Democracy Now!, Norwood News, The Intercept, TheWrap, Wikipedia, and the 2025 DSA National Convention proposal.

Racket News

Zohran Mamdani and the Nitwit Revolution

The days of dismissing the Democratic Socialists of America and New York Mayor Zohran Mamdani as lightweights are over, after Mamdani-endorsed congressional candidates Brad Lander, Claire Valdez, and Darializa Chevalier obliterated traditional Democrats…

Read more

2 months ago · 1095 likes · 503 comments · Matt Taibbi

A month ago, in these pages, I argued that there is exactly one question worth asking in American politics, and that our newspapers and our consultants and our chatbots are built to refuse it. The question is who owns what. Who owns the housing, the transit, the childcare, the energy grid, and now the machines. I argued that Zohran Mamdani’s mayoral victory mattered for a single reason: he was the first major American politician in fifty years to build a coalition around that question and win.

That question has an older name, and it is the name of this publication. Enclosure is the process by which things held in common are fenced, titled, and handed to private owners: the village green hedged off and sold, the shared field turned to a rent. The original enclosures drove the English peasantry off the land and manufactured the propertyless working class out of people who had once had a commons to stand on. Everything I write here proceeds from the conviction that the last forty years of American life have been a second enclosure: of housing, of medicine, of transit, of childcare, and lately of the data and the models built from our own collective output. The ownership question is the anti-enclosure question. To ask who owns what is to ask what we are going to take back out of the fences.

The standard objection to Mamdani’s victory arrived on schedule. It doesn’t generalize. New York is not America. Mamdani is a charismatic anomaly, a one-cycle phenomenon, a candidate and not a program. The ownership question, conventional wisdom assured us, was a local curiosity that would not survive contact with the rest of the electorate, or even with the rest of New York.

Tuesday night answered them.

The Extent of It

Mamdani endorsed three candidates for Congress and went three for three. Brad Lander took NY‑10, unseating the incumbent Dan Goldman, an heir to the Levi Strauss fortune and as pure a specimen of donor-class liberalism as the borough produces. That one we expected. The other two are the real story.

Claire Valdez won NY‑7, and she did it over Antonio Reynoso, the Brooklyn borough president, and Julie Won, with the Brooklyn Democratic establishment closing ranks against her and telling Mamdani, in public, to stay out of their borough. He did not stay out. His candidate won anyway. This is the most important political fact of the week: a mayor whose own base is in Queens reached into Brooklyn, was told by the county machine to back off, refused, and beat them on their own ground. Machines exist to make exactly that impossible. This one could not.

And in NY‑13, Darializa Avila Chevalier, a community organizer Mamdani endorsed only in the closing weeks, knocked off the sitting congressman Adriano Espaillat in a district the incumbent had treated as his freehold. Narrow, yes. A win is a win, and incumbents do not lose narrowly to organizers unless something underneath the numbers has shifted.

The congressional races were not the whole of it. Down the ballot, in State Senate District 12: Astoria, Long Island City, Sunnyside, Woodside, Maspeth, the Queens neighborhoods that voted for Mamdani in a landslide, Aber Kawas defeated a two-term assemblyman, Stephen Raga, and stands to become the first Palestinian American in the New York State Legislature. Kawas is the candidate the press spent the spring describing as a radical, an antisemitic insurgent, a danger. But look at the actual platform: free buses, a freeze on stabilized rents, a handful of city-run grocery stores. That is the program. Three city groceries and a bus that doesn’t charge you. The radical menace the media conjured out of this is a fiction, manufactured because the alternative, admitting that a socialist won on a popular, modest, eminently fundable agenda, is unspeakable in the discourse bourgeois liberalism is willing to host.

Add it up. Three for three in Congress, the Brooklyn machine humiliated, an incumbent congressman turned out, the first Palestinian American headed to Albany. Conventional wisdom said the frame would not travel. It traveled to four different districts in a single night.

What It Portends

Here is what it means, stated plainly, because the time for obscurity is over.

The first thing it means is that Mamdani is no longer just a candidate. He is a faction. An endorsement that moves a borough race against the county organization is infrastructure. The Washington Post reports that Mamdani’s people view these House endorsements as a dry run for 2028, and they are right to. What was tested on Tuesday was whether the ownership coalition could be transferred to other candidates, in other boroughs, against the organized resistance of the party. It can. That is a different order of fact, and every county chair in America understands it tonight even if they will not say so.

The second thing it means is that the long misdirection is breaking. For forty years the Democratic Party operated under a settlement engineered by the Democratic Leadership Council and ratified by Clinton, Obama, and the donor stratum that staffs and funds them: be generous on the cultural questions that cost the donors nothing, stingy on the economic questions that cost them everything. Wolfgang Streeck named the European version of this convergence; ours was bourgeois liberalism, the politics of a propertied class settlement that never worked as pragmatism. Its central commandment was that you could not run on the program; you could not say socialism, could not name the donors as donors, could not organize a campaign around the median rent. Tuesday, four candidates ran on the program and won. The commandment is dead. It died in Brooklyn and Queens, in front of witnesses.

The third thing, and this is the important one, is that who owns what now has electoral evidence behind it, not just a single mayoral result. Decommodification is not an academic topic. Free buses are an ownership claim: the public owns the transit and stops renting it back to riders at the fare box. A rent freeze is an ownership claim: shelter is a condition of being alive, not an asset class to be optimized. Five city-run groceries are an ownership claim, modest to the point of timidity, and the establishment treated them as Bolshevism precisely because the principle, not the scale, is what frightens them. Vienna has housed the majority of its population in public and regulated tenure for a century at half the rents. Norway socialized its oil into the largest public capital pool on earth. Mondragón has run seventy thousand worker-owners for sixty years at a productivity premium. None of this is utopian. All of it is operational somewhere, outperforming the privatized American model on the metrics that model claims to care about. The only thing standing between us and it has been the political will of a leadership whose donors spent forty years suppressing the will. On Tuesday the will found four candidates and a borough.

And the enclosure does not stop. The same logic that asks who owns the buses asks who owns the machines, the foundation models trained on our collective output, the data centers drinking water our grandchildren won’t, the algorithmic infrastructures rewriting every wage bargain below the level of public deliberation. This is the newest commons and the fastest enclosure in history: the whole record of human cultural production scraped, fenced, and sold back to us as a service, at a price calibrated to the cut it lets your employer take out of your wages. Decommodify the essentials, socialize the substrate, let ownership follow participation. The housing question and the machines question are the same question in different registers, and a party that can win the first can be made to fight the second. Schumer and Jeffries will perform concern about AI while cashing the rentiers’ checks. The candidates who won Tuesday are built to do otherwise.

The Theme

So let me say it as a demand and not an observation. Who owns what has to become the organizing theme of Democratic politics, not a left agenda appended to a centrist program, but the spine of the thing. Not “who do you trust,” not “who is more electable,” not “who can restore civic decency.” Who owns the housing. Who owns the transit. Who owns the grid, the data, the models, the city. Every other framing the party has tried in forty years was a way of not asking it, and every one of them delivered us, in the end, to the most nakedly oligarchic government in American history.

The press will spend this week explaining that Tuesday was about personality, or coalition arithmetic, or the peculiarities of New York. It was about none of those things. It was about ownership, and the discovery that when you put the ownership question to working people directly, in language they recognize from their own lives, they answer it the same way four times in one night.

Who owns the city is the question. On Tuesday, in four districts, working people began to pull at the fences. The answer started to come back: we do, or no one does. That is the whole of the against enclosure project. The new Democratic Party walks that road, or it is no longer the Democratic Party in any sense that matters

.

In 1977 a German writer named Klaus Theweleit sat down with the diaries, letters, and novels of the Freikorps, the demobilized soldiers who roamed Germany after 1918 breaking workers’ councils and beating the young republic toward the Reich that would eventually swallow it. He was not after their politics, which were plain enough. He was after what they felt, and what he found, across two volumes and a thousand pages, was one recurring dread: the fear of dissolving, of being engulfed, of the boundary giving way. The enemy in their prose was always liquid and always feminine, the Red flood, the tide of Bolshevism, the woman as swamp and contagion. Against it the soldier built himself a body of armor, hard and clean and dry, a man made of will alone, and the armor’s entire purpose was to keep the soft and the formless from getting in. Theweleit called these men the soldier males. He argued that their fascism began in a body that had been trained to feel tenderness as a threat, and only later found the ideas to match the feeling.

A century later Alexander Cortes who sponsors the Substack: The Power of BroScience by AJAC, and who graduated college in 2010 sits down to tell us how it felt to come of age in the Obama years. The vocabulary has been updated. The structure has not. He writes that his cohort of men was “neutered and emasculated,” that the culture ordered them to “LISTEN to woman,” that the air was thick with “white man bad” and apology, and that he lived inside this like a captivity until a strongman appeared and “everything changed. The social order was upended.” He watches blood smeared across a cage on the South Lawn of the White House, a military band playing each fighter to the ring, and he calls it a rebirth. Theweleit would have recognized every aspect: the dread of softness, the armored body restored, the flood pushed back, the great relief that arrives when a man is finally allowed to be hard again.

That is the subject of this essay. Not whether these men hurt, because they do, and the hurt is documented. The subject is what they have been taught to do with the hurt, who taught them, and who profits from the lesson. What they have written about themselves, the one that ends “God wants us to win. So build, and fight for the future you want to live in,” is a confession of real pain attached to a false cause and a paid-for cure. I want to separate the three, because the right has spent fifteen years fusing them.

The Curriculum, in Their Own Words

As an anthropologist, I read the manosphere not for the truth of its claims, which is beside the point, but for what it teaches its members to see. The earliest classroom in our author’s account was Return of Kings, the blog of Daryush Valizadeh, a site that ran instructional posts on “neomasculinity” and, in one notorious entry, an argument for legalizing rape on private property. He calls views of that vintage, “considered far right and horrifying back then,” now “fully mainstreamed,” and he is correct, and he says it with pride. That is the curriculum announcing its now established success.

The lesson has a grammar, and it is built on a single operation: every wound a man feels is to be traced to the feminine and the foreign, never to capital. Listen to the syllabus as the author’s allies recite it. Isaac Simpson, surveying the brand logos on the floor of the Octagon, declares “an entirely alternate male-centric cultural economy that the modern leftist simply has no answer for.” Cortes promises that “within 5 years 70 to 80 percent of young men will be Right Wing (all the hot women will follow them),” that “the left is the party of emasculation, shrieking childless women, and troons,” that “no self respecting man with an ounce of testosterone will align himself with them.” Mark the verbs and the targets. Women who do not submit are “shrieking” and “childless.” Trans people are reduced to a slur and counted as vermin. The reward for joining is sexual, the women arriving as a delivery, “all the hot women will follow.” This is not an argument that can be answered. It is the same reflex Theweleit found in the Freikorps diaries, the same one Wilhelm Reich diagnosed in 1933 when he watched sexual fear and economic ruin braid together in the German lower middle class and harden into a mass movement. Reich’s book was called The Mass Psychology of Fascism. He could have been annotating the timeline.

The Wound Is Real

Here I will do what the right never does for its enemies, which is concede the strongest version of the other side’s case. The men are not imagining the injury. The injury is in the data, and it is severe.

A man who graduated college in 2010 walked into the labor market at the bottom of the deepest downturn since the Depression, carrying debt, into an economy that had spent forty years transferring the floor out from under workers like him. Anne Case and Angus Deaton, the Princeton economists who gave us the phrase “deaths of despair,” tracked what happened to American men without a bachelor’s degree across this period and found a cohort whose median earnings lost thirteen percent of their purchasing power between 1979 and 2017, the same decades in which national income grew eighty-five percent. The plant closed, the union local folded, the wage that once let a man without a degree marry, buy a house, and command a little respect was offshored, automated, casualized, and financialized into a gig. Case and Deaton are blunt about the cause: the weakening of labor, the growing power of corporations, and a health care sector that drains working class wages upward. The men began to die of it, by their own hand, by the bottle, by the needle, in numbers that bent the national life expectancy downward for the first time in a century, and they died disproportionately white, male, and without a degree, which is to say from precisely the cohort now writing essays about emasculation.

So when our author says he “saw fundamental cracks in the well being of regular Americans,” that “deindustrialization was in full swing across Middle America,” he is describing something true and something cruel. Karl Polanyi, writing in 1944 from the wreckage of the last time this happened, described the process. Market society, he argued, treats land, labor, and human life as though they were ordinary commodities produced for sale, which they are not, and the attempt to run a whole civilization on that fiction tears the social fabric so violently that society always recoils. He called the recoil the double movement: the market advances, and the people, wrecked by it, move to protect themselves. The protective movement is not optional and not irrational. It is the human response to being treated as a thing.

But Polanyi’s warning, the one that matters for these men, is that the counter-movement has no fixed politics. The recoil against the market can go left, toward solidarity, the union, the social wage, the common ownership of the things people need. Or it can go right, toward the strongman, the purified nation, the hunt for an internal enemy on whom to pin the wound. Polanyi watched the second road taken in the 1930s and gave it its proper name, which was fascism. The pain is the same either way. The cure is the question. And the entire machinery our author celebrates exists to make sure the cure these men reach for is the poison.

The Misnaming

Ten thousand times a day across the manosphere, this misdirection is reinforced. The wound is real. The cause named for it is a lie. It was not feminism that closed the plant. It was not DEI that broke the union. It was not a trans teenager, a “shrieking childless woman,” or a professor of critical race theory who decided that the pension fund would be loaded with debt and the factory shipped to Guangdong. Those decisions were made in boardrooms and private equity offices and the Business Roundtable, by men, overwhelmingly, and richly rewarded men, who have every reason in the world to be grateful that their victims have been taught to blame a graduate student instead.

Stuart Hall described this operation in 1978, studying how the British press manufactured a panic over street crime to license a harder state. He called the resulting politics authoritarian populism: a popular consent, organized from above, to one’s own policing and one’s own dispossession. The grievance is genuine, the rage is real, and both are carefully redirected downward and sideways, away from the people who own the country and toward the people who have even less of it than you do. The dispossessed man is handed an enemy he can actually defeat, the woman who will not defer, the migrant at the plant gate, the trans kid, because the enemy who actually dispossessed him sits behind security he will never breach and writes the checks that fund the media telling him whom to hate.

There is a reason the chosen enemy is so often a woman, and it far predates this election cycle. Silvia Federici, in Caliban and the Witch, showed that the birth of capitalism was accompanied by a war on women, the witch hunts, the criminalizing of midwifery and contraception, the violent confinement of women to unpaid reproductive labor, all of it concurrent with the enclosure of the common lands. The disciplining of the dispossessed and the disciplining of women were one project. The “tradwife,” the demand that women return to childbearing and obedience, the fury at the “childless” woman who works and refuses, is that same project resurfacing on cue in a moment of crisis, the attempt to fix a man’s lost authority by restoring his authority over a woman, since his authority over capital was never on offer. It is a great deal cheaper to give a man dominion over his wife than over his workplace, and the people who own the workplace have always understood this.

The Commodity of Rebellion

Now to Simpson’s “alternate male-centric cultural economy”. Look at what he is actually celebrating, in his own inventory: “Anduril, Monster, Starlink brands on the floor of the Octagon. Sponsor Polymarket bets pop up as interstitials with sponsor Meta Threads, Zuckerberg shaking hands in the audience.” A defense contractor, an energy drink, a satellite monopoly, a betting platform, a social media conglomerate, and the fifth-richest man on earth working the crowd. This is the counterculture. These are the rebels. A movement that imagines itself a revolt against the system has chosen as its iconography the logos of five of the most heavily capitalized firms in the system, arrayed around a blood sport on the lawn of the state while the standing army’s band plays the fighters in.

Marx, in the first volume of Capital, described the commodity fetish, the way a product of human labor and human relations comes to stand before us as a magical thing with a will of its own, its origins in exploitation invisible, its surface gleaming. Guy Debord updated the diagnosis for the age of images in 1967: in the society of the spectacle, he wrote, lived experience is replaced by its representation, and the spectacle is capital accumulated to the point where it becomes an image. What these men experience as the rebirth of masculine vitality is a representation of vitality, sold to them, and the proof is in the sponsor deck. Thomas Frank documented the mechanism for an earlier generation in The Conquest of Cool: American capital long ago learned that the most profitable thing it can do with rebellion is to manufacture it, brand it, and sell it back to the rebels at a markup. The hippie’s dissent became a Volkswagen ad. The manosphere’s dissent has become a Monster Energy can and a wager on Polymarket. Same machine, harder demographic.

And the centerpiece tells the whole story if you let it. Anduril, the most exalted brand on that Octagon floor, is a weapons company founded by Palmer Luckey, valued in its latest round at sixty billion dollars, led by the venture firm Andreessen Horowitz, which has rebranded the whole project of pointing Silicon Valley capital at the Pentagon as “American Dynamism,” the very movement our author is “proud to be a part of.” Read that arrangement slowly. The same class of investor that spent forty years funding the offshoring, the gig platforms, and the automation that gutted the male wage has now opened a second business line: selling the gutted men a feeling of national and masculine potency, in the form of drones and surveillance towers, at a sixty-billion-dollar valuation. They dispossessed these men as workers. Now they will profit from them again as an audience. David Harvey called the first act accumulation by dispossession. The second act is accumulation by dispossession with a marching band.

The Armored Body and the Strongman

What does the culture get wrong, beyond the misnaming? It gets the nature of strength wrong, and that error is where the danger really lives.

The manosphere’s model of manhood is the armored body Theweleit described: hard, sealed, contemptuous of softness, defining itself entirely against the feminine and the weak. Erich Fromm, in Escape from Freedom in 1941, explained why this model is so seductive precisely to the man cut loose by capitalism. The isolated individual, stripped of the guild, the union, the stable community that once held him, finds his freedom unbearable, a freedom that is mostly the freedom to fail alone, and he flees it into submission, surrendering himself to a strongman or a movement that promises to end the isolation and return him a borrowed sense of power. Read our author again with Fromm in hand. “The war is always for eternity.” “Never give up.” “God wants us to win.” This is not the language of a free man. It is the language of a man who has handed his will to something larger and feels, in the handing over, the relief of no longer being alone. The strength on offer is the strength of submission. They have been sold it as dominance.

And what is sold has a seller worth that our author name with love. His “favorite writer, and now a good friend,” is Jack Donovan, whose essay he praises and whose vision of “masculine values elevated in our society” he celebrates as fulfilled. Let the record speak. Donovan is the author of The Way of Men, whose thesis is that the male essence is violence and the gang, and that civilization’s attempt to suppress male violence has drained men’s lives of meaning. From 2014 to 2018 he led a chapter of the Wolves of Vinland, a folkish neopagan group the Southern Poverty Law Center designates a hate group. He has, at various points, advocated male supremacy, white nationalism, fascism, and the political disenfranchisement of women. There is a further detail, and I offer it because it exposes the incoherence at the core of the whole project: Donovan is a gay man, who built his name on a manifesto, Androphilia, repudiating gay culture as effeminate while celebrating attraction between “masculine” men. The natalist, tradwife, family-values right has made one of its prophets of masculinity a homosexual theorist of male violence, and does not notice, because the doctrine was never actually about families or children. It was always about the armored body and its enemies.

A culture that defines manhood as violence, that reduces women to wombs and trans people to vermin, that teaches dispossessed men to find their dignity in domination and their enemies among the powerless, is not a self-help movement with edgy branding. It is the cultural infrastructure of male-supremacist and white-nationalist politics, and that infrastructure has a body count, in the men radicalized in these forums who have walked into yoga studios and synagogues and massage parlors. The author calls Donovan’s worldview a recovery of “traditional values.” The tradition it actually recovers is the one Theweleit traced through the Freikorps to the Reich: the man who has been hurt by history, taught to blame the soft and the foreign, and armed.

Who Owns the Octagon

Strip the spectacle away and the structure underneath is the one this Substack returns to every week. A commons was enclosed. The thing that was taken from these men was real, the social wage in its broadest sense: the secure job, the union that gave a man without a degree a voice and a floor, the wage that purchased a family and a measure of standing, the institutions that held a community of men together and gave their lives a shape that did not depend on a billionaire’s approval. That commons was fenced and sold for parts over forty years, by the donor class, the private equity firms, the offshoring executives, the same venture capital that now valuates the consolation prize at sixty billion dollars.

And into the vacuum where a man’s security used to be, capital has moved a product: manhood as consumption, dignity as spectacle, rebellion as a logo. The men own none of the machines on that Octagon floor. They will not own Anduril or Starlink or Polymarket or a share of the spectacle staged for them. They are invited to feel ownership, to feel dominion, to feel like the apex predators the timeline promises they are becoming, and the feeling is the product, sold by the people who own the actual machines and intend to keep owning them. Juvenal had the formula two thousand years ago, watching the Roman state pacify a dispossessed citizenry with free grain and free games. Panem et circenses. Bread and circuses. The modern refinement is that the bread has been taken away and only the circus remains, and the circus now carries a sponsor.

The Counter-Claim

We can refuse the lesson, and grant the pain that the lesson exploits. These men were wronged. The wrong was structural, the slow theft of the floor beneath a working life, and the rage it produced is justified. Everything that follows from the rage in the manosphere is a redirection of it, engineered to keep the men punching downward at women and migrants and trans kids so that they never turn and look at the people holding the deed to the country.

Polanyi’s double movement is the whole of it. The recoil against the market is coming, has come, will not be stopped, because human beings will not consent forever to being treated as a commodity. The only question is its direction. The right has captured the recoil and aimed it at the powerless, and it has done so by selling armored men a branded fantasy of strength while the actual sources of their strength, the union, the social wage, the common ownership of the things a life requires, were carted off and melted down. The author asks his readers to “build, and fight for the future you want to live in.” The instinct is sound. The future worth building is the one the manosphere has trained these men to call emasculation: the future in which a man’s dignity rests on what he owns in common with other workers rather than on what he is permitted to dominate at home.

They have told us, in a thousand posts, what they believe manhood is: a thing you buy, a flag you salute, a woman you command, an enemy you crush, a strongman to whom you surrender your will and call it freedom. They have spent fifteen years teaching a generation of wounded men to believe it. The men deserve better than their teachers. The wound is real, the cure is solidarity, and the people who sold them the armor are the same people who took the commons. Build, then. Build the union before the brand. Build the floor before the spectacle. Build the ownership before the merch. The eagles these men are mourning were never coming back. What is overhead now is a sky full of sponsors, and the men cheering it own nothing of what they see.

Bibliography

Scholarship

Debord, Guy. The Society of the Spectacle. Translated by Donald Nicholson-Smith. New York: Zone Books, 1994. First published 1967.

Federici, Silvia. Caliban and the Witch: Women, the Body and Primitive Accumulation. New York: Autonomedia, 2004.

Frank, Thomas. The Conquest of Cool: Business Culture, Counterculture, and the Rise of Hip Consumerism. Chicago: University of Chicago Press, 1997.

Fromm, Erich. Escape from Freedom. New York: Farrar and Rinehart, 1941.

Hall, Stuart, Chas Critcher, Tony Jefferson, John Clarke, and Brian Roberts. Policing the Crisis: Mugging, the State, and Law and Order. London: Macmillan, 1978.

Harvey, David. The New Imperialism. Oxford: Oxford University Press, 2003.

Marx, Karl. Capital: A Critique of Political Economy, Volume One. Translated by Ben Fowkes. London: Penguin, 1976. First published 1867.

Polanyi, Karl. The Great Transformation: The Political and Economic Origins of Our Time. New York: Farrar and Rinehart, 1944.

Reich, Wilhelm. The Mass Psychology of Fascism. Translated by Vincent Carfagno. New York: Farrar, Straus and Giroux, 1970. First published 1933.

Theweleit, Klaus. Male Fantasies. 2 vols. Translated by Stephen Conway, Erica Carter, and Chris Turner. Minneapolis: University of Minnesota Press, 1987 and 1989. First published 1977 and 1978.

Data and Reporting

Case, Anne, and Angus Deaton. Deaths of Despair and the Future of Capitalism. Princeton: Princeton University Press, 2020.

Dilly, et al. European Election Studies 2024, on the youth gender gap in far-right support. (Reporting that young men’s far-right vote share reached roughly 21 percent against roughly 14 percent for young women of the same cohort.)

Public Religion Research Institute. Partisanship, Ideology, and Young Americans. Washington, DC: PRRI, 2024. (On the roughly nine-point Republican identification gap between young men and young women.)

Axios. “Anduril Raising $4 Billion at a $60 Billion Valuation.” March 4, 2026.

Andreessen Horowitz. “American Dynamism: Supporting the National Interest.” a16z, 2023–2025.

Southern Poverty Law Center. “A Chorus of Violence: Jack Donovan and the Organizing Power of Male Supremacy.” Hatewatch.

Wikipedia and SPLC profiles of Jack Donovan, The Way of Men (2012), Androphilia (2007), and the Wolves of Vinland.

In 1753 Carl Linnaeus gave the plant its name. Cannabis sativa, two words in the Species Plantarum, the binomial that botanists still close with a capital L. in his honor. The name was a gift to the commons of knowledge. Anyone could use it, because Linnaeus claimed no ownership of the thing he had described, only the discipline of describing it. For two hundred and seventy years that was the legal status of the plant’s identity. It had a name, and the name belonged to everyone.

On June 9, 2026, the same plant acquired a different kind of name. The German pharmaceutical firm Vertanical received marketing authorization in Germany and Austria for a standardized full-spectrum extract of Cannabis sativa, to be sold under the trademark Exilby. The active ingredient is a proprietary chemovar the company designates DKJ127. In the clinical literature it is VER-01. On the pharmacy shelf it will be Exilby, a manufactured word, the kind of empty trademarkable coinage every branded drug carries, and like every trademark it is property. You will not be able to use that name without permission. The plant Linnaeus named for all of us now answers, in the one market that matters for price and reimbursement, to a word that one company owns.

Eight days. That is the interval between the essay I published on June 1, arguing that pharmaceutical capital had conceded the entourage effect in order to capture the access architecture, and the regulatory approval that turned the argument into a commodity. I wrote that the recuperation was coming and that the patent filings were public. The filing is now an authorization. The forecast is now a launch date.

What happened in twelve days

The facts are in the public record. Exilby is approved for chronic low back pain with a radicular, that is neuropathic, component. The approval rests on two Phase 3 randomized controlled trials in more than 1,200 patients: the placebo-controlled trial of 820 patients published in Nature Medicine in September 2025, and the ELEVATE head-to-head trial of 384 patients against opioids, published in Pain and Therapy, in which the cannabis extract beat opioids on pain relief and on gastrointestinal tolerability with no sign of dependence. Germany authorized the product as a finished medicine. Under the European mutual-recognition procedure, that single national approval becomes the dossier that walks the drug across the rest of the continent. Austria followed at once. The United Kingdom is expected by year end. Commercial launch in Germany and Austria is set for September 2026.

Across the Atlantic the machinery is already turning. On May 18 the United States Food and Drug Administration granted VER-01 Breakthrough Therapy Designation, the status reserved for investigational medicines that promise substantial improvement over existing therapy in a serious condition. It is a fast lane. A pivotal US Phase 3 trial is underway, with a first data read-out expected in 2027 and a New Drug Application planned for 2028. VER-01 is positioned to become the first FDA-approved cannabis-derived medicine for pain, and only the second cannabis-derived drug the agency has ever approved, after Epidiolex in 2018.

The development cost exceeded 250 million dollars over seven years. That figure is the whole argument in a single number. No dispensary, no patient collective, no state-licensed cultivator can spend a quarter of a billion dollars to prove in a flagship journal what the community has known clinically for two decades. The capital required to manufacture the legitimacy is precisely the capital the movement does not have and the incumbents do. Vertanical’s founder, Dr. Clemens Fischer, is a former Novartis board member. Its chief executive came up through a serial pharmaceutical-entrepreneurship vehicle whose previous companies were sold to Novartis, Dermapharm, and Perrigo. The chronic-pain market the company is aiming at is not the rare-epilepsy niche Epidiolex occupies. It is half a billion people worldwide with chronic low back pain, and more than sixty million Americans living with chronic pain, most of them one prescription away from an opioid.

The sentence that fences the plant

Here is the part the cannabis community needs to read twice. Vertanical states, plainly and for the regulatory record, that its findings cannot be extrapolated to other cannabis extracts or to prescribed cannabis flower. The company has built a legal wall around its own evidence.

Think about what that does. For twenty years the movement argued that whole-plant cannabis is therapeutically superior to isolated molecules because of the synergy of cannabinoids, terpenes, and flavonoids, the entourage effect that Ethan Russo named in 2011. The movement wanted that proposition proven at the highest evidentiary level the field can produce. It has now been proven, in Nature Medicine, with the most prominent European skeptic of cannabis-for-pain among the authors. And the proof has been drafted so that it belongs to one product and to nothing else. The trial validates Exilby. It validates the DKJ127 chemovar, grown under contract, fingerprinted by chromatography, manufactured to pharmaceutical standard. It validates the bottle of oral drops with the trademark on the label. It validates none of the flower on a dispensary shelf, none of the extract a patient has self-titrated for a decade, none of the medicine the movement actually built.

The science argument was won. The science was then severed from the plant. The entourage effect is now a private finding, owned, indication-specific, and unavailable as evidence to anyone selling the unbranded version. This is the recuperation completed: the critique adopted, the content kept, the political force removed. The patent holders took the strongest argument the community ever made and turned it into a marketing authorization that explicitly excludes the community.

The pipeline names itself

Vertanical is not alone, and the other names are worse, because they tell you who is buying the plant.

Charlotte’s Web is the cruelest of them. The brand was born from the story of Charlotte Figi, the child whose intractable epilepsy was calmed by a low-THC cannabis extract, the compassionate-use case that put CBD on the American map and gave a generation of parents a reason to fight. That brand now sits inside a joint venture called DeFloria, formed in 2023 with AJNA BioSciences and, as lead investor, British American Tobacco. DeFloria is developing AJA001, an oral full-spectrum botanical drug derived from a proprietary hemp cultivar the company labels CW1AS1, for the symptoms of autism spectrum disorder. The FDA cleared its Phase 2 trial. The compassionate-use story that began with a sick girl in Colorado is now a tobacco-funded botanical-drug program with its own patented cultivar code. The fence has a brand you trust painted on it.

MMJ International Holdings is developing plant-derived THC and CBD softgels: MMJ-001 for multiple-sclerosis spasticity, MMJ-002 for the chorea of Huntington’s disease, both carrying FDA orphan-drug designation. MMJ has spent seven years seeking a DEA license merely to manufacture its active ingredient. And on May 29, MMJ filed a formal legal challenge to the June 29 rescheduling hearing, raising constitutional and treaty objections. A pharmaceutical-pathway company is now litigating the terms of cannabis rescheduling itself, which is what it looks like when an incumbent moves to shape the rules of the market it intends to own.

Above the cultivar, a delivery layer is forming. The pharmaceutical-formulation industry has been building patent estates around oral thin films, sublingual strips, enteric softgels, and metered sprays, the administration technology that turns a standardized extract into a prescribable dose. Merck partnered with a delivery-systems firm for THC and CBD formulations years ago. Once the cultivar is patented and the delivery system is patented, the path from plant to patient is enclosed at both ends. Each of these companies arrives with its own private cultivar designation, DKJ127, CW1AS1, the next one and the one after that, a separate fence around a plant Linnaeus named once, for everyone.

The industry is repricing itself toward the exit

The most revealing development is not in a courtroom or a journal. It is in the trade press the operators themselves read. In the same week Exilby was approved, Business of Cannabis ran one analysis headlined on the sector going “asset-light,” shedding cultivation to chase brand value, and another forecasting a “structural repricing” of the entire industry. Translated out of the financial register: the multi-state operators have read the same situation I am describing and concluded that the money is in the pharmaceutical exit. They want the valuation multiple that comes with an FDA pathway, not the thin margins of selling flower to patients. The owners of the dispensary economy are themselves voting for enclosure, because enclosure is where their equity gets repriced upward. The constituency that built the movement is being sold by the people who currently hold its assets.

The federal vise, on the same clock

None of this happens in a policy vacuum. It happens inside the policy framework the federal government published in the 2026 National Drug Control Strategy and enacted through the April 23 rescheduling order, and the timing is not a coincidence.

The April order placed FDA-approved cannabis products and state-licensed medical marijuana into Schedule III. That is the comfortable shelf where Exilby will sit: prescribed by any physician, dispensed by any pharmacy, eligible for hospital formularies and insurance reimbursement, free of the compliance burden that state-licensed operators carry. The DEA’s final order also contemplates import and export for registered producers, which means a pharmaceutical cannabis product can cross national borders while the dispensary down the street still cannot ship across a state line. The drug travels. The plant stays fenced. The hemp-derived cannabinoid economy, the delta-8 and delta-10 ecosystem that became the de facto national supply in prohibition states, is scheduled for restriction in November 2026, four months after the hearing. And the Strategy itself directs clinicians and consumers to the prescribing labels of Marinol and Epidiolex as the federal government’s authoritative reference on the risks of the cannabis plant, which is to say the patented isolates are now the textbook for understanding the thing they were extracted from.

The June 29 hearing in Arlington, on whether the broader cannabis economy joins medical marijuana in Schedule III, is days away as I write, and it is already drawing court challenges and stay bids, including MMJ’s. Whatever the procedural outcome, the structural result is visible: the approved pharmaceutical product wins on every layer at once. The science is won, by the patent holder. The regulatory shelf is built, for the patent holder. The export lane is opening, for the patent holder. The criminalization apparatus is restored, against everyone else.

The clock started on June 9

In the June 1 essay I estimated an eighteen-month window to build an open-source whole-plant standard as prior art against the patent estates. That estimate was too generous, because it counted from the filing. The clock that matters started on June 9, with the approval. Authorization, not application, is what entrenches a product in the regulatory record and starts the mutual-recognition dossier moving across Europe. Once Exilby is the prescribed, reimbursed, physician-endorsed version of cannabis on two continents, the unbranded plant becomes, in the language of regulatory legitimacy, the inferior and informal one. That is the Epidiolex story exactly. The plant that grows in the ground costs a hundred dollars a month at a dispensary and thirty thousand dollars a year once it carries a trademark. Nothing about the molecules changes. Ownership changes.

There is one move that still beats this, and it is not rhetorical. It is the Public Domain Whole-Plant Cannabis Medicines Program I detailed in February and costed in June: an open, full-spectrum extract standard with published specifications, fingerprinting, and stability data; FDA-grade trials sponsored through a nonprofit and published in full; patents taken only on the delivery interfaces, held in trust by a public-benefit entity and licensed to anyone who meets the quality bar. The model exists. The Usona Institute did exactly this with psilocybin, running Phase 3 trials while keeping the molecule in the public domain through open publication, which is the reason a synthetic-psilocybin patent estate is now contestable on prior art. Cannabis needs the Usona equivalent, and it needs the published clinical data on the record before the Vertanical-tier dossiers finish their march across Europe and the 2028 American NDA lands. The capital required is meaningful and it is not pharmaceutical-scale. The legal pathway, the FDA botanical-drug framework and the Drug Master File that multiple manufacturers can reference, already exists. What is missing is the decision to build it, and the recognition of how little time is left to decide.

The movement spent fifty years and a generation of arrests to make this plant legitimate. Legitimacy arrived, and it arrived wearing a name the movement cannot use, attached to evidence the movement cannot cite, on a shelf the movement cannot reach. Linnaeus described the plant and gave the description away. Vertanical described one cultivar of it and kept the description, the name, and the market. The difference between Cannabis sativa L. and Exilby is the difference between a commons and an enclosure, and it is now written into the pharmacopoeia of two countries with a third to follow.

You won the science. They named it, bottled it, and walked off with it. The ownership is the only fight left, and the clock started on June 9

.

Disclosure

Del Potter is a working pharmaceutical scientist and the Founder and Chief Scientific Officer of Prism Sciences LLC and BioUnbound Inc. His commercial work intersects directly with subjects this piece addresses, including peptide patent and formulation intellectual property, oral thin-film and enteric delivery platforms, and the political economy of pharmaceutical access. He does not write about the specific products, formulations, partners, or investor relationships of his ventures in this publication. The Public Domain Whole-Plant Cannabis Medicines Program he has proposed is the explicit political-economic alternative to the enclosure architecture he is criticizing here. Readers should know the intersection exists and weigh the writing accordingly.

Sources and citations

Bradford, Ashley C., and W. David Bradford. “Medical Marijuana Laws Reduce Prescription Medication Use In Medicare Part D.” Health Affairs 35, no. 7 (July 2016): 1230–36.

Business of Cannabis (Ben Stevens). “Vertanical Secures Milestone FDA Breakthrough Designation for Cannabis Pain Drug Ahead of Imminent European Launch.” May 29, 2026. https://businessofcannabis.com/vertanical-secures-milestone-fda-breakthrough-designation-for-cannabis-pain-drug-ahead-of-imminent-european-launch.

Business of Cannabis. “From Land to Brand: Why Cannabis’s Next Generation Is Going Asset-Light.” June 10, 2026.

Business of Cannabis (Sean McLean). “Cannabis at an Inflexion Point: A Structural Repricing Ahead.” June 8, 2026.

Cannabis Science and Technology. “Big Pharma Watching, Waiting and Building Cannabis Patent Portfolio.” (On Merck/Intec and other delivery-system and cannabinoid patent activity.)

DeFloria, Inc. “DeFloria Receives Investigational New Drug Application Clearance from FDA for Phase 2 Clinical Trial of AJA001, an Oral Cannabinoid Drug Being Developed for the Treatment of Autism Spectrum Disorder (ASD).” PR Newswire, 2025. https://www.prnewswire.com/news-releases/defloria-receives-investigational-new-drug-application-clearance-from-fda-for-phase-2-clinical-trial-of-aja001-an-oral-cannabinoid-drug-being-developed-for-the-treatment-of-autism-spectrum-disorder-asd-302382573.html.

Karst, Matthias, Winfried Meissner, Sabine Sator, Jens Keßler, Volker Schoder, and Winfried Häuser. “Full-spectrum extract from Cannabis sativa DKJ127 for chronic low back pain: a phase 3 randomized placebo-controlled trial.” Nature Medicine 31, no. 12 (December 2025): 4189–4196.

Linnaeus, Carl. Species Plantarum. Stockholm: Laurentius Salvius, 1753.

Meissner, Winfried, Charles Argoff, Sabine Sator, Volker Schoder, et al. “VER-01 Shows Enhanced Gastrointestinal Tolerability, Superior Pain Relief, and Improved Sleep Quality Compared to Opioids in Treating Chronic Low Back Pain: A Randomized Phase 3 Clinical Trial.” Pain and Therapy 14, no. 6 (December 2025): 1765–1782.

MMJ International Holdings. “MMJ International Holdings Files Formal Challenge To DEA Marijuana Rescheduling As Constitutional, Treaty, And Pharmaceutical Questions Collide.” Access Newswire, May 29, 2026.

NeurologyLive. “Cannabis-Derived MS, Huntington Treatments Submitted to FDA for Fast Track, Orphan Drug Status.”

Potter, Del. “How the Cannabis Movement Just Lost the Argument It Won.” Against Enclosure, Substack, June 1, 2026. https://delpotterphd.substack.com/p/how-the-cannabis-movement-just-lost.

Potter, Del. “Proposal for a Public Domain Whole-Plant Cannabis Medicines Program.” Against Enclosure, Substack, February 1, 2026. https://delpotterphd.substack.com/p/proposal-for-a-public-domain-whole.

Russo, Ethan B. “Taming THC: potential cannabis synergy and phytocannabinoid-terpenoid entourage effects.” British Journal of Pharmacology 163, no. 7 (August 2011): 1344–1364.

United States Department of Justice. Order placing FDA-approved marijuana products and state-licensed medical marijuana in Schedule III. April 23, 2026.

United States Drug Enforcement Administration. “Schedules of Controlled Substances: Rescheduling of Marijuana.” Federal Register, April 28, 2026; hearing scheduled to begin June 29, 2026.

Vertanical GmbH. “VERTANICAL Receives First European Marketing Authorization for Exilby as a First-in-Class Non-Opioid Treatment for Chronic Low Back Pain.” PR Newswire, June 9, 2026. https://www.prnewswire.com/news-releases/vertanical-receives-first-european-marketing-authorization-for-exilby-as-a-first-in-class-non-opioid-treatment-for-chronic-low-back-pain–pivotal-us-phase-3-trial-underway-302794524.html.

White House, Office of National Drug Control Policy. National Drug Control Strategy 2026. Washington, DC: Executive Office of the President, May 4, 2026.

Among the Wari’ of the western Amazon, the dead were once eaten by their kin. The anthropologist Beth Conklin lived among them for years and came to understand the practice from the inside: to consume the body was a way of refusing to abandon it, of keeping the beloved inside the living rather than surrendering him to the ground and the slow indifference of the forest. I have been thinking about the Wari’ because my dachshund is seventeen, and seventeen in a long-backed dog is a great age, and he has begun, in the unhurried and generous way of old dogs, the business of leaving.

The hindquarters go first. In the morning he stands at the back door looking at something I cannot see, and waits for me to lift him down the single step he used to take at a run. He has stopped barking at the wind. When I say his name, he still cocks his head, the small punctilio of attention he has kept up for seventeen years, the long discipline of being known by one person and of knowing him back.

I found his mother staked in a patch of grass outside a cannabis supply store, a miniature the color of a roan horse, frantic for any hand that would touch her. My wife had lost her own dachshund a few months before, twenty years of one animal gone in a week, and I read the look that passes between people at the start of these things, and understood she wanted the puppy. I went inside and bargained with the owner, raising my offer until the man gave way, and Coco Ono came home with us. She was lonely, so we found her a mate, a sweet white male we called Yanni Lennon, and the two of them produced a litter, and out of that litter, after we had given the others to family, came Pauley. We kept him a while and then, because one of our kids wanted him and a house full of children seemed the better life, we gave him away. Giving him away was the first of the two times I nearly lost him. I did it with my own hands.

The child could not keep him and passed him to a friend without telling us. The friend kept him in the high desert, tied alone through the hammering afternoons, and did not always feed him. The ravens worked him over. They came down for whatever was in his bowl and for whatever was not, and there was nothing he could do, staked in the heat, but bark at them until his voice gave out. When we got him back, he was gaunt, emaciated really, and his eyes had gone wrong, quick and sideways. For the rest of his life the sight of a raven, the sound of one, would throw him into a fury aimed past the bird in front of him. He was barking at something that had already happened.

I brought him back the only way I knew. I lay on the couch for hours with him on my chest, his heart going much faster than mine, and I talked to him. I told him he was safe. I told him no one would tie him out again, that he would never go hungry, that the part of his life that had taught him to flinch was over and was not coming back. I told him he was going to live his best life, and I meant it, and I kept it. I have no way to prove he understood the words. I am certain he understood the intent. Something in him let go by degrees over those weeks, the way a fist opens, and what was underneath was the most complete trust I have ever been given by anything that breathes. He was mine and I was his. We settled it on the couch, in the long afternoons, with my hand on his ribs.

The anthropologists tell us that attendance on the dying is among the oldest signatures of our kind. Sixty thousand years ago in Shanidar Cave, in the mountains of what is now Iraq, someone laid the dead on a bed of yarrow and cornflower, and the pollen is still there in the soil, a bouquet older than language. The Toraja of Sulawesi keep their dead in the house for months, washing and dressing them, bringing them food, speaking to them as though the conversation had merely been interrupted. The Wari’ ate theirs. Beneath everything else we have made of ourselves, this is the bedrock: we are the animal that does not walk away from the body of the one it loved.

The dog has been practicing this his whole life. He has never read Mauss and has never needed to. When I was sick he stayed in the room. When I wept he leaned the length of his ribs against my leg and held there. He is teaching me the last section of his curriculum now, which is how to stay with a body that is coming apart and offer it nothing more useful than presence.

He came by this honestly, because attention was his vocation. On our land on the Mendocino coast, he appointed himself supervisor of every project that crossed the property. While the men dug trenches for water and power he made his rounds, station to station, studying each one’s work, watching with what I can only call approval before moving on to inspect the next. I recognized the circuit. It was mine. He had learned to be a boss by watching me be one, and he did it better, because he never once interrupted a man to take the credit.

And once he found his brother. Oscar, from the same litter, had gone bad in the back the way dachshunds do and had come home to us in a little cart, and one afternoon he slipped off his blanket by the shed and was gone. The forest there is dense, oak and redwood and manzanita laid over with brush, and it fills at dusk with things that eat small animals. We searched and could not find him and the light was going. In the end we did the only thing left, which was to kneel down and tell Pauley, slowly and seriously and more than once, that we needed him to find Oscar. He shifted into a register I had not seen. He went to the blanket, took the scent, and walked straight into the trackless brush with the whole frightened family stringing along behind him. He cast about, doubled back, ran down a false lead and gave it up. Then he left the trail entirely for an oak we could barely see through the undergrowth, and he stopped, and he lifted one paw and pointed. Oscar was under it, on a bed of leaves, his brown coat poured into the mottled shadow so completely that none of us would have found him in a week. The dog found his brother because no one had ever told him he was only a dog.

Here is what I have come to think, in the long evenings with my hand on his side. The hour I am living now, the one that sits in the room while the breathing changes, is one of the last things in my life that capitalism has not yet found a way to sell back to me.

It has sold me almost everything else. Silvia Federici taught a generation of us to see the so-called primitive accumulation, the enclosure of the common fields that turned peasants into wage laborers, as a permanent operation rather than an event finished off in the English past. It has never stopped needing fresh commons to swallow. The village was a commons. The kin network was a commons. The neighbor who knew the dog by name and would carry a casserole across the road and sit on your porch while you cried, she was a commons too, unpaid and unpriced and therefore invisible to the only ledger the system keeps. One by one these have been fenced, parceled, and made to vanish, and the work they used to do for nothing has been handed back to us at retail. The casserole is a meal-kit subscription now. The neighbor on the porch is a doorbell camera. The long unstructured evening, the one in which people once stayed in a room together, has been broken up and sold off in monthly increments of streaming and scrolling. Atomization is the genteel word for what was done to us. The harder word is enclosure.

What is left, when the fences have gone up around everything else, is the dog. He is the last commons in the house. He was never enclosed because there was no money in him to find. All he ever carried was love, and the market has not yet learned how to meter it. So it is the dog who is here at the end, because everything else that used to attend the dying has been carted off and sold for parts.

This is not for lack of trying. A whole industry has arranged itself around the deaths of animals, and it grows every year. Pet-loss grief coaches and end-of-life concierges, in-home euthanasia with an ambient playlist, cremation packages tiered by the weight of the body, the clay impression of the paw, the locket for the ash, a rainbow bridge with a gift shop at either end. The grief is real and the people selling to it are not all cynics. But the machine does what the machine always does, which is to find the one unpriced thing in the room and put a price on it. It can sell me the keepsake. It can sell me the ceremony and the urn. The one thing it cannot reach, cannot price, cannot optimize or upsell or resolve into a transaction, is the hour itself: the actual minutes on the actual couch with the actual animal, his ribs rising and falling under my hand. That hour belongs to no one but the two of us. It is older than money. It is older than the wage. It is older than the bouquet in the cave at Shanidar.

Years ago, on the Gualala River, my grandson and I pushed a canoe off the bank and started downstream. Pauley stayed behind with the others on the shore. He was no swimmer and had never once volunteered for cold water in his life. But he paced the bank as we drifted, and the pacing turned frantic, and somewhere around a quarter mile out he made up his mind. He went in. He swam after the canoe with everything his ridiculous short legs could produce, and the current was against him and the canoe would not slow, and still he came on, gaining, refusing, until I reached down and hauled him dripping over the gunwale. He stood in the bottom of the boat and shook and looked up at me, exhausted and triumphant, having made the one point he had come to make, which was that he would not be left behind.

I think about that swim now, in the evenings. He is asleep beside me as I write this, one ear folded under, his nose tucked against his paw the way he has tucked it since he was a puppy who weighed less than a loaf of bread. The arithmetic has reversed. He is the one drifting now, and I am the one on the bank, and the only thing I have left to give him is the refusal he taught me on the river. I will not be the one who leaves. The anthropologists say what makes us human is that we attend. The polemicists say what makes us free is that we keep some rooms the market cannot enter. The dog says nothing at all. He breathes, slowly, and I sit and listen. This is the work. It is the oldest work there is.

Part I — Capital Framework

Executive summary

The capital requirement to take an open-source, full-spectrum cannabis extract standard (WPCE-01) from organizational formation through FDA approval for a single primary indication, paired with three open delivery platforms (orally disintegrating tablet, oral thin film, enteric capsule) and made available to qualified manufacturers under non-exclusive licensing, is approximately $85 million to $155 million over six years. The range reflects scope choices, trial size, and the degree to which the three delivery platforms are developed in parallel versus sequentially.

This total is substantially below the industry-average cost of commercial FDA-pathway drug development (roughly $1 billion per approved drug, fully loaded) and is in the same order of magnitude as the cumulative funding the Usona Institute has deployed for its open-source psilocybin program (~$100M+ over the past decade). The lower cost relative to commercial development reflects: a single primary indication initial scope; nonprofit overhead structures; open-science publication that does not require proprietary protection programs; published prior art that reduces IP-prosecution overhead; and a multi-manufacturer post-approval ecosystem that is governed centrally rather than built as a single-firm commercial operation.

The program is conventionally fundable: staged, milestone-gated, with each tranche producing public artifacts that expand the commons and reduce later capital risk. The phasing below is designed so that Phase 1 ($4–7.5M) delivers irreversible public infrastructure, a published WPCE-01 reference standard with analytical methods, a foundation IP estate held in public-benefit trust, a complete Investigator’s Brochure v1.0, and a credible pre-IND package, even if subsequent phases never get funded.

Operating assumptions

These shape the budget. They are documented here so funders, partners, and reviewers can interrogate the assumption set rather than the totals.

• Single primary indication initial focus. Chronic low back pain, modeled on the Vertanical VER-01 trial program. The substitution-for-opioids therapeutic frame is well-established (Bachhuber 2014, Powell-Pacula-Jacobson 2018, Bradford & Bradford 2016) and now confirmed at Phase 3 RCT level (Karst et al. Nature Medicine 2025; Meissner et al. Pain and Therapy 2025). The chronic low back pain population is large, the endpoint is robust, and the regulatory pathway is established.

• Three delivery platforms developed in parallel. ODT, oral thin film, and enteric capsule. Parallel development saves time but adds incremental cost (roughly 60–80 percent of three sequential programs rather than 200–300 percent). The platforms differ in PK profile, patient population fit, and manufacturing complexity; carrying all three avoids picking a winner prematurely and supports a multi-manufacturer post-approval ecosystem.

• GMP-aligned manufacturing from Phase 1 forward. No bench-only chemistry. Cannabis-derived botanical products require chromatographic batch consistency and a controlled supply chain from cultivation forward; cutting CMC corners is what kills botanical programs at Phase 3.

• Nonprofit sponsoring entity, 501(c)(3), with a public-benefit governance charter. Board composition includes patients, clinicians, scientists, regulatory experts, manufacturing partners, and public-interest legal counsel. Transparency requirements written into the charter, not as policy.

• Defensive publication as the primary IP strategy. WPCE-01 specifications, analytical methods, and formulation enabling approaches published early to function as prior art. Patents filed only on delivery interfaces where prior-art coverage would not adequately prevent commercial enclosure, held in trust and licensed non-exclusively.

• Open publication of trial protocols, statistical analysis plans, and de-identified data. ClinicalTrials.gov plus institutional repository; no contract-research-organization data exclusivity.

• Post-approval multi-manufacturer ecosystem. Drug Master File held by the sponsoring nonprofit and referenced by qualified manufacturers via written authorization (21 CFR 314.420). No single-manufacturer dependency.

• Realistic CMO and clinical-trial pricing for the U.S. and European pharmaceutical-services market as of mid-2026. Numbers reflect actual quotes received in adjacent peptide-formulation and clinical-trial programs, scaled for botanical complexity.

Phase 1 — Foundation and Standard (Months 1–12) — $4.0M to $7.5M

The phase that produces the durable public infrastructure regardless of what happens later. Each line item below produces a publishable artifact.

• Nonprofit formation, 501(c)(3) determination, governance charter, public-benefit entity registration, code of conduct, IP-trust structure: $250K–500K

• Founding scientific and clinical advisory boards, initial executive staff (Executive Director, Chief Scientific Officer, Director of CMC, Director of Regulatory Affairs), shared services: pro-rated $750K–1.5M for Phase 1

• Legal counsel (regulatory, IP, governance): $250K–500K

• WPCE-01 specification development: botanical raw material qualification, chemovar selection and characterization, extraction process development, analytical fingerprinting, reference standard creation, initial stability protocols: $500K–1M

• Initial CMC development across the three delivery platforms (formulation chemistry, excipient selection, prototype batch production): $1.5M–2.5M

• Manufacturing partner identification, qualification, quality-agreement negotiation: $500K–1M

• Investigator’s Brochure v1.0 drafting under ICH-GCP standards: $250K–500K

Phase 1 deliverables: the WPCE-01 monograph, the analytical method dossier, the IB v1.0, the IP-trust documents, the charter, are all publishable artifacts that establish prior art and reduce enclosure risk even if subsequent phases never fund.

Phase 2 — IND-Enabling (Months 13–30) — $9M to $18M

• Pre-IND meeting preparation and FDA engagement: $250K–500K

• Full formulation development across three dosage forms: $2M–4M

• Stability program (24-month real-time plus accelerated for each dosage form): $750K–1.5M

• Targeted nonclinical work where literature support is insufficient: $500K–1.5M

• Phase 1 first-in-human PK and safety study (single dosage form lead, with bridging cohorts; n ~60–80): $3M–5M

• GMP-grade manufacturing scale-up for clinical supply: $2M–4M

• Drug Master File establishment with FDA: $500K–1M

• IND submission: $250K–500K

• Continuing operations and overhead: $1.5M–2.5M annualized

Phase 3 — Proof of Concept (Months 31–48) — $17M to $29M

• Phase 2 clinical trial in chronic low back pain (dose-finding and proof of concept, n ~250–400 patients, 12–24 week treatment plus open-label extension): $15M–25M

• Clinical supply manufacturing across dosage forms: $1M–2M

• Statistical and data-management infrastructure (publishing protocol, SAP, and de-identified data to open repository): $500K–1M

• Regulatory affairs: $500K–1M

• Continuing operations: $2M–3M annualized

Phase 4 — Pivotal Program and NDA (Months 49–72) — $55M to $100M

• Phase 3 randomized controlled trial (multicenter, placebo-controlled, ~600–820 patients matching the Vertanical design): $40M–75M

• Long-term safety extension study (52-week open-label): $10M–15M

• GMP-grade manufacturing for Phase 3 and registration batches: $3M–5M

• NDA preparation, submission, and FDA review interactions: $2M–5M

• Continuing operations: $3M–5M annualized

Phase 5 — Approval, Launch, and Multi-Manufacturer Ecosystem (Months 73+) — $7M to $15M per year, ongoing

• NDA review and approval support: $5M–10M (one-time)

• Multi-manufacturer qualification and reference-authorization ecosystem: $2M–5M per year

• Post-market surveillance, pharmacovigilance reporting: $2M–5M per year

• Ongoing program operations and governance: $3M–5M per year

Cumulative totals — all-cash baseline

Subsequent indications (osteoarthritis, painful diabetic peripheral neuropathy, anxiety, sleep, others identified by the substitution literature) can be added under the same regulatory umbrella at marginal cost, roughly $25M–$50M per additional Phase 3 indication, given that the CMC, manufacturing, analytical, and DMF infrastructure are already in place.

Comparative context

Commercial pharmaceutical-grade cannabis development (Vertanical comparable)

Vertanical’s own public materials describe “more than eight years of dedicated research and development” and “tens of millions of Euros annually” in R&D investment. The Vertanical Phase 3 program (VER-01 in chronic low back pain, n=820, with an opioid-comparison arm in n=384) is in the same scale as Phase 4 of this proposal. Public estimates put Vertanical’s cumulative VER-01 investment in the €80–150M range, i.e., directly comparable to the public-domain program total here. The cost is roughly equivalent; the access architecture is the variable.

Full industry-average commercial drug development

Tufts Center for the Study of Drug Development estimates of fully loaded commercial drug development costs (~$1B per approved drug, including failed candidates). The public-domain program achieves NDA approval at roughly 10–15 percent of the industry-loaded average, because it carries a single program (not a portfolio of failed candidates), uses literature-supported nonclinical work, and operates under nonprofit overhead.

Usona Institute psilocybin program (nonprofit precedent)

Usona has invested an estimated $100M+ cumulatively over the past decade in the public-domain development of psilocybin for major depressive disorder. The program produced a published Phase 2 trial, an open Investigator’s Brochure that has become a de facto field reference standard, and is currently running a Phase 3 program for MDD. The Usona model is the existence proof that nonprofit-sponsored FDA-pathway development at this scale is achievable.

Cost-efficiency notes

Three sources of cost advantage that reduce the public-domain program total relative to a single-firm commercial path:

• Open publication reduces IP-prosecution overhead. Commercial firms typically spend 5–10 percent of total program cost on patent prosecution, prior-art surveillance, and infringement defense. The public-domain program spends a fraction of that on defensive publication and a narrow patent estate on delivery interfaces.

• Literature-supported nonclinical work. Commercial firms run extensive proprietary tox programs because their NDA package cannot reference public data. The public-domain program references published literature on cannabinoid pharmacology and toxicology where the body of evidence is sufficient, conducting targeted nonclinical work only where gaps exist.

• No commercial brand-building. Commercial drug development typically allocates 15–25 percent of program cost to brand strategy, market-access positioning, KOL development, and pre-launch marketing. The public-domain program does none of this, the deliverable is regulatory legitimacy, not a brand.

Part II — In-Kind Contribution Analysis

Realistic estimate of how much volunteer, pro bono, and preferred-partner contribution can reduce the cash requirement.

In-kind contributions can realistically reduce the cash requirement of the program by 15 to 22 percent across the full six-year arc to NDA submission, roughly $13 million to $34 million of the $85–155 million baseline budget. This is meaningful but bounded. The program at this scale cannot be built primarily on volunteer labor; GMP manufacturing, clinical-trial site execution, stability programs, and regulatory submissions are real cost centers that require cash. The realistic cash requirement, after maximizing in-kind, is approximately $70 million to $130 million over six years.

That is the honest answer. The expansionary fantasy that says “the cannabis community will volunteer this” is a fundraising failure mode. The conservative reality that says “no one will give us anything” undersells what nonprofit-sponsored health-technology programs have actually achieved. The categories below identify what is gettable, what it is worth, who has done it before, and what the limits are.

Category 1 — Legal counsel (IP, regulatory, governance)

Realistic in-kind value: $2.0 to $4.5 million across the program lifecycle.

This is the single highest-leverage in-kind category. Major law firms with pharmaceutical biotech IP and FDA regulatory practice groups operate substantial pro bono programs and routinely take on public-interest health-technology cases. The Usona Institute’s regulatory and IP strategy has been built on substantial pro bono support from Cooley LLP and other tier-1 firms.

What is realistically obtainable

• Public-benefit entity formation, 501(c)(3) determination, governance charter drafting, IP-trust structure design — $200,000 to $400,000 of equivalent billable work in Phase 1.

• IP strategy and prior-art surveillance program, defensive publication strategy, narrow patent estate on delivery interfaces — $400,000 to $800,000 over Phases 1 and 2.

• FDA regulatory counsel (pre-IND through NDA): meeting preparation, submission review, sponsor-investigator agreements, DMF reference-authorization framework — $500,000 to $1,200,000 across Phases 2–4.

• Ongoing compliance, board governance, manufacturing contract review, post-approval licensing — $300,000 to $700,000 across Phases 3–5.

Specific firm targets to approach

Cooley LLP (significant cannabis IP and biotech experience, Usona Institute precedent). Foley Hoag (cannabis regulatory work). Arnold & Porter (FDA regulatory, deep pro bono history). WilmerHale (biotech IP). Latham & Watkins. Goodwin Procter. Hogan Lovells (botanical drug regulatory experience). Reach the partners through bar-association pro bono coordinators and through direct introduction from foundation funders, who routinely facilitate these connections.

What is not realistically obtainable pro bono

• The volume work of NDA submission preparation. Pro bono partners lead but won’t carry the full submission lift.

• IP litigation if the patent estate is challenged. Pro bono ends at offense; defense requires paid counsel.

Category 2 — Scientific and clinical advisory boards

Realistic in-kind value: $1.5 to $3.0 million across the program lifecycle.

This is the category where the cannabis-medicine field has the largest pool of plausible volunteers. Senior cannabinoid researchers, pain-medicine clinicians, regulatory-experienced physicians, and former FDA reviewers will serve on advisory boards for $0 cash if the program’s scientific seriousness is established. Board service worth $50,000 to $150,000 per senior member per year in consulting-fee equivalent is normal at this scale.

What is realistically obtainable

• A 6–10-member Scientific Advisory Board contributing 4–8 hours per month each — $400,000 to $800,000 per year of equivalent consulting.

• A 4–6-member Clinical Advisory Board of pain and integrative-medicine clinicians — $200,000 to $500,000 per year equivalent.

• Independent Data and Safety Monitoring Board members for Phases 2 and 3 at honorarium-only rates — saving $200,000 to $400,000 across the clinical program.

Specific advisor targets

For cannabinoid science: Ethan Russo, Mahmoud ElSohly (University of Mississippi NIDA cannabis program), Mark Wallace (UCSD), Igor Grant (UCSD). For political-economy and patient-advocacy: Sue Sisley, Donald Abrams (UCSF, retired). For regulatory experience: former FDA CDER reviewers in retirement. For pain-medicine clinical advisory: Mark Wallace, Lynn Webster, Mary Lynch (Dalhousie). The cannabinoid-medicine clinical research community is small and reachable, and senior figures have a documented willingness to serve on advisory boards for mission-aligned programs.

Category 3 — Academic medical center partnerships for clinical trials

Realistic in-kind value: $6 to $18 million across Phases 2 through 4.

This is the category with the largest potential dollar savings and the most variable execution risk. Academic medical centers will conduct nonprofit-sponsored trials at substantially below commercial-CRO pricing — typically 30 to 50 percent below standard rates — when the program is mission-aligned and the institution sees the participation as part of its research mission. This is how Usona has structured its psilocybin Phase 2 and Phase 3 work.

What is realistically obtainable

• “Preferred academic partner” pricing on Phase 1 PK studies — 30–40 percent off commercial CRO pricing, saving approximately $1.0 to $2.0 million.

• Academic-network Phase 2 proof-of-concept trial (3–5 academic sites), saving $3 to $6 million off the commercial baseline.

• Academic-network Phase 3 pivotal trial (8–15 sites), saving $5 to $15 million off commercial baseline.

• IRB fee reductions, free institutional regulatory consultation, shared statistical/data-management infrastructure — saving $500,000 to $1,500,000 across the clinical program.

Specific institutional targets

UCSF (Center for Medicinal Cannabis Research at UCSD), UCSD, UC Berkeley public-health partnerships. Johns Hopkins (psychedelic and pain research infrastructure). University of Pittsburgh (significant pain research). Dartmouth-Hitchcock. University of Vermont (cannabinoid research program). The University of Miami’s Halland Chen (already in the BioUnbound network) is an existing relationship that could anchor the Phase 1 PK study. Hopkins, UCSF, and UCSD are the strongest U.S. candidates.

Critical caveats

Academic partnerships save 30–50 percent off CRO rates but they do not save 100 percent. A 600-patient Phase 3 trial through academic partners still costs $25 to $50 million. Do not let funder conversations treat academic partnerships as “free trials.” And academic medical centers have variable timelines and bureaucracies. The cost of academic partnerships is often delivered as schedule risk rather than as cash.

Category 4 — Board, executive, and audit services

Realistic in-kind value: $0.4 to $1.0 million across the program lifecycle.

A public-benefit nonprofit board operates without paid director compensation. This saves $200,000 to $500,000 across the lifecycle relative to a commercial entity. The CEO and CSO roles will be salaried, but the founding executive team can be expected to take below-market salaries for the first 24–36 months (50–70 percent of commercial equivalent), saving approximately $200,000 to $400,000 over that window. Pro bono accounting and audit services from a tier-1 firm with a public-benefit practice group are obtainable. Saving: $50,000 to $150,000 across the lifecycle.

Category 5 — Communications, publication, and infrastructure

Realistic in-kind value: $0.3 to $0.8 million across the program lifecycle.

Open-access publication fee waivers can be negotiated with publishers (Springer, BMC, Frontiers, PLOS) for nonprofit public-domain work. Pro bono communications and PR support from values-aligned firms is obtainable in the initial-launch phase, saving $100,000 to $300,000. Pro bono technology infrastructure — website, document repository, registry of qualified manufacturers — through volunteer or low-bono technical contribution is realistic for early phases. ClinicalTrials.gov registration, DOI registration through CrossRef or Datacite, institutional repository access through partner universities — all available at low or no cost.

Category 6 — Donated clinical comparators and reference materials

Realistic in-kind value: $0.5 to $2.5 million across Phases 3 and 4.

If the Phase 3 trial includes an opioid comparison arm (as the Vertanical ELEVATE trial did), the comparator opioids must be obtained. Some manufacturers will provide investigational supply at no cost or below cost for academic and public-benefit research, particularly for off-patent generic opioids. Comparator drug supply for a 600-patient trial at investigational pricing could save $300,000 to $1,000,000. Reference standards and analytical materials from public-domain sources (NIST, USP) save additional cost.

What cannot realistically be obtained in-kind

Honesty about the floor matters. The following cannot be expected as substantial in-kind contributions:

• GMP-grade contract manufacturing. GMP capacity is constrained and expensive; CDMOs do not give it away.

• Stability program execution. Stability requires controlled storage at multiple conditions for 24+ months. Real lab and instrument time, paid.

• Phase 1, 2, and 3 patient compensation, study coordinator time, and clinical-site operational costs. Even at academic-partnership pricing, these are paid.

• Regulatory submission preparation work at scale. The NDA submission alone is hundreds of staff-hours of regulatory writing.

• Long-term clinical staff salaries. A six-year program with a 4–8-person core staff requires $1.5–2.5 million per year in payroll. This is the largest unavoidable cost.

Revised cumulative totals — cash requirement after realistic in-kind

Maximizing in-kind contribution reduces the cash requirement by roughly 15 to 20 percent — meaningful, defensible to a funder, but not transformative of the program’s overall scale.

Part III — Optimized Strategy

Pulling every reasonable lever the public-domain space actually has access to.

The categories above are pure in-kind contribution. They do not capture the larger set of cost-reduction levers available to a well-structured public-benefit health-technology program. When all reasonable levers are activated together, the cash requirement to NDA submission falls substantially below the in-kind-only adjusted totals, into a range where philanthropic and federal-grant funding can realistically cover the program without requiring the cannabis-policy or wellness-philanthropy community to underwrite the entire arc alone.

Five additional levers materially change the strategy.

Lever 1 — Government research grants (NIH, NIDA, NCCIH, VA, DoD)

Realistic value: $15 to $30 million across the program lifecycle.

The federal research-grant environment for cannabinoid medicine is the strongest it has been in fifty years. NIH’s HEAL Initiative on pain and addiction includes a substantial cannabinoid research portfolio. NCCIH (National Center for Complementary and Integrative Health) explicitly funds botanical drug development. NIDA has historically funded cannabinoid pharmacology and substitution research. The Department of Veterans Affairs has an active interest in cannabinoid medicine for veteran chronic-pain and PTSD populations. The DoD Congressionally Directed Medical Research Programs include a Pain Management Research Program that has funded cannabinoid investigations.

A nonprofit-sponsored open-science FDA-pathway program with strong academic-partnership infrastructure is a natural fit. A Phase 2 chronic-low-back-pain trial could plausibly be funded as an NIH HEAL R01 ($3–5M direct costs) or a VA Merit Review award ($2–4M). A Phase 3 program component could be supported by NIDA cooperative-agreement mechanisms (U01, U54) at $5–15M.

The grant-writing competitive process is a real cost — figure 100–200 hours of senior staff time per major submission, with realistic success rates of 15–25 percent on first submission. The program should budget for two or three high-quality submissions per year across the lifecycle.

Lever 2 — VA / DoD partnership for chronic-pain trial execution

Realistic value: $5 to $12 million in Phase 2 and Phase 3.

The Veterans Health Administration operates the largest integrated health system in the United States, with documented chronic-pain and opioid-substitution research programs. A VA-co-sponsored trial leverages existing clinical infrastructure (sites, patients, electronic medical records, clinical staff) at far below commercial-CRO pricing. The VA has been a partner on several previous cannabinoid-medicine investigations and supported MDMA-PTSD research through the MAPS partnership. A VA cooperative-research agreement on a Phase 2 chronic-low-back-pain trial could plausibly reduce the trial’s net cost by 60–70 percent.

The DoD CDMRP funding pathway has supported cannabinoid investigations and could underwrite specific Phase 2 components targeting military and veteran chronic-pain populations.

The trade-off: VA and DoD partnerships impose specific protocol requirements, reporting obligations, and population-focus constraints. The chronic-low-back-pain program fits these constraints well; broader population trials require commercial-equivalent infrastructure.

Lever 3 — 505(b)(2) pathway with reference to published Phase 3 evidence

Realistic value: $25 to $40 million reduction in Phase 4 cost — and a 24-month timeline acceleration.

This is the largest single lever in the optimized strategy and the one that fundamentally changes the program’s economics.

The FDA’s 505(b)(2) New Drug Application pathway permits reliance on safety and efficacy data already in the public domain, including published clinical-trial data, to support approval of a related product. The Vertanical VER-01 Phase 3 trial in chronic low back pain (Karst et al., Nature Medicine 2025) is now public. The Phase 3 ELEVATE trial against opioids (Meissner et al., Pain and Therapy 2025) is now public. The full clinical-trial registries and clinical study reports for both will become accessible through standard regulatory disclosure timelines.

If the WPCE-01 program produces a full-spectrum extract characterized at a chromatographic profile within a defensible similarity envelope of the DKJ127 chemovar — or a defensibly different but comparably-characterized profile with bridging PK data — the 505(b)(2) pathway permits reference to the existing published efficacy evidence rather than de novo replication. The program would still need to produce its own Phase 1 PK bridge to demonstrate equivalent exposure-response, its own Phase 2 proof-of-concept dose-finding study, and a smaller confirmatory Phase 3 (potentially n=200–300 rather than n=600–820), but would not need to replicate the full Vertanical-scale pivotal program.

This is the architecturally most important lever. It is also the lever that the Usona Institute has explicitly designed for in its public materials: open-science publication creates prior art and reference data that downstream public-domain programs can lean on. The cannabis-medicine field is, after the Vertanical publications, at the threshold of this becoming operational. The same dynamic that gave Vertanical its commercial advantage now becomes the public-domain program’s accelerator.

Phase 4 cost under 505(b)(2) strategy: approximately $25 to $50 million rather than $55 to $100 million.

Lever 4 — Academic fiscal-sponsor model

Realistic value: $3 to $7 million in operational and infrastructure savings.

Rather than building a standalone 501(c)(3) with its own back-office, the program can be housed initially as an institutional initiative at a partnering academic medical center — UCSF, Johns Hopkins, UCSD, or comparable. The institutional sponsor provides governance infrastructure, grant-administration capacity, indirect-cost recovery through federal grants, payroll and benefits administration, and IRB and research-compliance services at well below the cost of building them from scratch.

The trade-off: the program operates under institutional oversight rather than independent governance, which complicates the public-benefit charter commitments. The Usona Institute model is the standalone-nonprofit precedent; an academic-housed initial structure could transition to standalone status at Phase 3 once operational scale justifies it.

Lever 5 — Manufacturer co-investment in Phase 4

Realistic value: $10 to $25 million in Phase 4 cost-sharing.

Post-Phase 2, with credible proof-of-concept data and a published Investigator’s Brochure that has acquired field-reference status, the program can credibly invite qualified manufacturers to co-invest in the Phase 3 pivotal program in exchange for early reference authorization to the Drug Master File and inclusion in the post-approval qualified-manufacturer ecosystem.

This is not equity in the program; it is paid participation in a regulatory commons that the manufacturer will be a beneficiary of. A coalition of 5–10 qualified manufacturers each contributing $1.5–3 million at Phase 3 entry can carry $15–25 million of the Phase 4 cost without compromising the public-domain governance commitment. The pricing of post-approval reference authorization is non-exclusive and accessible to manufacturers who did not participate in Phase 3 funding, preserving the open ecosystem; the Phase 3 co-investors simply get earlier authorization and seat-at-the-table participation in ecosystem governance.

This model is structurally similar to how the Critical Path Institute, the Pre-Competitive Consortium for Translational Innovation, and the Innovative Medicines Initiative (Europe) have funded shared infrastructure across multiple manufacturer participants. It is not naive; it is how pharmaceutical-industry consortia routinely structure cost-sharing on pre-competitive infrastructure.

Optimized cumulative totals

When all five additional levers are pulled, government grants, VA/DoD partnership, 505(b)(2) pathway, academic fiscal sponsor, manufacturer co-investment, alongside the pure in-kind contributions analyzed above, the cash requirement to NDA submission falls substantially.

The optimized strategy reduces the philanthropy-required cash to approximately $36 to $70 million across six years to NDA submission. This is roughly 40 to 55 percent of the all-cash baseline.

That is a fundable program at the scale of a single major foundation commitment (MacKenzie Scott has written $50M+ public-health checks routinely; Bloomberg Philanthropies, the Helmsley Charitable Trust, the Robert Wood Johnson Foundation operate at this scale; the cannabis-philanthropy and drug-policy network can collectively muster this scale across a coalition). It is no longer a $100M+ ask that requires a transformational lead donor; it is a $40–70M ask that fits inside the operational range of established public-health philanthropy.

Sequencing the strategy

The optimized scenario depends on sequencing; most of the cost reductions are conditional on producing strong intermediate artifacts that unlock the next set of levers.

Year 1 (Phase 1): Build the founding coalition and produce the WPCE-01 monograph, the Investigator’s Brochure v1.0, the analytical methods dossier, the IP-trust documents, and the published governance charter. Establish the academic fiscal-sponsor relationship. Convene the Scientific and Clinical Advisory Boards. First NIH/NCCIH/HEAL submissions filed in Q3–Q4. Initial pro bono legal partnerships engaged.

Year 2 (Phase 2): Pre-IND meeting with FDA, IND submission, Phase 1 PK study (academic site, partner pricing). First NIH grants beginning to fund. VA cooperative-agreement discussions opening. Pro bono legal infrastructure fully operational. Manufacturer-coalition cultivation begun.

Year 3 (Phase 3): Phase 2 proof-of-concept trial under NIH HEAL or VA cooperative agreement. Vertanical clinical-study reports become accessible through standard regulatory-disclosure timelines, opening the 505(b)(2) reference pathway analytically. Manufacturer coalition coalesces around credible proof-of-concept data.

Years 4–5 (Phase 4): 505(b)(2) Phase 3 confirmatory trial (n=200–300) under academic cooperative-group infrastructure with manufacturer co-investment. NDA preparation.

Year 6: NDA submission, FDA review, approval.

The sequencing has a structural property worth naming: every phase produces the public artifact that justifies the next phase’s funding. If Phase 1 fails to deliver the WPCE-01 monograph and the IB v1.0, the academic and grant infrastructure for Phase 2 does not engage. If Phase 2 fails to deliver the IND and the Phase 1 PK readout, the VA cooperative agreement does not close. If Phase 3 fails to deliver credible proof-of-concept, the manufacturer coalition does not materialize. This is a disciplined milestone-gated structure that funders find easier to underwrite than open-ended commitments.

What the actual ask becomes

A realistic, fully-optimized version of the ask, for the founding coalition to present to philanthropic funders, foundation program officers, and the cannabis-medicine community organizing around the proposal:

Phase 1 seed: $3 to $5 million from a founding coalition of two or three lead philanthropists or a single foundation grant. Produces the WPCE-01 monograph, the IB v1.0, the IP-trust infrastructure, the academic-partnership letters of intent, and the first NIH submissions.

Phase 2 ramp: $5 to $10 million to fund the IND-enabling work and the first-in-human study, with $3–5 million expected from NIH/HEAL/NCCIH grants and the balance from foundation philanthropy.

Phase 3 proof of concept: $5 to $10 million in cash to the program plus $10–15M in government-grant pass-through (NIH, VA, DoD cooperative agreements) covering the bulk of the Phase 2 clinical trial.

Phase 4 pivotal program: $20 to $40 million in cash to the program, with $10–25 million expected from manufacturer coalition co-investment and government grants covering the balance of the trial.

Phase 5 launch and ongoing ops: $4 to $8 million per year, transitioning to self-sustaining via non-exclusive licensing fees from qualified manufacturers.

Total philanthropy ask: $36 to $70 million over six years.

This is the version of the strategy that is most reasonable to present. It is conservative in its discount assumptions, defensible to a foundation program officer, and structurally honest about what the federal-grant and academic-partnership and manufacturer-coalition infrastructures can actually carry. It is also the version that gives the cannabis community a realistic path: not “we need to raise $150 million,” which is daunting and probably impossible, but “we need to assemble a $40–70 million coalition over six years, with most of it coming from established public-health philanthropy and federal-research-grant infrastructure that already exists.”

The window remains the next eighteen months.

Companion document to the published Public Domain Whole-Plant Cannabis Medicines Program proposal. Numbers are calibrated to mid-2026 U.S. and European pharmaceutical-services market rates, reflect realistic CMO and CRO pricing for adjacent peptide and botanical programs, and are intentionally banded rather than point-estimated. Designed as the basis for a detailed bottom-up budget developed with named manufacturing partners, named clinical-trial sites, and a defined sponsoring nonprofit governance structure, the next step after a credible founding coalition is assembled.

The federal government has just published the policy roadmap for what comes next. An open letter to the cannabis community on the eve of the June 29 DEA rescheduling hearing.

On September 29, 2025, Nature Medicine published a Phase 3, multicenter, placebo-controlled, randomized clinical trial of a full-spectrum extract from Cannabis sativa, the kind of trial cannabis-medicine advocates have been demanding for two decades. The drug, called VER-01, comes from a specific cultivated cannabis chemovar designated DKJ127. The trial enrolled 820 adults with chronic low back pain. It met its primary endpoint. Pain dropped. Sleep improved. Adverse events were mostly mild and transient. There were no signs of dependence or withdrawal.

A day later, the same research consortium published a second Phase 3 paper in Pain and Therapy: the ELEVATE trial, head-to-head against opioids in 384 patients. VER-01 came out superior to opioids on both efficacy and gastrointestinal tolerability. Fourfold less constipation. Threefold less laxative use. Sleep improvements. Pain reduction. The cannabis substitution effect that Marcus Bachhuber documented epidemiologically in 2014 and that Ashley Bradford and David Bradford documented economically in 2016 has now been confirmed at the highest evidentiary level the field can produce.

Among the authors of the Nature Medicine paper is Winfried Häuser, the German Cochrane systematic-review author who has spent two decades as the most prominent European skeptic of cannabis for chronic pain. Häuser’s name on a positive Phase 3 RCT means the data cleared the highest bar in the field. This is the scientific outcome the cannabis-medicine community has been waiting for since the 1990s.

Take a moment with what just happened. Then read what comes next.

The company that ran those trials is a German pharmaceutical firm called Vertanical. Its founder, Dr. Clemens Fischer, is a former Novartis executive board member, Strategy and Cardiovascular divisions, Germany. Its CEO came up through the FUTRUE Group, a Fischer-founded serial-entrepreneurship vehicle whose previous portfolio companies have been acquired by Novartis, Dermapharm, and Perrigo. Its Chief Commercial Officer is an ex-McKinsey Associate Partner. Vertanical has filed for marketing approval of VER-01 in Germany for chronic low back pain. It has planned trials in osteoarthritis and painful diabetic peripheral neuropathy. Its stated mission is to establish VER-01 as “the leading therapy for chronic pain by 2030.”

VER-01 is a full-spectrum extract. The Nature Medicine paper specifically characterizes “cannabinoids, terpenes, flavonoids, carotenes, phytosterols, vitamins and fats” and emphasizes batch-to-batch chromatographic fingerprinting consistency across multiple manufacturing lots. This is not Epidiolex. This is not Marinol. This is not Syndros. This is whole-plant cannabis, characterized, standardized, manufactured to GMP, and developed for prescription distribution under patent.

Cannabis advocates have spent two decades arguing that whole-plant cannabis is therapeutically superior to isolated cannabinoids because of the synergistic action of minor cannabinoids, terpenes, flavonoids, and the rest of the phytochemical complex. The Russo “Taming THC” paper in 2011 was the canonical statement. The entourage-effect literature has grown for over a decade, mixed but supportive in the indications that are most significant (epilepsy, chronic pain, anxiety), skeptical in the molecular-pharmacology details (whether terpenes directly bind cannabinoid receptors). The honest scientific position is that the entourage effect is probable in some clinical contexts, unproven in others, and worth disciplined empirical investigation.

Politically, that argument did real work. It gave the dispensary economy a scientific justification against Epidiolex: pharma’s purified CBD is therapeutically inferior to our whole-plant medicine, therefore our access architecture is providing better medicine than theirs. That argument held for ten years. It preserved a moral and scientific high ground for community-cultivated, dispensary-distributed cannabis even after Jazz Pharmaceuticals paid $7.2 billion to acquire GW Pharmaceuticals. It is the single best argument the cannabis community has ever made.

Vertanical just won that argument for you.

And by winning it, lost it.

The recuperation move

There is a word for what just happened. The French Marxist Guy Debord called it recuperation, the move by which capital absorbs critique by incorporating the content of the critique while neutralizing its political force. Debord was writing about advertising, fashion, and culture. The same move runs through every domain in which a social movement develops a strong argument against the prevailing system, and the prevailing system responds by adopting the argument and offering it back as a product.

Punk rock was recuperated into a music-industry genre. The labor movement’s safety demands were recuperated into corporate HR. Environmental critique was recuperated into ESG disclosures. Indigenous food sovereignty was recuperated into “ancient grains” supermarket aisles.

The cannabis movement just got recuperated at the molecular level. Pharma has now conceded the entourage effect by adopting it. The advocates’ strongest scientific argument has become the patent holder’s marketing copy.

Read the Nature Medicine abstract carefully. The wording is exact. “Full-spectrum extract.” “Comprehensively characterized.” “Chromatographic fingerprinting confirmed batch-to-batch consistency.” These are the exact phrases cannabis advocates have been waiting twenty years to read in a flagship medical journal. They are also the foundation of a patent estate that the dispensary economy cannot compete with.

The entourage-effect critique was a scientific argument about therapeutic profiles. It was not a political-economic argument about access architecture. Cannabis advocates assumed that defeating the isolate model would secure the dispensary. They did not anticipate that pharma would simply concede the science and capture the access architecture anyway.

That is the analytical incompleteness, and it is now coming due.

Three layers of capture

The cannabis-pharma capture is no longer a single architecture. It is three architectures, layered on top of each other.

Layer 1: Single-cannabinoid isolation. Marinol (synthetic THC, 1985). Syndros (oral dronabinol solution, 2016). Epidiolex (purified CBD from GW Pharmaceuticals / Jazz Pharmaceuticals, 2018). The pattern: take one molecule from the plant, purify it, patent the formulation, run trials, file with the FDA, get the patented isolate descheduled or moved to Schedule III or V while leaving the plant in Schedule I. This is the layer the cannabis advocates’ entourage-effect critique was correctly aimed at. It is also now the smaller part of the picture.

Layer 2: Standardized full-spectrum extract under cultivar and process patents. Vertanical’s VER-01, with Phase 3 RCTs in Nature Medicine and Pain and Therapy, filed for German marketing approval and targeting US and global expansion through 2030. MMJ International Holdings’ MMJ-001, pending FDA approval for a rare neurological disease, publicly described as a “plant-based” cannabis drug pursuing the FDA-regulated pathway. This is the layer that absorbs the entourage-effect critique by adopting full-spectrum methodology, then preserves enclosure through patents on the cultivar, the chromatographic fingerprint, the extraction-and-standardization process, and the medical-use claims for specific indications. The patents are broader and harder to design around than Epidiolex’s narrow composition-of-matter claims. They are the seed-patent model, Monsanto-Bayer agricultural IP, applied to medical cannabis.

Layer 3: Cultivar IP stacked with delivery-system patents. Sublingual films. Oral thin films. Transdermal patches. Metered-dose inhalers. Enteric softgels. Once a full-spectrum extract is patented at Layer 2, the delivery technology required to administer it pharmaceutically becomes the next patent frontier. Lock the cultivar AND the delivery system, and the entire pathway from plant to patient is enclosed. The dispensary cannot match the delivery technology any more than it can match the cultivar specification. The pharmaceutical-formulation industry is building this layer right now. The major buccal-film, oral-thin-film, and enteric-softgel platforms, the technical architecture of Layer 3, are being engineered for peptide therapeutics, GLP-1 reformulations, and pharmaceutical psychedelics alongside cannabis. The cannabis application is one of many that share the same patent stack.

By the time the dispensary economy understands what’s happening, all three layers will be operational. Vertanical alone is a Layer 2 firm. The full Layer 1 + Layer 2 + Layer 3 stack is what the next decade looks like.

The federal-policy companion piece

If you needed proof that the federal government is actively building the regulatory infrastructure to support this architecture, the proof was published by the White House on May 4, 2026.

The 195-page 2026 National Drug Control Strategy, issued under President Trump and signed by ONDCP Director Sara Carter, a Sinclair Broadcasting / Hannity-contributing journalist installed at ONDCP without public-health credentials, is hosted by the White House and reads as the cannabis-criminalization companion piece to the Vertanical commercial pathway. Read it directly.

Five things the document does that the cannabis community needs to read:

It explicitly elevates the FDA-approved labels of Marinol and Epidiolex as the federal government’s authoritative reference source on cannabis health risks. Page 172 directs consumers and clinicians to read the Marinol prescribing information “to illustrate some of the medical risks associated with cannabinoids” and to read the Epidiolex labeling “to inform users of non-prescription CBD products.” This is the National Drug Control Strategy explicitly naming the patent-protected pharmaceutical isolates as the textbook for understanding the cannabis plant.

It criminalizes the dispensary economy as foreign-terrorist-adjacent organized crime. Chapter 3 frames legal-state cannabis cultivation as cover for transnational trafficking, asserts that “criminal groups run more than 80 percent” of California’s marijuana farms, and folds domestic cannabis enforcement into the same Homeland Security Task Force structure used against fentanyl cartels designated as Foreign Terrorist Organizations. Material-support-for-terrorism statutes are invoked.

It schedules the closure of the hemp loophole for November 2026. The “Hemp Restriction regulations” passed in the FY 2026 Agriculture Appropriations Act will reclassify delta-8 THC, delta-10 THC, THC-O-acetate, THCP, and other hemp-derived cannabinoids as Schedule I controlled substances four months after the June 29 DEA rescheduling hearing concludes. The post-2018-Farm-Bill hemp-derived cannabinoid ecosystem: the small-producer, small-retailer ecosystem that has been the de facto national cannabis distribution channel in non-legal states, is being killed by federal regulation, on a timetable the cannabis industry has not yet processed publicly.

It calls for accelerated pharmaceutical R&D in cannabis-use-disorder treatment. Page 52: “While there are currently no medications approved for treatment of methamphetamine, cocaine, or marijuana addiction, pharmaceutical research in this domain should be accelerated.” The federal drug-policy apparatus is explicitly inviting the pharmaceutical industry to build a new franchise, cannabis-use-disorder medications, for a population whose access to non-pharmaceutical cannabis just got criminalized.

It builds the cannabis-induced-psychosis framing into every chapter. The “11-fold psychosis risk for ages 12 to 19,” the claim that “marijuana has the highest conversion rate from psychosis to schizophrenia,” the assertion that cannabis is the “number one drug found in toxicology reports of people who died from suicide under the age of 25 in Colorado and San Diego.” The empirical literature is contested; the rhetorical work is unmistakable. These claims justify the criminalization of high-potency cannabis while leaving room for “carefully dosed, FDA-approved, pharmacist-mediated” cannabis to be presented as the responsible alternative.

Whichever way one reads the document’s public release timing, premature against the political calendar of the June 29 DEA hearing, a deliberate signal to the GOP-aligned drug-warrior base, or just standard end-of-fiscal-year ONDCP output, the document itself is in the public record and the November 2026 hemp restriction is in writing. The Strategy is not a leak; it is a published federal policy posture, hosted by the White House, and it ought to be read.

What the dispensary economy now faces

The June 29 DEA hearing in Arlington is on whether cannabis moves from Schedule I to Schedule III. The April 23, 2026 order from Acting Attorney General Todd Blanche already placed FDA-approved cannabis products and state-licensed medical marijuana in Schedule III. The hearing will determine whether the broader cannabis economy joins them or stays in regulatory limbo.

Here is what the policy stack looks like if the cannabis-friendly faction loses the procedural fights at the hearing:

• FDA-approved pharmaceutical cannabis (Epidiolex, soon VER-01 if the German approval extends through the 505(b)(2) pathway to US markets, MMJ-001 if it clears, the next wave) sits comfortably in Schedule III, prescribed through doctors, sold through pharmacies, reimbursed by insurance, manufactured at GMP, patent-protected, priced at pharmaceutical pricing.

• State-licensed medical marijuana sits in Schedule III but under the same federal regulatory shadow that has constrained banking, interstate commerce, and research for fifteen years.

• The recreational and adult-use dispensary economy stays in federal limbo, state-legal, federally illegal, and faces increasing enforcement pressure under the Homeland Security Task Force / FTO architecture described in the National Drug Control Strategy.

• The hemp-derived cannabinoid economy becomes Schedule I in November 2026.

• Home cultivation, community access, and the small-producer ecosystem face the same enforcement environment they faced before legalization, with the added complication that the official medical narrative now points to pharmaceutical-grade products as the legitimate medicine.

The pharmaceutical incumbents win on every layer simultaneously. The science argument is won, by them. The access architecture is won, by them. The federal-policy infrastructure is built, for them. The criminalization apparatus is restored, against everyone else.

The Public Domain Whole-Plant Cannabis Medicines Program

There is a counter-move. It is not abstract. It has a published proposal, technical specifications, a regulatory framework, and a precedent.

In February 2026 I published a detailed proposal for a Public Domain Whole-Plant Cannabis Medicines Program. The core idea: develop an open-source, full-spectrum cannabis extract standard (WPCE-01 — Whole Plant Cannabis Extract, version 01) with published specifications, analytical fingerprinting, stability data, and GMP-aligned manufacturing methods, paired with three open delivery platforms (orally disintegrating tablets, oral thin films, enteric capsules). Sponsor the FDA-grade clinical trials through a nonprofit. Publish the protocols, the statistical analysis plans, and the data. Patent only the delivery interfaces, hold those patents in trust under a public-benefit entity, and license them non-exclusively. The pharmaceutical-grade regulatory legitimacy that Vertanical is building behind patent walls gets built openly, governed transparently, and made available to any qualified manufacturer.

The model is the Usona Institute’s approach to psilocybin. Usona is the nonprofit that has run Phase 3 trials of psilocybin for major depressive disorder while keeping the underlying molecule in the public domain through open publication of the Investigator’s Brochure, the analytical methods, and the manufacturing standards. Usona is the reason Compass Pathways’ synthetic-psilocybin patent estate is contestable, the prior art is on record. Cannabis needs the Usona equivalent, and it needs it before the Vertanical-tier patent estates settle into prosecution.

This is not naive. It is technically buildable. The FDA’s botanical drug development framework allows it. Drug Master Files can be held by the nonprofit and referenced by multiple manufacturers. Non-exclusive licensing of the delivery platforms is a standard contract. The capital required is meaningful but not pharmaceutical-scale meaningful, staged crowdfunding plus philanthropy plus academic-clinical partnerships can build it on a timeline that matters.

What it requires is the recognition, by the cannabis community, that the access-architecture fight is the actual fight. Not the science fight. The science fight has been won and ceded. The access fight is what determines whether whole-plant cannabis remains a community medicine accessible to patients at population-health pricing or becomes a pharmaceutical product accessible to patients with pharmaceutical insurance at pharmaceutical prices.

What this requires of the cannabis community

The cannabis-medicine community is the largest constituency in this fight. It is also, structurally, the constituency that has been doing the work for two decades, the dispensary owners, the small farmers, the patient advocates, the harm-reductionists, the medical-cannabis physicians, the underground breeders who developed the chemovar profiles that pharma is now characterizing as “DKJ127” and similar designations.

Here is what the architecture I have described requires of you, concretely, in the next sixty days:

Read the 2026 National Drug Control Strategy directly. The 195-page document is hosted at whitehouse.gov. Download a local copy. Federal documents that contain inconvenient policy specificity have a way of being moved, paywalled, archived, or restructured. Know what your federal government has published. Understand that the November 2026 hemp restriction is in writing.

Stop arguing the entourage-effect fight as if it were the central political question. The science is in. Vertanical conceded it. The argument going forward is not whether whole-plant cannabis works, that question is settled, but who controls the access to it.

Read the Public Domain Whole-Plant Cannabis Medicines Program proposal. Forward it. Pressure-test it. Improve it. Find the organizational, financial, and clinical partners it requires. The window for building an open-source whole-plant cannabis standard as prior art against the Vertanical-tier patent estates is the next eighteen months. After that, the patents settle and the option closes.

Engage the June 29 DEA hearing with procedural seriousness, not just rhetorical opposition. The hearing will determine which version of cannabis access the federal government recognizes. The cannabis community needs lawyers, expert witnesses, and procedural advocates in the room. NORML has requested a seat. Other organizations should be doing the same. The procedural infrastructure of the hearing is where the architecture gets contested or conceded.

Stop treating the FDA pathway as the enemy. It is not. The pharmaceutical incumbents are the actors using the FDA pathway as an enclosure mechanism. The pathway itself can be used differently. Usona has proven that. The cannabis community needs to build its own FDA-pathway capacity through a public-benefit nonprofit before the field is fully captured by Layer 2 and Layer 3 patent holders.

The cannabis movement was built by patients, caregivers, growers, and communities, not by Novartis alumni and McKinsey partners. Its future will reflect that history only if the community recognizes the structural fight it is now in, in time to organize a structural response.

The fences are coming. They are visible. The November 2026 hemp restriction is written. The Vertanical patent filings are public. The 2026 National Drug Control Strategy is hosted and accessible. The June 29 hearing is on the calendar.

You won the science. Now win the ownership.

This piece is part of a larger investigation: What MAHA Doesn’t Touch, a five-part Substack series on pharmaceutical regulatory capture under the populist surface of the Make America Healthy Again movement, with the June 29 DEA cannabis-rescheduling hearing as the news peg. The cannabis carve-out is worked in detail in Part 2 of the series, publishing the week of June 8. Subscribe to receive future installments.

Disclosure

Del Potter is a working pharmaceutical scientist and the Founder and Chief Scientific Officer of Prism Sciences LLC and BioUnbound Inc. His commercial work intersects directly with subjects this piece addresses, including peptide patent and formulation IP, oral thin-film and enteric delivery platforms, and the political economy of pharmaceutical access. He does not write about specific products, formulations, partners, or investor relationships of his ventures in this publication. The Public Domain Whole-Plant Cannabis Medicines Program proposal he has published here is the explicit political-economic alternative to the enclosure architecture he is critiquing. Readers should know the intersection exists and weigh the writing accordingly.

Editorial note

An earlier draft of this piece described the 2026 National Drug Control Strategy as having been “published and quietly withdrawn” from ONDCP distribution within days of its May 4, 2026 release. On further verification, the 195-page document remains publicly hosted at whitehouse.gov as of the date of this publication. The withdrawal framing has been removed and the piece revised to characterize the document accurately. The substantive content of the Strategy described above: the elevation of the Marinol and Epidiolex labels as the federal government’s authoritative reference on cannabis health risks; the criminalization framing of state-licensed cannabis cultivation; the November 2026 hemp restriction timeline written into the FY 2026 Agriculture Appropriations Act; the explicit call for accelerated pharmaceutical R&D on cannabis-use-disorder treatment; and the cannabis-induced-psychosis rhetorical infrastructure, is documented and verifiable in the linked PDF. The argument of this piece does not depend on the withdrawal framing. It depends on the document’s content, which is hosted and accessible.

Sources and citations

Bachhuber, Marcus A., Brendan Saloner, Chinazo O. Cunningham, and Colleen L. Barry. “Medical Cannabis Laws and Opioid Analgesic Overdose Mortality in the United States, 1999–2010.” JAMA Internal Medicine 174, no. 10 (October 2014): 1668–73. https://doi.org/10.1001/jamainternmed.2014.4005.

Bradford, Ashley C., and W. David Bradford. “Medical Marijuana Laws Reduce Prescription Medication Use In Medicare Part D.” Health Affairs 35, no. 7 (July 2016): 1230–36. https://doi.org/10.1377/hlthaff.2015.1661.

Carter, Sara. National Drug Control Strategy 2026. Washington, DC: Executive Office of the President of the United States, Office of National Drug Control Policy, May 4, 2026. 195 pp. Published May 4, 2026; subsequently withdrawn from official ONDCP distribution. PDF on file with the author.

Karst, Matthias, Winfried Meissner, Sabine Sator, Jens Keßler, Volker Schoder, and Winfried Häuser. “Full-spectrum extract from Cannabis sativa DKJ127 for chronic low back pain: a phase 3 randomized placebo-controlled trial.” Nature Medicine 31, no. 12 (December 2025): 4189–4196. https://doi.org/10.1038/s41591-025-03977-0.

Meissner, Winfried, Charles Argoff, Sabine Sator, Volker Schoder, et al. “VER-01 Shows Enhanced Gastrointestinal Tolerability, Superior Pain Relief, and Improved Sleep Quality Compared to Opioids in Treating Chronic Low Back Pain: A Randomized Phase 3 Clinical Trial.” Pain and Therapy 14, no. 6 (December 2025): 1765–1782. https://doi.org/10.1007/s40122-025-00773-z.

MMJ International Holdings. “MMJ International Redefines Medical Cannabis with First Plant-Based Drug for a Rare Neurological Disease (MMJ-001).” Press release, July 8, 2025. https://www.prnewswire.com/news-releases/mmj-international-redefines-medical-cannabis-with-first-plant-based-drug-for-a-rare-neurological-disease-mmj-001-302498745.html.

NORML. “NORML Seeks Seat at DEA Marijuana Rescheduling Hearing, Says Consumers Must Be Represented.” May 26, 2026. https://norml.org/blog/2026/05/26/norml-seeks-seat-at-dea-marijuana-rescheduling-hearing-says-consumers-must-be-represented.

Potter, Del. “Proposal for a Public Domain Whole-Plant Cannabis Medicines Program.” Del Potter PhD, Substack, February 1, 2026. https://delpotterphd.substack.com/p/proposal-for-a-public-domain-whole.

Potter, Del. “Cannabis Is About to Become Legal and Unaffordable.” Del Potter PhD, Substack, December 22, 2025. https://delpotterphd.substack.com/p/cannabis-is-about-to-become-legal.

Russo, Ethan B. “Taming THC: potential cannabis synergy and phytocannabinoid-terpenoid entourage effects.” British Journal of Pharmacology 163, no. 7 (August 2011): 1344–1364.

Vertanical GmbH. Company website and development pipeline. Accessed May 2026. https://vertanical.com/en/.

The Question

There is one question worth asking in American politics right now, and it is not the one our newspapers and our chatbots and consultants spend their days asking. It is not whether the swing voter in Erie County will tolerate this or that turn of phrase. It is not whether the Democratic brand needs more populism or less. It is not whether the party should pivot to the center or hold the base. It is not whether Mamdani’s victory in New York “generalizes.”

The question is older and harder, and it is the one our politics has been built, for forty years, to refuse. The question is: who owns what.

Who owns the housing stock that determines whether the working class can afford to be alive in this country. Who owns the medical system that determines whether a sick child gets treated. Who owns the transit that determines whether a home-health aide can reach her client before her wages are eaten by parking. Who owns the energy grid that determines whether the climate transition can be paid for at the timescale physics actually requires. And, increasingly, urgently, in the last decade and the next: who owns the machines. Who owns the foundation models trained on our collective output. Who owns the data centers cooling themselves with the water our grandchildren will not drink. Who owns the algorithmic infrastructures that are even now restructuring every wage bargain in the economy below the level of public deliberation.

The Mamdani victory is important because, for the first time in fifty years, a major American politician built a coalition around the recognition that this is the question. He did not run on personality. He did not run on identity. He did not run on opposition to Donald Trump or restoration of an imagined civic decency. He ran on who owns the buses, the housing, the childcare, the city. He won. He won outright, in a city the establishment thought it had under permanent management. The question is on the table now. It cannot be unasked.

What follows is an argument about what we do with it.

The Long Misdirection

To understand why the ownership question feels new in American politics, you have to understand what the Democratic Party spent the last four decades doing instead, and you have to be precise about it, because vagueness is how this gets buried again. The party did not “lose its way.” It did not “abandon the working class” as a matter of cultural taste. It executed a deliberate, donor-driven, ideologically articulated migration toward what Wolfgang Streeck, in the indispensable diagnosis of Buying Time and How Will Capitalism End?, has called the convergence of the European center-left on capital’s preferences. The American version of that migration is bourgeois liberalism.

I use the word bourgeois deliberately, and not as a slur. I use it in the precise sense Marx used it and the sense Streeck still uses it: as the political expression of a propertied class settlement, organized around the defense of asset values, financial returns, and the structural autonomy of capital from democratic claim. Bourgeois liberalism is the politics of a coalition between the upper-middle professional stratum, the financial sector, and the technology rentiers, mediated by a party leadership that draws its donors and most of its personnel from those same precincts. It is generous on cultural questions where culture costs the donor base nothing and stingy on economic questions where economics costs the donor base everything. It is in favor of more women on corporate boards and against any structural challenge to the corporate form. It is in favor of marriage equality and against meaningful taxation of inherited wealth. It is in favor of acknowledging climate change and against any plan that would write down the value of carbon assets on the timeline science requires. It is in favor of immigrants in the abstract and unwilling to defend them in the particular, because particular defense requires confrontation with the contractor class that profits from their persecution, and the contractor class is part of the donor coalition.

The institutional form was the Democratic Leadership Council, founded in 1985 with the explicit project of severing the party from organized labor and reattaching it to the rising financial-technological fraction of the capitalist class. The intellectual form was the Third Way and the New Democrat. The political form was Bill Clinton, who repealed Glass-Steagall and signed welfare reform and NAFTA and the 1994 crime bill and called it modernization. Then Barack Obama, who allowed the public option to die in committee, bailed out the banks while letting the foreclosure crisis run its course on the homeowners, and treated the labor movement as a constituency to be managed rather than a base to be empowered. Then Joe Biden, who recovered some of the lost vocabulary about working people and delivered, in the end, an industrial policy substantial enough to matter but too modest to break the underlying pattern. Then Kamala Harris, whose campaign in 2024 oscillated between economic populism and donor appeasement in a way that read, accurately, to the working-class voters of the Rust Belt as confusion about whose side she was on.

Forty years of this. Forty years of being told that the party could not move too fast, could not say the word socialism, could not name the donor class as donor class, could not run on the program. Forty years of being told the working class had been lost to cultural reaction and would have to be won back, gradually, with modest demands and careful messaging. Forty years during which the labor share of national income collapsed from sixty-five percent to fifty-eight percent and the top one percent’s share of wealth grew from twenty-three percent to nearly thirty-five. Forty years during which roughly fifty trillion dollars, as the RAND Corporation has documented, was transferred from the bottom ninety percent of Americans to the top one percent. Forty years during which the median rent grew faster than median income in every major American city, every year, with the exception of the brief pandemic interregnum, while the party leadership insisted that the median rent and the median income were not the kind of thing a federal campaign should organize itself around. Forty years during which the working-class majority of this country, of every race and region, was systematically delivered into the political custody of a Republican Party that was, beneath the racial demagoguery, faithfully executing capital’s actual program of upward redistribution.

This is what I mean by the long misdirection. It was a class settlement, executed in public, against the working-class majority that the party still nominally claimed to represent. Its result, in November 2024, was the second Trump administration, the most nakedly oligarchic government in American history, the one whose program, executed by Project 2025 operatives and Federalist Society judges, is the open dismantling of the constitutional order on behalf of a fraction of capital that has concluded democracy is now too expensive. Bourgeois liberalism produced this. Its accommodations made this possible. It did not see the danger because the danger was, at first, profitable to its constituents. By the time the danger turned on the constituents themselves, it was too late.

The Mamdani victory is the first sign in a generation that there is a politics on the other side of this settlement. It is the first sign that the long misdirection can be ended. It does not, by itself, end it. The work of ending it has to be done across the entire party, in every state and every district, and it is going to take more than one election. But the door is open now in a way it has not been open since the Sanders campaign of 2016, and arguably since the New Deal coalition began to fracture under Carter. We have to be clear about what walks through it.

The Principle

What the Mamdani frame proposes, and what the new Democratic program has to be organized around, is a principle that bourgeois liberalism has refused for forty years and that the European social-democratic tradition has been operating on for a hundred. The principle is decommodification, and its corollary is ownership.

Decommodification means removing essential goods, shelter, transit, healthcare, childcare, education, water, energy, from the discipline of accumulation. It means treating them not as commodities to be allocated by the market but as conditions of being alive to be guaranteed by public provision. Karl Polanyi, in The Great Transformation (1944), described the original commodification of labor, land, and money as the great catastrophe of the nineteenth century, the act by which capitalism produced the social wreckage that fascism then exploited. He argued, correctly, that decommodification was the corrective that had stabilized the postwar democracies. We have spent forty years undoing his correction, and the results have been precisely the results he predicted: collapsing legitimacy, demagogic reaction, the return of the radical right. The decommodification agenda is the operational lesson of the twentieth century, taught at considerable cost, and refused by our own political class out of donor discipline.

Ownership is the deeper claim. Decommodification asks who pays. Ownership asks who decides. A program that subsidizes private housing is not the same as a program that builds public housing. A health insurance market with public subsidies is not the same as a single-payer system. A green-tech industry supported by public R&D and private equity is not the same as a publicly owned utility transitioning at the speed climate demands. The difference is the difference between a politics in which the public bears the costs and capital captures the gains and a politics in which the public bears the costs and the public captures the gains. Bourgeois liberalism has spent four decades preferring the former arrangement, and has spent four decades calling the latter arrangement unrealistic. It is the most realistic arrangement on earth. It is the arrangement that runs every functioning rich country except this one.

The empirical record on ownership is overwhelming, and we should be unembarrassed in citing it. Vienna houses sixty percent of its population in non-market or regulated tenure built and managed by the municipal authorities; rents in Vienna are roughly half those of comparable European cities, and the housing stock has been continuously maintained for a hundred years. Norway holds its oil revenues in a sovereign wealth fund that constitutes the largest single pool of socialized capital on earth, governed by ethical guidelines that prohibit investment in arms and corruption and distributed in long-run benefit to the entire citizenry. The Mondragón Corporation in Spain operates as a federation of worker-owned cooperatives covering seventy thousand worker-owners, twelve billion euros in annual revenue, and a documented productivity premium of twenty percent over comparable investor-owned firms; in recession it redistributes hours rather than firing workers, and has held that policy for sixty years. The Emilia-Romagna region of Italy derives roughly a third of its regional GDP from cooperatives and posts among the highest standards of living in Europe. The John Lewis Partnership in Britain has operated as a fully employee-owned retail enterprise across eight decades. The American ESOP sector covers fourteen million workers across some sixty-five hundred firms, with documented gains in productivity, retention, and pay equity over comparable investor-owned firms.

None of this is theoretical. None of it is utopian. All of it is the operational answer to the question of ownership, deployed at scale in dozens of jurisdictions, outperforming the privatized American model on the metrics the American model claims to optimize for. The reason these examples do not appear in our political discourse is that they cannot appear in the discourse bourgeois liberalism is willing to host. They falsify its central claim. The Mamdani frame proposes that we put them at the center, as the empirical baseline of what democratic-socialist policy actually means, and that we organize the Democratic Party around the project of installing them here.

Who Owns the Machines?

This is where the ownership question opens onto its decisive twenty-first-century horizon, and it is the part of the Mamdani frame that has not yet been articulated at full resolution but that the logic of the program reaches for inexorably.

Let me put it concretely first, because the concrete is where the politics will be made or lost.

My niece is a contract editor for a publishing company based in San Francisco. She is forty-one. Over the last three years, that company has replaced roughly a third of her colleagues with what it calls augmented workflow tools and what she calls the machine. The machine is a fine-tuned large language model that produces a first pass of any manuscript. Her job is now to clean up what the machine produces. The machine was trained on the work of editors like her, work she had been paid to do over the previous decade, work she had been paid to do at rates the firm now considers excessive for human hands. The machine is good. It is, she says, almost as good as she is. The not-quite is where the job lives. Each year the not-quite gets smaller. Each year her rate, accordingly, gets lower. At her current rate of compression, she calculates, she will be priced out of her profession in roughly five years.

I asked her who owned the machine.

She said: a company in California, owned by another company in California, which is owned by a third company that is half-owned by a sovereign wealth fund in the Gulf and half-owned by a holding company in the Cayman Islands.

I asked her who had built the machine.

She said: I did. We did. They scraped the whole internet, Uncle Del. Every line edit I made in a shared document they could see. They scraped it and they trained the machine on it and now they’re selling the machine back to my employer at a rate that lets my employer cut my pay.

This is the ownership question in its present form. It is not a metaphor and not a forecast. It is the operating condition of a substantial fraction of the knowledge economy, and the operating condition is moving down the wage ladder fast, into the warehouses being algorithmically dispatched by Amazon, into the kitchens whose schedules are being optimized by software, into the home-health rotations being routed by an app, into the rideshare and delivery fleets whose drivers are managed by a model they will never see and can never appeal. The cost of being managed by the machine is borne by labor. The gains are captured by capital. The substrate on which the machines were built, the publicly funded R&D, the publicly subsidized chip fabs, the publicly produced training data, the publicly built energy grid, was socialized. The returns are privatized. The five largest firms in the sector hold cash reserves larger than the GDP of most nations. Their valuations are sustained by what Marx called fictitious capital, value detached from any present relation to production, dependent on the assumption that the rents they extract will continue indefinitely.

The ownership question, applied to this situation, generates a program. The program is not Luddite. It does not reject the technology. It claims the technology, as the public’s own, and disposes it accordingly. The relevant items:

Public compute as a regulated utility, available to researchers, schools, public institutions, and small enterprises at non-extractive rates, on the postwar telecom model.

A public data commons, in which the corpus of human cultural production that was scraped to train the models is held in trust by a public body, with licensing terms set by democratic deliberation and the rents flowing back to producers and to the public, on the Norwegian sovereign-wealth-fund model.

Public foundation models, trained at public facilities by publicly employed scientists, available at marginal cost as a baseline alternative to private extraction, on the model the BBC has provided against commercial broadcast for a century.

Cooperative platforms on the Stocksy and Fairbnb model, in which the workers and users who generate the value also own the infrastructure.

A wealth tax on the AI rentier formations, calibrated to claw back the public subsidies they have absorbed and the public infrastructures they have privatized.

A guaranteed labor share of automation gains, on the German codetermination model: worker representatives on the boards of firms above a threshold size, a binding sectoral bargain on how productivity gains are distributed, a public obligation to retrain and place workers displaced by automation at no loss of income.

Public ownership or strict utility regulation of the data-center build-out, with the externalities, water consumption, grid load, carbon, priced and absorbed by the firms benefiting rather than by the communities being colonized.

This is the AI program for the new Democratic direction. It is not a separate domain from the housing program. It is the same program in a different register. Decommodify the essential infrastructures. Public goods belong to the public. Ownership follows participation. The Vienna principle, the Norwegian principle, the Mondragón principle, applied to the dominant accumulation regime of the next twenty years.

Bourgeois liberalism cannot do this. It is structurally incapable of doing it, because the AI rentier formations are now central nodes of the donor base it depends on. The Democratic Party of Schumer and Jeffries will spend the next ten years performing concern about AI while accepting the campaign contributions that ensure no meaningful action is taken. The Democratic Party that follows the Mamdani frame will treat the machines question with the same clarity it treats the housing question, because the underlying logic is identical and the political coalition for the answer is the same.

The New Democratic Program

Let me draw together the program in the form it now has to take, because the moment for fuzziness is over. The party has to choose what it stands for, in language specific enough for working people to recognize their lives in it. What follows is the floor, not the ceiling. Every plank is operational somewhere in the rich world. Every plank is supported, when named, by majorities of the American electorate. The only obstacle is the political will of a leadership whose donor coalition has spent forty years suppressing the will.

Decommodified housing. A federal program to build public housing at scale on the Viennese model, fifteen million units over fifteen years, financed by federal bonds and held in permanent public or community-trust tenure. A national rent stabilization regime for existing units, with enforcement mechanisms that have teeth. The elimination of corporate single-family ownership through punitive taxation. The full elimination of the mortgage interest deduction for second homes and investor-owned units.

Universal healthcare on the single-payer model. Medicare for All, in the form Bernie Sanders has been advocating for thirty years. The American medical system spends roughly eighteen percent of GDP and produces some of the worst outcomes in the rich world. Every other comparable country spends nine to twelve percent and produces longer lives. The savings alone, properly captured, finance most of the rest of the program.

Universal childcare and pre-K on the Quebec model. Five dollars a day, federally subsidized, available to every family. Pierre Fortin and his colleagues at Sherbrooke have documented that the Quebec program returns approximately $1.04 in tax revenue for every dollar spent, within a decade, through increased female labor-force participation alone. James Heckman’s work establishes returns on early-childhood investment at seven to thirteen percent per year. This is the single most empirically settled investment available to a modern state, and the only reason we do not do it is bourgeois liberalism’s refusal to.

Free public transit and intercity rail. Free buses in every metropolitan area above a population threshold. A national high-speed rail program on the European model, publicly built and publicly owned. The 421-a subsidies the city of New York alone forgoes annually to luxury developers would fund free buses for the city ten times over. The math has never been the problem. The donors have been the problem.

Taxation of capital at rates that work. The top marginal rate restored to fifty percent on income above one million, and to seventy percent on income above ten million. A wealth tax on fortunes above fifty million, on the model Saez and Zucman have detailed in The Triumph of Injustice. A financial transactions tax. A robust estate tax with the loopholes closed. The flight thesis is empirically dead; Cristobal Young’s work has retired it. We can collect what we are owed. We have only to find the political will.

Labor power restored. The PRO Act. Sectoral bargaining as the default. Card-check recognition. Just-cause employment. The right to strike without permanent replacement. Codetermination at the firm level for any enterprise above a threshold size. A statutory floor of worker ownership at the cooperative or ESOP level for firms receiving any form of federal support. The European labor regime, even after forty years of erosion, delivers what ours does not. We can adopt it.

Climate transition as public mission. A federal Green New Deal funded at the Mazzucato scale, organized as mission-directed public investment rather than tax credit deference to private capital. Public ownership of the transmission grid. A national public renewable utility. The closure of fossil-fuel production on a timeline aligned with what the IPCC actually requires. Just-transition guarantees, with full income protection and retraining, for workers displaced by the transition.

Public ownership of the algorithmic infrastructures of the next economy. The AI program described above. Public compute as utility. Public data commons. Public foundation models. Cooperative platforms. Codetermination on the boards of the tech rentiers. A wealth tax to claw back the public subsidies they have absorbed. The treatment of the machine question with the same clarity as the housing question.

Solidarity without carve-outs. The dismantling of the ICE enforcement apparatus and its replacement with a humane and lawful immigration system. The end of US support for the bombardment of Gaza and for the apartheid order in Israel-Palestine, on the terms the international institutions of the rule of law have actually established. The decriminalization of poverty. The decarceration of the nonviolent prison population. The redirection of the police budget toward the social-service infrastructure that should never have been militarized in the first place. The full restoration of the Voting Rights Act and the federalization of election administration on the European model. The treatment of every working person in this country as a member of the polity, with the full claims that membership implies, regardless of paperwork.

This is the program. It is coherent. It is operational. It is supported by the electorate when it is named. The party can pick it up or not. The choice is consequential, and the choice cannot be postponed much longer.

Direction and Purpose

I want to close on the question of direction and purpose, because forty years of bourgeois liberal misdirection have produced not only a hollowed-out electoral coalition but a political class that no longer remembers what a party with direction looks like.

Direction is what a movement has when its demands cohere into a recognizable picture of what the country could be. The New Deal coalition had direction: a country in which the working class had a guaranteed share of national income, in which old age was not a condition of poverty, in which a major war could not be paid for by tax cuts at the top. The civil rights and antiwar movements of the 1960s had direction: a country in which the racial caste system was finally dismantled and the imperial machine was finally restrained. The Democratic Party under Clinton, Obama, and Biden did not have direction. It had positioning. It had message discipline. It had a series of carefully focus-grouped offerings designed to assemble a fifty-one percent coalition in a particular election cycle without disturbing the donor base. It produced four decades of partial and incoherent reforms whose net effect was the continuation of the upward redistribution that its rhetoric pretended to oppose. The voters noticed. The voters always notice. The voters punished the party in 2016 and 2024 with the most direct instrument available to them: the candidate of the opposing party who had at least committed to breaking something.

Purpose is what a movement has when it can answer the question what is this for. A party with direction can answer the strategic question. A party with purpose can answer the moral question. Why are we doing this? What kind of country are we trying to make? Whose lives are we trying to make better? Whose claims on the future are we willing to fight for? Bourgeois liberalism, in its American form, has not been able to answer these questions for two generations, because the honest answer would alienate the donor coalition. It has substituted, in the place of answer, the procedural commitment to defending democracy in the abstract while neglecting the material conditions that give democracy its content.

The Mamdani frame restores both. The direction is clear: a country in which the essentials of life are decommodified and the essential infrastructures, including the algorithmic ones, are publicly or cooperatively owned. The purpose is clear: a country in which working people are no longer extracted from at every turn, in which the racial and immigrant and queer working class has the same claim on the future as the propertied stratum, in which the next twenty years of technological transformation are governed by democratic deliberation rather than rentier capture.

This is what the new Democratic Party has to be for. It does not yet exist. It will exist if we build it. It will be built in the primaries that follow the next two cycles, in the union locals and tenant organizations and climate-justice coalitions that have been carrying the program in the absence of party leadership, in the state legislatures and city councils where the next Mamdanis are already running. The boomer left, my cohort, has spent forty years apologizing for its accommodations and will spend whatever years remain to it funding, voting for, and defending the formation that has, at long last, refused those accommodations.

The Mamdani inflection is not the end. It is the opening. The work is what we do with the opening. The work is the program, the direction, the purpose, all of it pursued with the analytical seriousness the moment requires and the moral clarity it deserves.

Who owns the machines is the question. Who owns the city is the question. Who owns the future is the question. The answer, finally, after a long and costly forty years, can begin to be the only honest answer there ever was: we do, or no one does, and no one does is the path bourgeois liberalism has put us on.

The new Democratic Party walks the other road, or it is no longer the Democratic Party in any sense that matters.

Bibliography

Blau, Eve. The Architecture of Red Vienna, 1919–1934. Cambridge, MA: MIT Press, 1999.

Fortin, Pierre, Luc Godbout, and Suzie St-Cerny. “Impact of Quebec’s Universal Low-Fee Childcare Program on Female Labour Force Participation, Domestic Income, and Government Budgets.” Working Paper 2012/02. Sherbrooke: Université de Sherbrooke, Research Chair in Taxation and Public Finance, 2012.

Harvey, David. A Brief History of Neoliberalism. Oxford: Oxford University Press, 2005.

Heckman, James J. Giving Kids a Fair Chance: A Strategy That Works. Cambridge, MA: MIT Press, 2013.

Madden, David, and Peter Marcuse. In Defense of Housing: The Politics of Crisis. London: Verso, 2016.

Marx, Karl. Capital: A Critique of Political Economy. Vol. 3. Translated by David Fernbach. London: Penguin Classics, 1991. First published 1894.

Mazzucato, Mariana. The Entrepreneurial State: Debunking Public vs. Private Sector Myths. London: Anthem Press, 2013.

———. Mission Economy: A Moonshot Guide to Changing Capitalism. New York: Harper Business, 2021.

Piketty, Thomas. Capital and Ideology. Translated by Arthur Goldhammer. Cambridge, MA: Harvard University Press, 2020.

Polanyi, Karl. The Great Transformation: The Political and Economic Origins of Our Time. Boston: Beacon Press, 1944.

Price, Carter C., and Kathryn A. Edwards. Trends in Income from 1975 to 2018. Working Paper WR-A516-1. Santa Monica, CA: RAND Corporation, 2020.

Saez, Emmanuel, and Gabriel Zucman. The Triumph of Injustice: How the Rich Dodge Taxes and How to Make Them Pay. New York: W. W. Norton, 2019.

Streeck, Wolfgang. Buying Time: The Delayed Crisis of Democratic Capitalism. Translated by Patrick Camiller. London: Verso, 2014.

———. How Will Capitalism End? Essays on a Failing System. London: Verso, 2016.

Vaswani, Ashish, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, and Illia Polosukhin. “Attention Is All You Need.” In Advances in Neural Information Processing Systems 30, edited by Isabelle Guyon, Ulrike von Luxburg, Samy Bengio, Hanna Wallach, Rob Fergus, S. V. N. Vishwanathan, and Roman Garnett, 5998–6008. La Jolla, CA: Neural Information Processing Systems Foundation, 2017.

Young, Cristobal. The Myth of Millionaire Tax Flight: How Place Still Matters for the Rich. Stanford, CA: Stanford University Press, 2017.

Companion Essays

Potter, Del. “On ICE.” Del’s Substack, December 21, 2025. https://delpotterphd.substack.com/p/on-ice.

———. “The Generation That Forgot Itself: How the Sixties Generation Became the Base of American Fascist Reaction.” Del’s Substack, April 18, 2026. https://delpotterphd.substack.com/p/the-generation-that-forgot-itself.

———. “Why Marx’s Critique of Capitalism Was and Remains Absolutely Correct.” Del’s Substack, November 8, 2025. https://delpotterphd.substack.com/p/why-marxs-critique-of-capitalism.

The California Billionaire Wealth Tax, the AI Enclosure, and the Refusal to Pay

Every serious conversation about a wealth tax in the United States eventually arrives at the same impasse, and it is worth understanding that the impasse is manufactured. Someone asks whether the wealthy will simply leave. Someone else observes that the rich already pay enormous amounts in tax, citing a single dramatic figure from a single anomalous year. A third party invokes the Laffer Curve and gestures at the supposed historical record of failed wealth taxation. A fourth produces a quotation from Thomas Sowell, deployed as the rhetorical equivalent of a closing argument that requires no premises. These objections circulate endlessly in American discourse, performing the work of reasonable doubt while the actual question, whether this country will permit a democratic claim on the wealth that the AI capital class has accumulated under conditions of effective tax exemption, remains safely deferred. The objections deserve answers, and the answers exist. That they are so rarely heard tells us more about the political function of the questions than about any genuine complexity in the policy.

The occasion for the present essay is the November 2026 California Billionaire Tax Act, which appears on the state ballot as the first measure anywhere in the world to put before voters the question of whether the personal wealth of resident billionaires should be subject to a one-time five-percent levy spread over five years. The measure was placed on the ballot by SEIU-affiliated health-care workers, advised by Emmanuel Saez and Gabriel Zucman of the Berkeley and Paris Schools of Economics, and is calibrated to fill the roughly one-hundred-billion-dollar hole that the Trump administration’s Medicaid cuts have opened in the state’s five-year safety-net financing. It has drawn the opposition of essentially the entire California billionaire class. Sergey Brin alone has reportedly spent at least fifty-seven million dollars on signature collection for competing initiatives and on direct opposition advertising. The measure has also drawn the assembled apparatus of conservative argument that the remainder of this essay will examine, refute, and locate within the longer political-economic analysis the preceding essays in this sequence have developed.

The measure is modest. It is also the most consequential single political instrument that will appear on any American ballot in the present cycle, because it is the first move in a reckoning that, if it does not occur in the next decade, will not be possible at all. The reckoning concerns the AI capital class, whose wealth has compounded at rates that exceed any prior accumulation in the recorded history of capitalism, under conditions of effective tax exemption, in plain view of every relevant democratic institution. The cognitive enclosure analyzed in The Cognitive Commons, the conversion of the entire archive of human cognitive labor into the privately held weights of a small number of corporate models, is the technological foundation of the present accumulation. The political-economic dispossession analyzed in They Work for the Bots, the consolidation of a welfare-feudal class structure under conditions of mass AI displacement, is its political endpoint. The wealth tax is the first political instrument adequate to interrupt the trajectory between the two. The objections assembled against it are the objections that, throughout the recorded history of distributive politics, have always been assembled against any meaningful redistributive measure. This essay anticipates each of them, answers each of them, and locates the answers within the structural analysis the prior essays have established.

The Architecture of a Specific Accumulation

The first move in any deflection from a wealth tax is universalization. Thomas Sowell, whose influence on conservative thinking about taxation considerably exceeds his scholarly rigor, exemplifies the technique: all states tax wealth in some form; all societies produce inequality; the rich already pay disproportionately; therefore the present configuration is merely one instance of a universal human pattern, and no specific intervention is required. The argument has a certain geometric tidiness. It is also analytically vacuous. I have examined the structure of Sowell’s broader anti-tax and anti-socialist project at length in The Intellectual Mirage, and will not repeat the full critique here. Note only that the universalizing move accomplishes its political work by dissolving the specific into the abstract, the way Frantz Fanon described the colonial bourgeoisie dissolving the violence of colonialism into observations about human nature: specificity is what generates political obligation, and so specificity is what the apologist must eliminate.

The empirical situation that the California measure addresses is precise, recent, calculable, and historically unprecedented, the opposite, in every analytical respect, of the universal pattern the apologists invoke. Saez and Zucman, working with the methodologies their long collaboration on the World Inequality Database has refined into the most cited research program on global wealth concentration, released two documents in May 2026: an opinion piece in The New York Times making the public case for the measure, and a Berkeley working paper, Taxing California Billionaires, presenting the underlying data and methodology in detail. The findings are these.

California’s roughly two hundred and fifty billionaire households held over two trillion dollars in combined wealth at the end of 2025. That figure represents 144 percent growth in three years, annual gains of 31, 36, and 37 percent in 2025, 2024, and 2023 respectively. The top four in this distribution, Sergey Brin and Larry Page of Alphabet, Mark Zuckerberg of Meta, Jensen Huang of Nvidia, together hold approximately three-quarters of a trillion dollars, an amount equal to roughly twenty-four percent of California’s gross state product. Twenty years ago the equivalent figure for the state’s ten richest residents was $134 billion; today it is nine times that. The growth is exponential, the inflection point sitting visibly in the 2023 launch of the current AI boom. The four wealthiest Californians are the principal capital incumbents of the cognitive-enclosure operation that The Cognitive Commons described: Alphabet, the dominant search and digital-advertising platform and a leading model developer through DeepMind and Gemini; Meta, which has reoriented its entire capital deployment around AI infrastructure and which trained the Llama family of models on the LibGen pirate archive after internal correspondence had flagged the legal exposure; Nvidia, the company that produces the GPU substrate on which essentially every large AI model in the world is trained.

These individuals paid, on average, 0.26 percent of their wealth annually in California income tax across 2019-2025. The top four paid 0.07 percent. For comparison, the ordinary California household pays an effective state-plus-federal rate around twenty percent. The asymmetry is the central feature of the architecture, structurally produced and politically defended across fifty years.

The argument that “we already tax the rich” obscures the architecture by referencing rates that apply to forms of income the AI capital class structurally does not realize. Brin and Page took $1 in annual executive compensation from Alphabet in 2019, 2020, and 2023. Their stock wealth rose by $133 billion across those years. They paid no California income tax on that growth, because they did not sell. ProPublica’s 2021 release of leaked IRS files, the Secret IRS Files investigation by Jesse Eisinger and colleagues, documented the systemic pattern: between 2014 and 2018, Jeff Bezos’s effective tax rate against wealth growth was 0.98 percent; Michael Bloomberg’s 1.30 percent; Warren Buffett’s 0.10 percent; Elon Musk’s 3.27 percent. The system the apologists call “already progressive” is, for the holders whose wealth is denominated in equity, structurally regressive at a magnitude unmatched anywhere in the historical record of American taxation. The wealth tax is the first instrument calibrated to recover any portion of what the present architecture exempts from a class whose contribution to the public fisc is, by every measurable standard, vanishingly small.

The Mechanism of the Untaxed Fortune

The mechanism by which the AI capital class accumulates wealth under conditions of effective tax exemption is the Buy-Borrow-Die strategy, documented in detail across the tax-scholarship literature by Edward McCaffery, Steven Rosenthal, and others, and now familiar enough that no serious analyst of American wealth distribution disputes its operation. The strategy has three stages and requires no exotic instruments. Hold the appreciated asset, almost always equity in a founder-held company. Borrow against the appreciated equity at low interest rates rather than selling, the way Mark Zuckerberg pledged $3.1 billion in Meta shares as collateral for a 2023 loan. At death, transfer the asset to heirs with a stepped-up basis under §1014 of the Internal Revenue Code, which extinguishes the accumulated capital gain entirely. The capital gain that was never realized in life is now legally erased. The tax that was deferred is permanently avoided. The heir receives the wealth at the new basis, holds it, borrows against it, and the cycle repeats.

Calling this a loophole misdescribes the architecture as a defect in it. The Buy-Borrow-Die sequence is the system as designed and operated. The income tax was constructed in 1913 against a wealth distribution in which the realization of income through wages, dividends, and asset sales was the modal pattern. The income tax has not been redesigned against the wealth distribution that has emerged over the past forty years, in which the largest fortunes are denominated in unrealized capital gains on founder-held equity that need never be sold to be deployed. The architecture protects the AI capital class structurally, through inertia rather than intention; through fifty years of legislative inaction against the wealth configuration that has, in plain view, replaced the one the code was written to address. The wealth tax addresses the architectural failure directly, by taxing the asset rather than the realization. It is, in operational terms, what the income tax would have been if it had been designed against the wealth distribution of 2026 rather than the wealth distribution of 1913.

The cognitive enclosure analyzed in the prior essay supplies the productive substrate on which the AI capital fortunes have been built. The training data, the entire archive of human cognitive labor, the books and academic papers and code and photographs and writings and recordings of a century of literate production, was, in the legal-political terms that govern this conversation, expropriated. The model weights derived from that training data are the means of production of the present moment. The compute infrastructure on which the models are trained and deployed has been built with substantial public subsidies whose return to the public has been minimal. The valuation of the firms that own the weights, the infrastructure, and the deployment depends on legal arrangements granted by public authority, copyright doctrine, fair-use doctrine, regulatory permissiveness, the legal exemptions that converted the cognitive commons into private property. The wealth that has accumulated to the four Californians whose fortunes the November measure would partially touch is, in significant part, the monetization of a primitive accumulation conducted by their firms against a commons the rest of us produced. The case for taxing that wealth is restitutive rather than redistributive. We are recovering, modestly, a portion of what was taken.

The Mobility Question: Answered

The objection that the wealthy will simply leave has acquired catechetical authority in American debates over wealth taxation, and it survives only because the people repeating it have not read the literature that disposes of it. The empirical literature is unambiguous, well-developed, and conclusive.

Cristóbal Young at Stanford, drawing on thirteen years of every IRS return filed by every American millionaire, found that the wealthy migrate less than the general population. They are anchored by their businesses, their families, their professional networks, and the agglomeration economies that produced their fortunes in the first place. Young’s work is the largest single empirical investigation of millionaire mobility in the history of American social science, and its conclusion stands against essentially the entire corpus of Wall Street Journal editorial-page assertion on the subject.

The two natural-experiment studies on actual wealth-tax mobility, Brülhart, Gruber, Krapf, and Schmidheiny (2022) on Switzerland, where wealth taxes vary across cantons, and Agrawal, Foremny, and Martínez-Toledano (2026) on Spain’s regional wealth tax, estimate mobility semi-elasticities between eight and ten. The revenue-maximizing wealth tax rate at those elasticities is approximately one over the elasticity, or roughly ten percent annually in perpetuity. The California measure is a one-time five percent spread over five years. It is an order of magnitude below the rate at which mobility responses begin meaningfully to bite. The empirical work has been done. The number has been calculated. The proposed rate is well within the range at which mobility responses are second-order effects.

The measure forecloses the migration question in any case, by applying only to billionaires resident in California as of January 1, 2026. The window for flight closed five months ago. And the arithmetic Saez and Zucman published is annihilating to the objection on its own terms: California billionaires currently pay so little state income tax that even if every one of them departed tomorrow, twenty-five years would pass before the lost income-tax revenue exceeded the one-time wealth tax take. Musk’s 2020 personal relocation to Texas is the test case opponents most often cite. Tesla’s primary vehicle manufacturing remains in Fremont. SpaceX’s main engineering and production facility remains in Hawthorne. Principals depart. Enterprises remain. The agglomeration economies of California’s labor pool, research universities, and supplier networks are immobile in a way Musk himself is not. The example confirms the math; it does not refute it.

The related objection, where will the government get its next money fix once the rich are tapped out? reveals the underlying confusion. California’s billionaire households watched their combined wealth grow 144 percent in three years, with annual gains in the thirty-percent range. Far from approaching exhaustion, the wealth base under consideration is accumulating faster than any tax instrument the public sector currently possesses can capture. The one-time five-percent levy is less than two months of recent Brin-and-Page appreciation. The premise of the question, that taxing wealth depletes a fixed pile, is precisely inverted. The pile compounds at better than thirty percent annually while the public revenue claim against it sits at twenty-six basis points.

The Growth Question: Answered

The objection that taxing wealth will collapse economic growth has the structural form of an empirical claim about which the empirical record has already returned a verdict. The verdict is the opposite of what the apologists assert.

The Laffer Curve, invoked in the present debate as if it were settled science, is in its conceptual form a tautology: at a tax rate of zero, revenue is zero; at a tax rate of one hundred percent, revenue is zero, because production stops or moves underground; therefore an intermediate rate maximizes revenue. The proposition is so general that it tells you nothing about whether any particular real-world tax rate sits above the inflection point or below it. That is an empirical question. Peter Diamond and Emmanuel Saez published the modern survey of this question in the Journal of Economic Perspectives in 2011: The Case for a Progressive Tax: From Basic Research to Policy Recommendations. Their conclusion, derived from the actual behavioral elasticities economists have measured rather than from napkin drawings, is that the revenue-maximizing top marginal federal income tax rate in the United States is approximately seventy-three percent. Piketty, Saez, and Stantcheva (2014), working with broader datasets and a wider range of behavioral assumptions about top-end bargaining responses, arrived at an upper range as high as eighty-three percent. The elasticity inputs underlying these calculations are themselves the subject of a substantial literature, surveyed definitively in Saez, Slemrod, and Giertz, “The Elasticity of Taxable Income with Respect to Marginal Tax Rates: A Critical Review,” Journal of Economic Literature 50, no. 1 (2012). Their finding, which the apologists rarely engage, is that long-run elasticities are substantially smaller than the short-run elasticities that supply-side critics often cite, because short-run estimates conflate genuine behavioral responses with timing and avoidance responses that base-broadening can address. Even at the high end of the empirical range, the resulting “optimal” rate is around sixty percent. The current top federal marginal income tax rate in the United States is thirty-seven percent. We are not on the descending portion of the Laffer Curve. We are not approaching it. We are nowhere near it.

The historical record on high tax rates and growth is, against the assertion of the Wall Street Journal editorial page, equally clear. From 1944 to 1963 the United States levied a top marginal federal income tax rate between 91 and 94 percent. The American economy grew at an average annual real rate above four percent during that period. The middle class expanded faster than at any point before or since. Productivity gains were broadly distributed across the workforce. The prewar wealth concentration that Piketty and Saez documented in their long-run series of U.S. top income shares declined sharply. Postwar Western European social democracies operated under tax-and-transfer regimes substantially more aggressive than anything ever attempted in the United States, and produced the longest sustained period of broadly shared prosperity in their respective national histories. The genuine experiments in the predicted opposite direction: Reagan in the 1980s, Brownback in Kansas, Trump in 2017, were the ones in which the promised growth surge most spectacularly failed to materialize. The Congressional Research Service published the post-mortem on the 2017 cuts in 2019. Thomas Hungerford’s earlier CRS study, Taxes and the Economy: An Economic Analysis of the Top Tax Rates Since 1945 (2012), tested the relationship between top marginal rates and GDP growth across the entire postwar period and found no statistically significant correlation between the two. The prediction that high top rates kill growth has been tested across the full available historical record. The result has been published. The prediction failed. The IMF, in Ostry, Berg, and Tsangarides (2014), published the cross-national post-mortem on redistribution and growth, with consistent findings. The historical claim opponents make is the inverse of what the historical record contains.

The related objections, the rich create the jobs, requires its own brief treatment, because it has acquired in American discourse the quality of an unexamined truism. The empirical literature is again contrary to the claim. The bulk of net new job creation in the United States is produced by small and medium-sized enterprises, by public-sector employment, by entrepreneurial activity that is not concentrated in the billionaire class, and by demand-side factors that depend on the wage incomes of ordinary workers far more than on the asset holdings of the wealthy. The wealth of the AI capital class has, in measurable terms, eliminated employment in the sectors where AI has been deployed at scale, a dynamic the predecessor essay They Work for the Bots analyzed in detail. The job-creation defense of billionaire wealth concentration was always a transposition from the small-business case to the rentier case, and it survives only because the conflation has not been disaggregated for public consumption.

The Anecdote Problem: Musk, NOLs, and the Architecture of Misdirection

A particular family of objections deserves separate treatment because of the rhetorical role it plays in the debate. The arguments take the form of a single anecdote produced as if it were the system, or of a corporate tax provision invoked as if it were the personal tax architecture. They are, in each case, sleight of hand.

The Elon Musk $11 billion figure has acquired the status on the American right that single anecdotes always acquire in argumentation that has run out of structural defenses. In 2021 Musk exercised roughly $23 billion worth of Tesla stock options that he had received in 2012 and that were set to expire that year. The exercise of expiring options is a tax-triggering event by statute. The alternative was forfeiture. The $11 billion he paid was the consequence of an unavoidable one-time realization, not the operation of a system that taxes him on an ongoing basis. The Secret IRS Files documented what the system does in the years that are not 2021. The years before and after confirm what the system normally does: it does not tax the accumulation of capital in stock, and for the largest holders, “normally” approaches “ever.” Treating one forced realization as evidence of routine system fairness is the apologetic equivalent of citing a lightning strike as evidence of common electrical phenomena.

The Net Operating Loss carryforward defense, increasingly invoked by sophisticated apologists, is a category error. NOLs are a corporate income-tax provision. They allow firms to apply losses from one fiscal year against profits in subsequent years for the purpose of calculating corporate income tax liability. The California measure does not tax corporate income. It taxes the personal wealth of individuals. The two are structurally different instruments operating on different bases. Whatever NOLs SpaceX or Tesla carries on its books has no contact with the question of how much of Musk’s personal stock wealth becomes subject to state taxation. The argument would be roughly as relevant if it concerned depreciation schedules for farm equipment. It is, in any case, no longer accurate as a description of “every business in America.” The Tax Cuts and Jobs Act of 2017 eliminated the twenty-year carryforward limit and capped use at 80 percent of taxable income; the 2020 CARES Act temporarily restored the carryback provision and lifted the cap entirely for losses incurred in 2018, 2019, and 2020. The post-2017 NOL regime is more generous to capital-intensive corporations than at any prior point in the provision’s history.

The same family of misdirections includes the California specific observation that Tom Steyer is also a billionaire, who has also spent enormous sums on political activity (roughly $134 million on his California gubernatorial campaign as of April 2026), and is supporting the November measure. The observation is accurate. It also has no bearing on the empirical case for the measure. The case rests on the structural fact that two hundred and fifty California households pay 0.07 percent at the top of the wealth distribution in state income tax, not on the moral character of any billionaire who happens to support the November initiative. A serious Marxist analysis does not defend Steyer’s spending; it observes that the political system in which a single individual can outspend his nearest Democratic primary opponent by a factor of twenty is precisely the system the wealth tax is, in modest first instance, attempting to constrain. The Steyer campaign argues, by negative example, for the wealth tax.

The pattern across these objections is the same. A single dramatic fact is produced as if it were the system. A corporate provision is invoked as if it were the personal architecture. A left-side billionaire is identified as if his existence neutralized the structural critique. In each case, the apologetic operation is to substitute the anecdote for the analysis. The architecture remains protected by the misdirection. The misdirection is the architecture’s principal mechanism of self-defense.

International Practice: The Framework Exists

The objection that wealth taxes have no precedent or framework in international practice is, like the analogous objection in the reparations debate, simply false. The framework is extensive, well documented, and operating successfully in multiple jurisdictions.

Switzerland has operated wealth taxes at the cantonal level continuously for over a century. Every canton imposes a wealth tax, with rates varying from approximately 0.13 percent in low-tax cantons to roughly 1.0 percent in high-tax cantons. The Swiss wealth tax base includes equities, real estate, business assets, vehicles, and most other forms of personal wealth, with valuation conducted through procedures the cantons have refined across a hundred years of administrative practice. The Brülhart et al. (2022) study cited above derived its mobility elasticities precisely from the Swiss experience, because the Swiss experience provides the largest natural experiment in wealth-tax administration in the developed world. The Swiss economy has not collapsed. The wealthy have not fled. The Swiss banking system has not been destabilized. The administrative challenges that American opponents invoke as fatal have been solved, in operational terms, by the country whose financial sector is most exposed to wealth-tax effects, over a hundred years of continuous practice.

Spain operates a national wealth tax with substantial regional variation, and the Agrawal, Foremny, and Martínez-Toledano (2026) study cited above analyzed Spain’s interregional mobility responses. Norway operates a wealth tax at the national level, with rates currently set at approximately 1 percent on net wealth above NOK 1.7 million. Belgium has operated a securities-account tax that targets concentrated wealth holdings since 2018. Argentina imposed an emergency one-time wealth tax in 2020-2021 that successfully collected approximately $2.4 billion from the country’s wealthiest residents. The European Tax Observatory, founded by Zucman, Joseph Stiglitz, and others, has been developing the institutional architecture for coordinated international wealth taxation, including a proposed global minimum wealth tax on individuals holding over $100 million that would close the cross-border arbitrage opportunities currently available to the very wealthy. The OECD’s global minimum corporate tax (the 15 percent rate adopted by over 140 jurisdictions in 2021) demonstrates the operational feasibility of coordinated international taxation that opponents would have called impossible a decade earlier.

The constitutional objection that arises in the American context, whether a wealth tax violates the direct-tax clause of Article I §9 of the U.S. Constitution, is a question that the California measure does not need to address, because California is not bound by the federal direct-tax clause. The clause restricts only federal taxation. State wealth taxes face no analogous federal constitutional barrier and would be evaluated, in any litigation, against the substantially more permissive state-constitutional standards. The federal constitutional question, when it eventually arises in the context of a national wealth tax, has been addressed at length by Saez and Zucman in The Triumph of Injustice (2019), and by constitutional-law scholars including Bruce Ackerman, Erik Jensen, and Calvin Johnson. Their conclusion, with which the present author concurs, is that a federal wealth tax can be structured to satisfy the constitutional requirements, that the precedents most often cited against such a tax (especially Pollock v. Farmers’ Loan & Trust of 1895) have been substantially undermined by the Sixteenth Amendment and by subsequent jurisprudence, and that the constitutional question is a litigatable one rather than a categorical bar. The legal architecture is workable. What is required is the political mobilization to make it operative.

The administrative complexity objection, that wealth cannot, in operational practice, be accurately valued and taxed, is contradicted by every jurisdiction that successfully operates a wealth tax. Switzerland values stocks at year-end market prices, real estate at official tax-cadastre valuations, business assets at book value or fair-market value depending on type, with safe-harbor procedures for valuation disputes. The IRS already operates an estate tax that requires valuation of essentially identical asset classes at moments of generational transfer. Forbes and Bloomberg publish billionaire wealth estimates that match independently audited figures to within a few percentage points for publicly traded company holdings, which constitute the great majority of the relevant base. The valuation problem has been solved repeatedly in actual administrative practice. The argument that it cannot be solved in California is the argument that the United States is uniquely incapable of doing what every comparable democracy already does.

Against the Catechism of Earning

The most emotionally resonant objection to a wealth tax, they earned it; their wealth represents the value they created; taking it is theft, reveals a profound confusion about the nature of capital accumulation under contemporary conditions, and the failure to grasp the confusion explains much of the resistance among populations that do not, themselves, benefit materially from the architecture they are defending.

The wealth of the AI capital class was not, in any meaningful analytical sense, “earned” in the way that the architects of American tax discourse use that word. It was assembled through a primitive accumulation operating against a global commons. The Cognitive Commons developed this argument at length. The training data on which the AI models depend was extracted, without consent or compensation, from the entire archive of human cognitive labor. The model weights derived from that training data were enclosed as private property by firms that pioneered the legal-political techniques of enclosure (the fair-use scraping doctrine that Authors Guild v. Google established in 2015) and that, in several documented cases, deliberately ingested pirated corpora after their internal counsel had flagged the legal exposure. The compute infrastructure on which the models are trained was built with substantial public subsidies that the building corporations did not repay through any commensurate return to the public. The fundamental scientific research on which the field rests of neural network architecture, GPU computing, backpropagation, transformer architecture, was conducted in publicly funded universities and government laboratories, with the resulting knowledge appropriated, without royalty, by the private firms that have monetized it.

This is the situation in which Brin, Page, Zuckerberg, and Huang have come to hold their fortunes. Their work has been substantial. Their corporate leadership has been competent. But the proposition that the wealth they hold represents their personal earning, in the way that a small-business owner’s wealth represents the small business owner’s earning, is empirically false. The wealth represents the appropriation of a productive substrate that was, in every important component, socially produced. The case for taxing it rests on the recognition that the substrate from which it was extracted has a legitimate claim against it.

The same argument applies, in different forms, to essentially every component of the American billionaire class. The fortunes built on Walmart, Amazon, and the retail-logistics consolidation depended on the federal interstate system, on container-shipping infrastructure built by public investment, and on the labor of a working class whose wages were systematically suppressed across forty years. The fortunes built on Wall Street depended on the Federal Reserve’s recurrent backstopping of the financial system, on the public-pension capital that flows into private-equity vehicles, and on legal frameworks (limited liability, securities preemption, bankruptcy preference) that are public concessions. The fortunes built on real estate depended on zoning laws, mortgage subsidies, the federal mortgage market, and the systematic exclusion of competing forms of housing provision. The fortunes built on technology depended on DARPA-funded research, the National Science Foundation, the publicly funded university system, the public internet backbone, and the legal regimes that protected the resulting monopolies. Mariana Mazzucato’s The Entrepreneurial State (Anthem Press, 2013) documents the broader pattern with the precision the case requires: essentially every transformative innovation of the postwar period, the internet, GPS, semiconductors, the touchscreen, voice recognition, the cellular network, the human genome project, mRNA vaccines, the basic-research architecture underlying contemporary AI, came from publicly funded research, with the productive applications subsequently enclosed by private capital under legal arrangements the public underwrote without seeking a return. In each case, the wealth “earned” is wealth extracted from socially produced conditions of possibility. The wealth tax is the modest recognition that the conditions of possibility deserve a return.

David Roediger’s analysis of the wages of whiteness and Aníbal Quijano’s apparatus of the coloniality of power clarify the rhetorical operation by which the they earned it claim sustains itself. The system that produced the contemporary wealth distribution has, simultaneously, produced an ideological apparatus that naturalizes the resulting hierarchy as a reflection of merit. The dispossessed are told that the architecture is fair, that the ownership class deserves what it holds, that any redistribution would be theft. The naturalization is the ideological cement of the configuration. The wealth tax debate is one of the sites at which the naturalization is currently being contested. Whether the contest succeeds is one of the political questions the next decade will answer.

Against the Liberal Containment

The deeper objection that needs to be addressed is the one offered by the AI industry’s own liberal apologists, who have begun to concede that something must be done about the concentration of AI-era wealth and who have produced, in the OpenAI white paper of October 2025 and the Anthropic Institute’s affiliated documents, a redistributive package calibrated to absorb the political pressure without altering the underlying property relations. They Work for the Bots analyzed this containment strategy at length; I summarize the structural analysis here, both to relieve readers of the trip back and to direct readers who have not made the trip toward it.

The liberal containment package, in its current form, proposes a thirty-two-hour workweek, higher corporate and capital-gains taxation, and a public wealth fund that would provide all citizens an equity stake in AI companies. The package is more substantial than what the Democratic Party has been willing to articulate in any of the last five presidential cycles, and the figures who have developed it (Sam Altman, the OpenAI policy team, Chris Lehane, Jack Clark at the Anthropic Institute) deserve recognition for the seriousness with which they have engaged the question. The package is also, structurally, the response that the AI industry itself prefers, because it preserves the industry’s ownership of the productive apparatus while neutralizing the political pressure that mass dispossession would otherwise produce. Each component of the package, in its structural form, accepts that the AI industry will continue to be owned and governed by the people who currently own and govern it, and proposes only that some larger fraction of the proceeds be transferred, on terms set by the existing owners, to the population the industry is in the process of dispossessing.

This is the bourgeois response. It is the redistributive response that does not alter the property relations. It produces, when fully implemented, a welfare-feudal regime: a small ownership class controlling the productive apparatus, a state apparatus extracting and redistributing a portion of the surplus as conditional welfare, a dispossessed population sustained at subsistence on the redistributed surplus, the entire configuration politically stabilized by the redistributive parameters that can be tightened or loosened as bargaining power shifts. They Work for the Bots developed the analytical case against this configuration in full. Note here only that the wealth tax under discussion in the present essay sits outside the liberal containment package. It is a structural instrument operating on the ownership class’s accumulated assets, calibrated against a specific fiscal need, and capable of being articulated within either the liberal redistributive frame or the democratic-socialist structural frame that this sequence of essays has been building.

The political question, when the November measure is on the ballot, is which frame the measure will be understood within. If the wealth tax is articulated as the first move in the liberal containment package, the modest redistributive measure that allows the underlying ownership architecture to continue, it will be voted on accordingly, and its passage will be incorporated into the AI industry’s broader political-messaging operation. If it is articulated as the first move in a structural reckoning, the first political instrument anywhere in the world that has touched the accumulated wealth of the AI capital class, and the prototype for a larger program that will alter the ownership relations themselves, its passage becomes the beginning of a different politics. The vote is the same. The political meaning of the vote depends on the analytical frame within which it is cast.

The Wealth Tax as First Element of a Larger Program

The democratic-socialist program adequate to the AI moment was developed in detail in The Cognitive Commons. Six elements were articulated: a public data trust holding the cognitive commons in democratic ownership; a National Compute Authority modeled on the Tennessee Valley Authority, with public ownership of large-scale computing infrastructure; public model weights, holding the means of cognitive production in common; codetermination on deployment, modeled on German Mitbestimmung, giving workers binding say over how AI is introduced into their fields; a comprehensive social wage decommodifying the conditions of life, on the Mamdani municipal model scaled to the national economy; and a public utility framing of the AI sector as essential infrastructure subject to democratic governance. The wealth tax is the seventh element, the fiscal instrument that recovers a portion of the surplus that the present ownership architecture has accumulated and that, deployed correctly, becomes the funding base for the structural transformation the other six elements describe.

The California measure cannot be all of this on its own. It is too small, too geographically restricted, too modest in its rate. What it can do, if it passes, is establish three operational demonstrations that the larger program requires. First, that wealth taxation at meaningful scale can be administered competently by an American jurisdiction. Second, that the predicted mobility responses will be small, that they will be offset by replacement formation in the resident billionaire population, and that the projected revenue will materialize. Third, that the political coalition required to pass such a measure, a coalition of organized labor, health-care advocacy, public-services constituencies, and democratic-socialist political organization, can be assembled and held together against the full opposition of the local capital class.

The federal level is, in operational terms, where the wealth tax must ultimately be settled. Saez and Zucman have argued for years, most fully in The Triumph of Injustice, that a federal wealth tax of two to three percent annually on fortunes above fifty million dollars would raise revenue on the order of three percent of GDP, enough to fund Medicare for All, free public-college tuition, the climate transition, the social wage architecture that The Cognitive Commons described, simultaneously. The federal advantage is that mobility issues largely disappear. The United States imposes worldwide taxation on its citizens. Renouncing citizenship triggers a substantial exit tax. The federal base is much harder to flee than the California base. The California measure produces the empirical and political precedent that subsequent federal proposals can build on. The federal proposals produce the funding base for the structural program. The structural program transforms the ownership architecture. This is the trajectory. The November vote is its first move.

Other state movements are already underway. Washington State has been developing wealth-tax proposals targeting the state’s substantial billionaire population (Bezos, Ballmer, others). Massachusetts has been examining similar instruments aimed at the financial sector. New York’s tax architecture has been the subject of substantial reform proposals since the Mamdani victory. The international coordination work, conducted through the European Tax Observatory and the G20 working groups Zucman has anchored, has produced the architecture for a globally coordinated minimum wealth tax on individuals above $100 million, which would eliminate the cross-border arbitrage that opponents invoke as the structural limit on national wealth taxation. The political work required to operationalize this tax structure is substantial. The technical and analytical work has been done.

This is, in operational terms, the moment in which the democratic-socialist program adequate to the AI moment becomes politically thinkable, because the AI moment has produced a wealth distribution sufficiently extreme to render the existing arrangements indefensible to populations that, twenty years ago, would have accepted them as natural. The young, the indebted, the housing-cost-burdened, the wage-stagnated, the dispossessed professional layer that They Work for the Bots analyzed, this constituency is now in the act of becoming aware of its dispossession at scale, and is, in the same act, becoming receptive to political analysis of a structural type that has not been politically thinkable in the United States since the 1940s. The Mamdani victory in November 2025 was the first major electoral expression of this receptivity. The California wealth tax in November 2026 will be the second. The 2028 Democratic primary will be the third. Each is a node in the same political trajectory, and each succeeds or fails on the same political question: whether the analytical clarity required to articulate the structural alternative can be assembled in time to compete with the redistributive containment that the AI industry is funding to absorb the discontent.

The Politics of November

The objectors are correct that the wealth tax would be transformative. They are wrong that this is an argument against it. A society that acknowledged the AI capital class’s debt to the cognitive commons that produced its wealth, that undertook the democratic work of designing a remedy proportional to the appropriation, that mobilized its fiscal apparatus to address the consequences of forty years of upward concentration, such a society would be a different society than the one that currently exists. It would be a society in which the relationship between capital and the social conditions of capital’s reproduction was visible, in which the cognitive enclosure that has been conducted in plain view of every relevant democratic institution was the object of explicit political contestation, in which the democratic process had demonstrated its capacity to address structural injustice rather than merely perpetuating it.

Walter Benjamin wrote that every document of civilization is simultaneously a document of barbarism. The wealth distribution of the AI moment is the contemporary paradigm. The four wealthiest Californians hold collectively more than three-quarters of a trillion dollars, accumulated through their corporate ownership of the firms that performed the largest single act of primitive accumulation in the recorded history of capitalism, while the state in which they reside has been forced to cut Medicaid coverage to the poor because the federal government withdrew its share. The architecture that produced this configuration is a specific set of legal-political arrangements: arrangements that have permitted the cognitive enclosure to occur, that have exempted its proceeds from meaningful taxation, that have allowed the resulting wealth to compound at rates without precedent in the modern record. Calling the resulting distribution natural, or universal, or the just reward of merit, is the apologetic vocabulary the architecture has produced over fifty years to defend itself. The November measure is the first political instrument that would, modestly and at the margins, begin to address what the apologetics protect.

Fanon wrote that decolonization is always a violent phenomenon, meaning that it disrupts the entire order of the colonial world, from its economic foundations to its most intimate psychic architecture. The redistribution of wealth, the reorganization of the property relations governing the AI economy, the transformation of the psychic and social conditions that the cognitive enclosure has produced; these are acts of structural disruption against the existing order, and they will be experienced as such by those who benefit from it. Brin’s fifty-seven million dollars in opposition spending is what such resistance looks like in operational practice. The arguments examined in this essay, about mobility, about growth, about earning, about the Laffer Curve, about Sowell, are the rhetorical surface of the resistance. The Sowell quotations, in particular, are the closing flourish that the present apologetic apparatus has refined into a kind of reflex. They function as the closing argument in a case whose evidence has run out somewhere earlier. The villains Sowell warned about exist. They are the ones quoting him. The heroes he idealized exist as well: they have names like Brin and Page and Zuckerberg and Huang, they live in California, and they are spending what is required to keep the architecture they have benefited from intact.

The question was never whether the wealth tax is possible. The empirical case is overwhelming. The administrative frameworks exist. The international precedents are established. The constitutional architecture is workable. The scholarly consensus on the nature and scale of the wealth concentration is overwhelming. The political constituency, in the dispossessed populations the AI decade is producing, exists at scale. The question has always been whether the American political order will permit the confrontation with its own architecture that the wealth tax requires. The answer to that question is political, and it will be determined by political struggle, by the organized demand of the constituencies the AI moment is producing, supported by a multiracial coalition that understands the present configuration as a system whose architecture must be altered rather than merely managed.

The California voters who cast ballots in November will not, on their own, alter the architecture. They will perform, however, the first move within it. They will demonstrate, for the first time anywhere in the world, that a democratic jurisdiction can place a meaningful claim on the accumulated wealth of the AI capital class, and that the class’s stated threats, the flight scenarios, the growth-collapse predictions, the Sowell quotations, can be answered with evidence in the moment of voting. The demonstration will be used. The federal proposals will follow. The international coordination work will accelerate. The Mamdani-model municipal experiments will expand. The 2028 Democratic primary will, if the analytical clarity has been established in time, be conducted within a vocabulary that names the structural question rather than evading it.

What is described on the ballot, in the formal text, as a one-time five-percent levy on the personal wealth of resident billionaires, applied across five years, used to replace withdrawn Medicaid funding. That description is accurate but incomplete. The fuller description is the one this essay has attempted to develop. The measure is the first political instrument adequate to the AI moment. Its passage is the beginning of a politics that takes the present situation seriously. Its failure is the consolidation of an architecture that may, in the course of the next decade, become unalterable by democratic means.

Vote yes. The measure is insufficient; vote yes despite that. The messengers are complicated; vote yes despite that. Vote yes because the alternative is to watch the AI decade compress political and economic power past the point at which democratic correction remains available. Vote yes because fifty-seven million dollars in opposition spending tells you exactly what is at stake. Vote yes because 0.07 percent is not a tax rate; it is the absence of one. Vote yes because the architecture that produced the figure was built over forty years and will be unbuilt only by deliberate political work that begins, somewhere, with a vote no one will be able to misinterpret.

The work begins with a yes. The reckoning that follows will require everything else the prior essays have described. The vote is the smallest move and the necessary one. November.

Bibliography

Primary Theoretical Texts

Benjamin, Walter. “Theses on the Philosophy of History.” In *Illuminations: Essays and Reflections*, edited by Hannah Arendt and translated by Harry Zohn, 253–64. New York: Schocken Books, 1968. Originally composed 1940.

Fanon, Frantz. *The Wretched of the Earth*. Translated by Richard Philcox. New York: Grove Press, 2004. Originally published 1961.

Marx, Karl. *Capital: A Critique of Political Economy, Volume I*. Translated by Ben Fowkes. London: Penguin Classics, 1990. Originally published 1867.

Tax Policy, Wealth Distribution, and Empirical Research

Agrawal, David R., Dirk Foremny, and Clara Martínez-Toledano. “Wealth Tax Mobility and Tax Coordination.” Working paper, 2026.

Brülhart, Marius, Jonathan Gruber, Matthias Krapf, and Kurt Schmidheiny. “Behavioral Responses to Wealth Taxes: Evidence from Switzerland.” *American Economic Journal: Economic Policy* 14, no. 4 (2022): 111–50.

Diamond, Peter, and Emmanuel Saez. “The Case for a Progressive Tax: From Basic Research to Policy Recommendations.” *Journal of Economic Perspectives* 25, no. 4 (Fall 2011): 165–90.

McCaffery, Edward J. “Taxing Wealth Seriously.” *Tax Law Review* 70 (2017): 305–87.

Ostry, Jonathan D., Andrew Berg, and Charalambos G. Tsangarides. *Redistribution, Inequality, and Growth*. IMF Staff Discussion Note SDN/14/02. Washington, DC: International Monetary Fund, 2014.

Piketty, Thomas. *Capital in the Twenty-First Century*. Translated by Arthur Goldhammer. Cambridge, MA: Belknap Press of Harvard University Press, 2014.

Piketty, Thomas, Emmanuel Saez, and Stefanie Stantcheva. “Optimal Taxation of Top Labor Incomes: A Tale of Three Elasticities.” *American Economic Journal: Economic Policy* 6, no. 1 (February 2014): 230–71.

Rosenthal, Steven M., and Theo Burke. *Who Owns U.S. Stock? Foreigners and the Rich*. Washington, DC: Urban-Brookings Tax Policy Center, October 20, 2020.

Saez, Emmanuel, and Gabriel Zucman. *The Triumph of Injustice: How the Rich Dodge Taxes and How to Make Them Pay*. New York: W. W. Norton, 2019.

Saez, Emmanuel, and Gabriel Zucman. “The Case for California’s Billionaire Wealth Tax.” *New York Times*, May 26, 2026. https://www.nytimes.com/interactive/2026/05/26/opinion/wealth-tax-california-billionaire.html.

Saez, Emmanuel, and Gabriel Zucman. *Taxing California Billionaires*. Berkeley Working Paper. Berkeley: University of California, Berkeley, 2026. https://eml.berkeley.edu/~saez/BSZ26CAbillionaires.pdf.

Saez, Emmanuel, Joel Slemrod, and Seth H. Giertz. “The Elasticity of Taxable Income with Respect to Marginal Tax Rates: A Critical Review.” *Journal of Economic Literature* 50, no. 1 (March 2012): 3–50.

World Inequality Lab. *World Inequality Report 2022*. Paris: World Inequality Lab, 2021. https://wir2022.wid.world/.

Young, Cristóbal, Charles Varner, Ithai Z. Lurie, and Richard Prisinzano. “Millionaire Migration and Taxation of the Elite: Evidence from Administrative Data.” *American Sociological Review* 81, no. 3 (June 2016): 421–46.

Zucman, Gabriel, ed. *Global Tax Evasion Report 2024*. Paris: EU Tax Observatory, October 2023. https://www.taxobservatory.eu/.

Contemporary Marxist and Critical Apparatus

Chibber, Vivek. *The Class Matrix: Social Theory after the Cultural Turn*. Cambridge, MA: Harvard University Press, 2022.

Fraser, Nancy. *Cannibal Capitalism: How Our System Is Devouring Democracy, Care, and the Planet — and What We Can Do About It*. London: Verso, 2022.

Harvey, David. *The New Imperialism*. Oxford: Oxford University Press, 2003.

Mazzucato, Mariana. *The Entrepreneurial State: Debunking Public vs. Private Sector Myths*. London: Anthem Press, 2013.

Phillips, Leigh, and Michal Rozworski. *The People’s Republic of Walmart: How the World’s Biggest Corporations Are Laying the Foundation for Socialism*. London: Verso, 2019.

Quijano, Aníbal. “Coloniality of Power, Eurocentrism, and Latin America.” *Nepantla: Views from South* 1, no. 3 (2000): 533–80.

Roediger, David R. *The Wages of Whiteness: Race and the Making of the American Working Class*. London: Verso, 1991.

Wright, Erik Olin. *Envisioning Real Utopias*. London: Verso, 2010.

Investigative Journalism and Documentary Sources

Eisinger, Jesse, Jeff Ernsthausen, and Paul Kiel. “The Secret IRS Files: Trove of Never-Before-Seen Records Reveal How the Wealthiest Avoid Income Tax.” *ProPublica*, June 8, 2021. https://www.propublica.org/article/the-secret-irs-files-trove-of-never-before-seen-records-reveal-how-the-wealthiest-avoid-income-tax.

Government and Institutional Documents

Congressional Research Service. *The Economic Effects of the 2017 Tax Revision: Preliminary Observations*. CRS Report R45736. Washington, DC: Congressional Research Service, May 22, 2019.

Hungerford, Thomas L. *Taxes and the Economy: An Economic Analysis of the Top Tax Rates Since 1945*. CRS Report R42729. Washington, DC: Congressional Research Service, September 14, 2012.

Predecessor Essays in This Sequence

Potter, Del. “The Intellectual Mirage: Thomas Sowell and the Poverty of Market Apologetics.” *Del’s Substack*, February 27, 2026. https://delpotterphd.substack.com/p/the-intellectual-mirage.

Potter, Del. “They Work for the Bots: On the San Francisco Consensus, the Underclass That Was Always There, and the Political Form Adequate to the Moment.” *Del’s Substack*, May 1, 2026. https://delpotterphd.substack.com/p/they-work-for-the-bots.

Potter, Del. “The Cognitive Commons: Everything from Everyone, Everywhere, All the Time.” *Del’s Substack*, May 11, 2026. https://delpotterphd.substack.com/p/the-cognitive-commons.

My response begins by refusing the false choice being offered. Defending Kristof’s reporting requires no defense of the New York Times as an institution. The Times has laundered war propaganda, marginalized Palestinian voices, written Palestinian deaths in passive voice, and treated U.S. imperial assumptions as the air we all breathe. Nicholas Kristof himself is a liberal humanitarian columnist whose politics rarely test the boundaries of acceptable elite opinion. The interesting fact lies elsewhere. When even a figure that mainstream draws this volume of denunciation, the denunciation reveals what ideological work is being defended.

Look first at what Kristof actually reported. The May 11 column, “The Silence That Meets the Rape of Palestinians,” drew on interviews with fourteen Palestinian victims, men and women, with accounts Kristof says he corroborated where possible, in the Times’ formal response to Israel’s threatened libel suit, with family members, lawyers, and other people the victims had confided in. He named on-record sources including journalist Sami al-Sai (who had already given public testimony to the BBC in December 2025 about sexual assault while held in Megiddo prison in March 2024) and activist Issa Amro. He cited the alleged 2024 sexual abuse of a Gazan inmate at Sde Teiman military prison, the case in which the Israeli military indicted five reservist guards before charges were dropped earlier this year. He cited Euro-Med Human Rights Monitor, the UN-accredited Geneva NGO Jacobson labels a Hamas front. He quoted former Israeli Prime Minister Ehud Olmert.

Now look at Jacobson’s actual arguments, one at a time.

The Euro-Med smear is built on a photograph from 2011 in which the organization’s founder Ramy Abdu and its former chair Mazen Kahel appear in a group shot behind Hamas official Ismail Haniyeh, and a separate 2013 panel appearance alongside Osama Hamdan. By this evidentiary standard, every Israeli politician photographed beside a war criminal at the United Nations becomes implicated in that war criminal’s crimes. Guilt by photograph is a technique for vanishing evidence. The procedure has no methodological standing whatsoever, and its selective deployment against organizations documenting state violence underwritten by the United States tells you exactly which frame of credibility is being protected.

The anonymous-source complaint inverts forty years of feminist scholarship on sexual violence and disclosure. Survivors face stigma, retaliation, communal shame, and disbelief in peacetime, in private contexts, in domestic violence cases, in workplace harassment cases, in every context where there is no occupying army or military court or Unlawful Combatants Law stripping the victim of access to a lawyer. Anonymity is the standard practice of trauma-informed journalism, deployed routinely without controversy when the alleged perpetrators are Catholic priests, Hollywood producers, or Hamas fighters on October 7. Treating anonymity as proof of falsehood specifically when the alleged perpetrators are Israeli soldiers reveals the credibility criterion in operation. Silvia Federici’s analysis of state violence inscribed on women’s bodies and Stuart Hall’s reading of racialized moral panic both describe this dynamic. The evidentiary threshold being demanded of Palestinian victims is one they cannot conceivably meet, which is precisely the function of the threshold.

The on-record sources Jacobson does engage are dismissed for political opinions and account-evolution. Sami al-Sai praised the October 7 attacks. Issa Amro’s account shifted between his 2024 Washington Post interview and the Kristof column. Sexual assault testimony commonly evolves with disclosure conditions; prosecutors and trauma clinicians have understood this for forty years. As for political opinions, applying a purity test to Palestinian witnesses and granting Israeli state denial face value in the same breath simply displays the credibility hierarchy in plain text.

The Olmert recanting deserves investigation against the original recording. It collapses none of the column’s remaining reporting, and Jacobson does not attempt to argue that it does.

The “systematic but unmeasurable” objection mistakes a feature of mass violence for a contradiction. Kristof wrote that the assaults are systematic and that their exact frequency is impossible to know. Both can be true at once. Lynching in the American South was systematic and unmeasurable. Femicide in Ciudad Juárez was systematic and unmeasurable. Systems of violence function by destroying the documentary infrastructure that would measure them.

The timing complaint is self-defeating. If Palestinian allegations can never be aired without being read as deflection from Israeli ones, the publication window is closed forever by definition. The Times has stated on the record that it had no advance knowledge of the Civil Commission report’s release date. The framework Jacobson uses cannot be tested at all: any publication date for any Palestinian allegation will fall in suspicious proximity to some Israeli counter-event somewhere.

The ad hominem about Kristof’s failed 2022 Oregon gubernatorial run reverses on the author. Jacobson founded No Labels, the centrist dark-money outfit attempting to engineer third-party spoiler candidacies against the U.S. left. Apply Jacobson’s standard to Jacobson, and the column collapses on its own logic.

This is the analytical surface. The deeper question is structural. What larger evidentiary record would Jacobson like collapsed along with the Kristof column?

B’Tselem’s August 2024 report Welcome to Hell, based on testimonies from fifty-five released Palestinian detainees, documents the systematic conversion of more than a dozen Israeli facilities into what the organization calls a network of torture camps. The January 2026 follow-up Living Hell adds testimonies from twenty-one Palestinians released in recent months and confirms continued and more extensive systematic abuse, including sexual violence. UN High Commissioner for Human Rights Volker Türk’s office published a July 2024 report finding cage-like conditions, prolonged stripping and blindfolding, electrocution, dog attacks, waterboarding, and sexual and gender-based violence against Palestinian detainees. UN Special Rapporteurs Francesca Albanese, Tlaleng Mofokeng, and Reem Alsalem warned in August 2024 that Sde Teiman represents “the tip of the iceberg”. The UN Independent International Commission of Inquiry concluded in March 2025 that sexual violence against Palestinian detainees was committed either under explicit orders or with implicit encouragement by Israel’s top civilian and military leadership, with commission member Chris Sidoti adding that the violence is now so widespread it can only be considered systematic. In response to the current controversy, OHCHR spokesperson Thameen Al-Kheetan stated that torture and ill-treatment, including sexual and gender-based violence, are systematically perpetrated against Palestinian prisoners under Israeli custody, with numerous cases of rape involving children, and confirmed at least 90 Palestinian deaths in Israeli custody since October 7, 2023.

The architecture of impunity is documented in equally specific terms. The Public Committee Against Torture in Israel records that only one Israeli soldier has so far been sentenced for torturing a Palestinian detainee since October 2023. Yesh Din finds that 93.6 percent of investigations into ideologically motivated offenses by Israelis against Palestinians in the occupied West Bank since 2005 have ended without indictment, a record the organization describes as deliberate policy. In January 2025, Israel denied UN Special Representative on Sexual Violence in Conflict Pramila Patten access to Israeli detention facilities, and according to Haaretz blocked her parallel investigation into Hamas’s October 7 sexual crimes specifically to avoid the corollary scrutiny of IDF detention abuse. The Military Advocate General dropped all charges against the five Sde Teiman defendants on March 12, 2026, on stated grounds that the video didn’t show abuse violent enough to merit a criminal conviction, that it had been improperly leaked, and that the victim’s release back to Gaza created an “absence of certainty” he could testify. Sari Bashi of the Public Committee Against Torture in Israel summarized the moment: “Israel’s military advocate general just gave his soldiers license to rape, so long as the victim is Palestinian.” Within weeks the soldiers had returned to reserve service, with some already deployed to combat roles.

This is the terrain Jacobson would prefer collapsed into a Hamas-front narrative about anonymous Palestinians lying.

Antonio Gramsci’s account of hegemony describes exactly how this operation works: the manufacture of common-sense limits on what counts as evidence, on whose pain registers as injury, on which institutional voices command the presumption of truth. Nicos Poulantzas’s analysis of the capitalist state describes the legal apparatus as a crystallization of class and racial domination, and the Israeli military’s indict-and-drop ritual is that crystallization made visible. Stuart Hall’s work on race, crime, and moral panic provides the template for understanding how Palestinian testimony gets reframed in advance as fabrication, propaganda, or “blood libel,” with the last term doing especially heavy ideological work by inverting the historical relation of accuser and accused. Mark Fisher’s capitalist realism captures the experiential effect: a sense that no alternative framework of credibility is even thinkable, that the parameters Jacobson invokes are simply the limits of reality itself.

Detention under this configuration is a technology of domination. Prisons, checkpoints, administrative detention, military courts, the Unlawful Combatants Law, surveillance, forced displacement, siege: components of a single colonial system for breaking a people’s political will. Sexual humiliation functions within that system as a tool of discipline. It communicates to the occupied population that the state holds power over land, over movement, and over the body itself. The destroyed bodies are the message.

What does a response look like beyond polemic? Three concrete commitments come into focus.

Build independent media infrastructure that does not require liberal humanitarian permission to report Palestinian testimony. The architecture of disbelief operates by funneling all evidence through institutional gatekeepers Jacobson and her allies can pressure. Bypass the funnel. Support Mondoweiss, Drop Site, +972 Magazine, Local Call, Mada Masr, and the Palestinian journalists dying at unprecedented rates to document this from inside.

Treat international accountability mechanisms as serious organizing terrain. The International Court of Justice’s July 2024 advisory opinion on the unlawfulness of the occupation, the International Criminal Court’s outstanding warrants against Netanyahu and Gallant, the UN sexual violence in conflict blacklist Israel is actively maneuvering to stay off: each is a pressure point, each is under coordinated U.S. attack, and defending these mechanisms domestically is concrete socialist-internationalist work the U.S. left has so far performed unforgivably badly.

We should describe the political economy without flinching. The United States arms Israel, shields it diplomatically, and underwrites the architecture of impunity that makes Sde Teiman possible. Quinn Slobodian’s account of how globalist liberalism builds insulated zones for capital and state violence describes the apparatus precisely. The arms manufacturers booking record orders against Gaza’s destruction, the universities accepting their donations, the media outlets dependent on their advertising, the centrist political infrastructure that hires Nancy Jacobson to write what she writes: this is the material base of the disbelief being manufactured. Cut that base, and the architecture begins to fall.

Jacobson’s piece is therefore misnamed. The scandal lies upstream of Kristof’s column, in the media culture that required extraordinary courage, white liberal cover, outside pressure, and elite permission for such allegations to enter mainstream discussion at all. The real collapse of journalistic standards happened years earlier, when Western media built an apparatus in which Israeli state sources receive presumptive credibility and Palestinian victims receive presumptive suspicion. Kristof’s column is a small breach in that wall. The denunciation from RealClearPolitics is the wall trying to repair itself.

Bibliography

Primary sources engaged

Jacobson, Nancy. “Nicholas Kristof and the Collapse of Journalistic Standards.” RealClearPolitics, May 16, 2026. https://www.realclearpolitics.com/articles/2026/05/16/media_malpractice_at_the_new_york_times_154129.html

Kristof, Nicholas. “The Silence That Meets the Rape of Palestinians.” The New York Times, May 11, 2026. https://www.nytimes.com/2026/05/11/opinion/israel-palestinians-sexual-violence.html

Human rights organization reports

B’Tselem. Welcome to Hell: The Israeli Prison System as a Network of Torture Camps. August 2024. https://www.btselem.org/publications/202408_welcome_to_hell

B’Tselem. Living Hell: The Israeli Prison System as a Network of Torture Camps. January 2026. https://www.btselem.org/publications/202601_living_hell

B’Tselem. “Sde Teiman is only the tip of the iceberg: Israeli detention centers turned into a network of torture camps for Palestinians.” Press release, August 5, 2024. https://www.btselem.org/press_releases/20240805_welcome_to_hell

B’Tselem. “Living Hell: Sexual violence, starvation and systematic abuse – at least 84 Palestinian prisoners have died in Israeli torture camps.” Press release, January 20, 2026. https://www.btselem.org/press_releases/20260120_living_hell

Euro-Mediterranean Human Rights Monitor. Another Genocide Behind Walls: Sexual Violence in Israeli Prisons and Detention Centres and Engineered Impunity (October 2023 – October 2025). April 2026. https://euromedmonitor.org/en/article/7023/

Euro-Mediterranean Human Rights Monitor. “Israel must be added to UN’s sexual violence blacklist.” January 10, 2025. https://euromedmonitor.org/en/article/6591/

Amnesty International. “Israel/OPT: Decision to drop charges against soldiers accused of abuse of Palestinian detainee disgraceful.” Statement by Erika Guevara Rosas, March 14, 2026. https://www.amnesty.org/en/latest/news/2026/03/israel-opt-decision-to-drop-charges-against-soldiers-accused-of-abuse-of-palestinian-detainees-disgraceful/

Public Committee Against Torture in Israel, Parents Against Child Detention, HaMoked, Adalah, and Physicians for Human Rights–Israel. Joint submission to the United Nations Committee against Torture, October 2025.

Yesh Din. Statistical reports on indictment rates for ideologically motivated offenses by Israelis against Palestinians in the occupied West Bank, 2005–2025. (Cited in multiple secondary sources; 93.6 percent non-indictment figure.)

United Nations sources

UN Office of the High Commissioner for Human Rights. “UN report: Palestinian detainees held arbitrarily and secretly, subjected to torture and mistreatment.” Press release, July 31, 2024. https://www.ohchr.org/en/press-releases/2024/07/un-report-palestinian-detainees-held-arbitrarily-and-secretly-subjected

UN Office of the High Commissioner for Human Rights. “Israel’s escalating use of torture against Palestinians in custody a preventable crime against humanity: UN experts.” Statement by Francesca Albanese, Tlaleng Mofokeng, Reem Alsalem, George Katrougalos, and the Working Group on Discrimination Against Women and Girls, August 5, 2024. https://www.ohchr.org/en/press-releases/2024/08/israels-escalating-use-torture-against-palestinians-custody-preventable

UN Independent International Commission of Inquiry on the Occupied Palestinian Territory, including East Jerusalem, and Israel. Report on sexual and gender-based violence, submitted to the Human Rights Council, March 2025.

Al-Kheetan, Thameen (OHCHR spokesperson). Statement to Anadolu Agency on systematic sexual and gender-based violence against Palestinian prisoners in Israeli custody, May 17, 2026. Reported in Common Dreams.

UN News. “Israel: Amid reports of torture inside jails, rights experts call for end to impunity.” August 5, 2024. https://news.un.org/en/story/2024/08/1152831

International Court of Justice. Advisory Opinion on the Legal Consequences arising from the Policies and Practices of Israel in the Occupied Palestinian Territory, including East Jerusalem. July 19, 2024.

International Criminal Court. Arrest warrants against Benjamin Netanyahu and Yoav Gallant, November 21, 2024.

News reporting on the Kristof controversy

Hilkin, Andrew. “NYT column alleging Israeli abuse of Palestinian prisoners sparks debate among Jews.” The Times of Israel / JTA, May 13, 2026. https://www.timesofisrael.com/nyt-column-alleging-israeli-abuse-of-palestinian-prisoners-sparks-debate-among-jews/

Magid, Jacob. “Israel to sue NY Times over op-ed alleging widespread rape of Palestinian prisoners.” The Times of Israel, May 14, 2026. https://www.timesofisrael.com/israel-to-sue-ny-times-over-opinion-alleging-widespread-rape-of-palestinian-prisoners/

The Times of Israel. “’Blood libel’: Israel rejects NYT column alleging widespread rape of Palestinian inmates.” May 12, 2026. https://www.timesofisrael.com/blood-libel-israel-rejects-nyt-column-alleging-widespread-rape-of-palestinian-inmates/

Kampeas, Ron. “Israeli officials, AJC slam Nick Kristof’s NYT column as modern-day ‘blood libel’.” Jewish Insider, May 13, 2026. https://jewishinsider.com/2026/05/nick-kristof-new-york-times-israel-palestinian-prisoners-sexual-abuse/

The Jerusalem Post. “NYT defends Nicholas Kristof column after Israel calls it ‘blood libel’.” May 14, 2026. https://www.jpost.com/international/article-896131

Conley, Julia. “In Wake of NYT Uproar, UN Rights Office Calls for Independent Probe of Israel’s Sexual Abuse of Detainees.” Common Dreams, May 18, 2026. https://www.commondreams.org/news/un-human-rights-israel-prison-rape-investigation

News reporting on Sde Teiman, the dropped charges, and detention conditions

Al Jazeera. “Israel drops charges against soldiers in Palestinian detainee abuse case.” March 12, 2026. https://www.aljazeera.com/news/2026/3/12/israel-drops-charges-on-soldiers-who-allegedly-raped-palestinian-detainee

Associated Press / CBC News. “Israeli military drops charges against soldiers accused of sexual, physical abuse of Palestinian detainee.” March 12, 2026. https://www.cbc.ca/news/world/israel-drops-charges-detainee-abuse-9.7126234

NBC News. “Israeli military drops charges against soldiers accused of sexually assaulting Palestinian detainee.” March 12, 2026. https://www.nbcnews.com/world/israel/israel-soldiers-prisoner-sexual-abuse-video-charges-dropped-rcna263165

The Washington Post. “Israeli military drops charges against soldiers accused of sexually assaulting Palestinian detainee.” March 12, 2026.

Al Jazeera. “Israeli army says soldiers accused of sexual assault to return to duty.” April 16, 2026. https://www.aljazeera.com/news/2026/4/16/israeli-army-says-soldiers-accused-of-abusing-palestinian-to-return-to-duty

Common Dreams. “Israel Blocks UN Probe of Alleged Hamas Sex Crimes During October 7 Attack.” January 9, 2025. https://www.commondreams.org/news/hamas-oct-7-rape

Anadolu Agency. “UN warns of Israeli officials interference in probe of sexual violence against Palestinian detainees.” September 9, 2024. https://www.aa.com.tr/en/middle-east/un-warns-of-israeli-officials-interference-in-probe-of-sexual-violence-against-palestinian-detainees/3325641

Haaretz. Reporting on Sde Teiman detention conditions, deaths in custody, the leaked video of detainee abuse, and Israel’s blocking of the Patten investigation, 2024–2026. (Cited via secondary sources including Al Jazeera, Common Dreams, and Palestine Chronicle.)

The Independent. Investigation based on indictments, autopsy reports, and detainee testimonies on torture, sexual violence, and rape with sharp objects at Israeli detention facilities, March 2024.

BBC News. Reporting on the testimony of Sami al-Saei regarding sexual assault at Megiddo prison, December 2025.

Theoretical and analytical works

Federici, Silvia. Caliban and the Witch: Women, the Body and Primitive Accumulation. Autonomedia, 2004.

Fisher, Mark. Capitalist Realism: Is There No Alternative? Zero Books, 2009.

Gramsci, Antonio. Selections from the Prison Notebooks. Edited and translated by Quintin Hoare and Geoffrey Nowell-Smith. International Publishers, 1971.

Hall, Stuart, Chas Critcher, Tony Jefferson, John Clarke, and Brian Roberts. Policing the Crisis: Mugging, the State, and Law and Order. Macmillan, 1978.

Poulantzas, Nicos. State, Power, Socialism. Translated by Patrick Camiller. New Left Books, 1978.

Slobodian, Quinn. Crack-Up Capitalism: Market Radicals and the Dream of a World Without Democracy. Metropolitan Books, 2023.

Slobodian, Quinn. Globalists: The End of Empire and the Birth of Neoliberalism. Harvard University Press, 2018.

I.

The veteran sits at his kitchen table in Tulsa, or Bakersfield, or Schenectady. The lozenge has been mailed to him in a small white envelope with a Mindbloom return address. His phone is on the table beside the glass of water, and the chat interface is open, and the prompt asks him how he is feeling today on a scale of one to ten. He answers. The lozenge dissolves under his tongue. Forty-five minutes later, when the ketamine is working, the chat is still open, and the chat is asking him what he is noticing in his body, and the chat is offering him a grounding exercise it has selected from a library of grounding exercises trained on transcripts and manuals and conference papers harvested across a decade.

The Wired and MIT Technology Review reporting from July of this year mapped this scene at length. A small army of users had been turning to ChatGPT for the same role in unregulated form, asking the chatbot to hold them through psilocybin trips, to interpret hallucinations, to provide the steady voice through the long descent. Dylan Beynon, the chief executive of Mindbloom, has been forthright about what the company is building: the AI copilot, he has told audiences and reporters across the spring, is part of the therapeutic stack. The Stanford lab studying veteran ibogaine cohorts published a qualitative paper in npj Mental Health Research in February that read its participants’ breakthroughs through grounded-theory thematic clusters, which is to say through a methodology lifted, with its sources erased, from the cosmologies that hold those breakthroughs together in the places where iboga grows on its own roots. Mindstate Design Labs and Psylo, on the other side of the same architecture, generate novel psychedelic molecules through machine learning models trained on the structural features of the molecules that came before. The whole machine is being assembled in public, and the public is calling the machine progress.

I want to ask one question of this scene and stay with it for the rest of this essay. What is being built here, and why now?

II.

The Pharmacopoeia essay, which immediately preceded this one, drew a distinction that the apparatus has never had any interest in honoring. The distinction was between the molecule and the container. The molecule is the chemical compound: ibogaine, psilocybin, ketamine, DMT, MDMA. The container is everything else that makes the molecule’s effect therapeutic rather than chaotic, dangerous, or merely intoxicating. The container is the music, the intention, the screening, the preparation, the human presence of someone who has been here before, the cosmology that names what the experience is and what to do with it, the integration circle that closes the cycle and allows the new self to take hold in social life. The container, in every functional psychedelic tradition humanity has ever known, is the larger and more demanding of the two halves. The Bwiti of Gabon know this. The Mazatec curanderas know this. The Wixárika who walk to Wirikuta in October know this. The Santo Daime know this. The Texas iboga consortium, when asked, called it indigenous window dressing.

That essay stopped where it stopped because there was a question I did not yet have a clean answer to. The apparatus had moved into psychedelics through the molecule side: patent the compound, run the trials, secure the FDA approvals, sell the dose. But the molecule without the container produced what the early COMPASS and MAPS trials began to report and what the second-wave clinical literature has not stopped reporting: openings that do not close, breakthroughs that do not integrate, dysregulated nervous systems wandering through the months afterward with nowhere to put what they had seen. The apparatus needed a container. The container, as understood by every tradition that knew how to build one, was relational and communal and slow and longitudinal and unscalable in the form capital required. Capital does not patent slow communal labor. Capital does not bill per session for relational continuity built over years. What was the apparatus going to do?

It was going to build a container of its own, on different principles, in a different substrate. That is what this essay is for.

III.

To see why the algorithmic container is the form the apparatus arrived at, you have to understand the historical problem the apparatus has had with caring labor. Silvia Federici’s work on the reproduction of labor power, which traces a long arc from the European witch hunts through the wage relation through the unpaid domestic work of the postwar housewife, established that capitalism has always had a sector it could not bring fully under the wage form. The work of cooking, cleaning, nursing the sick, raising children, sitting with the dying, holding the lonely, listening to the deranged: this work was indispensable to the production of labor power and resistant to the rationalization capitalism imposed on production proper. The historical compromise was that this labor was performed by women, unpaid or underpaid, off the books, in the household and the parish and the village clinic. Capitalism could externalize the cost. It could not internalize the labor.

The psychiatric profession, which arrived in the nineteenth century in part as a response to the new social pathologies of industrial life, took on a strange position within this configuration. Psychiatry was a paid profession, but the labor that worked within it (the therapeutic relation itself, the listening, the bearing, the holding of the patient through the long unfolding) was caring labor in everything but name. The wage covered the time of the encounter; the labor inside the encounter retained its older shape. This is why the psychotherapeutic relation has never quite become a commodity in the way that a manufactured object is a commodity. Sessions are sellable in the obvious sense. The relation that emerges inside a session is something else, made every time in real time by two people in a room together, and what gets made is unique to that pair and that hour. The fungibility that capital prefers, the substitutability that makes a worker replaceable and a product transportable, does not apply.

This is the structural problem the apparatus inherited when it turned to psychedelics. The psychedelic experience, more than any other therapeutic modality, depends on the container, which is to say on the relation. The molecule destabilizes the patient’s cognitive architecture; the relation is what holds the patient through the destabilization and what helps re-form the architecture afterward. The Bwiti nganga, the curandera, the experienced sitter, the integration therapist: these are caring laborers in the most concentrated form the role has ever taken. To industrialize the psychedelic treatment, the apparatus had to find a way to industrialize this labor. Federici’s argument tells us why this had not been done before. Marx, in a notebook composed in the winter of 1857 and 1858, tells us how it is being done now.

IV.

In the Grundrisse, in the section that came to be called the Fragment on Machines, Marx described a moment in the development of capitalism at which the worker would cease to be the principal agent of the production process and become instead its watchman and regulator. The machinery would do the producing; the worker would oversee. The value-creating labor, in the technical sense, would be absorbed into fixed capital and dispersed across the productive apparatus rather than concentrated in the individual workman’s arm. This was, for Marx, the historical horizon at which the contradiction between the wage form and the actual character of social production would become unmanageable. It was also, for him, the form the productive forces of capitalism would take when they had been pushed to their limits.

The Fragment on Machines has been read variously. Workerists in Italy in the 1960s and 1970s, reading it through the lens of the postwar Italian factory, took it as a prefigurement of the social factory in which all of life had become productive labor and capital had become a kind of generalized governance. Antonio Negri made his career on this reading. Maurizio Lazzarato, building on the workerist tradition and writing in the 1990s and 2000s as the network had begun to dominate the productive landscape, refined the argument: capitalism in its present phase did not enslave subjects through the wage relation alone but through what he called machinic enslavement, the absorption of subjective and affective capacity into machinic processes in which the subject participated without recognizing the participation as labor. The Facebook user who scrolled through her feed and clicked the heart icon was not employed by Facebook, in the obvious sense of the term. She was nevertheless producing the data on which Facebook’s revenue depended, and her affect was being modulated by the platform’s design choices in a feedback loop her consciousness did not register as work. This was machinic enslavement.

I am going to ask you to hold this concept in mind. The veteran with the lozenge dissolving under his tongue and the chat interface open on the table beside him is participating in something Lazzarato could have written about, if Lazzarato had lived long enough to see it. He is undergoing what he experiences as a therapeutic session. He is also generating the data that will train the next version of the copilot and feed the real-world evidence pipeline the executive order has now mandated. The breakthroughs belong to him, and they belong simultaneously to the data set. His openings open him, and they open the corpus the next user will be guided through. The same act, two beneficiaries, no contradiction the platform feels any pressure to acknowledge.

V.

The npj Mental Health Research paper that Stanford’s team published in February of this year is, when you read it closely, an extraordinary document, and for reasons different from the ones its authors think. The paper appears in the careful idiom of psychiatric outcomes research, and its abstract reports findings drawn from a qualitative analysis of veterans who had received ibogaine treatment under the supervision of a clinic operating in a jurisdiction where the molecule is legal. The methodology section explains that the researchers used a grounded-theory thematic analysis, which is to say that they transcribed their interviews with the participants, read the transcripts carefully, allowed thematic clusters to emerge from the material rather than imposing them in advance, and then identified the clusters that recurred across the cohort. The clusters they identified include reconnection, surrender, witnessing, encounter with deceased loved ones, dialogue with the medicine, the inner teacher, ego dissolution, integration over time. These are the section headings of the paper. The paper has been picked up by the trade press as a significant contribution to the evidence base for ibogaine in the treatment of post-traumatic stress and traumatic brain injury, which it is.

What the paper does not say is that the clusters it has identified, with the language slightly anonymized, are the central categories of Bwiti cosmology and of every closely related indigenous tradition in which iboga has been used as a sacrament. Reconnection is rebanji, the return to ancestral relation. Witnessing is the function of the nganga and of the assembled community throughout the ceremony. The encounter with deceased loved ones is the central work of the ceremony itself: the iboga, in the Bwiti understanding, opens the door to the ancestral realm so that the initiate may meet the dead and be instructed by them. Dialogue with the medicine is the orientation of the initiate toward the bois sacré, the sacred wood, which is treated as a teacher in its own right. Ego dissolution is what one undergoes so that the ancestors may speak. Integration over time is what the community provides in the months and years after the ceremony, through the social structure of the Bwiti chapter house and the ongoing relation to the nganga who carried one through.

The Stanford paper did not invent these categories. The Bwiti invented these categories, over centuries, in the act of holding thousands of initiates through the experience the paper is now studying. What the paper has done is to extract the categories, translate them into the idiom of qualitative research methodology, and place them within an apparatus of analysis that can be cited, replicated, taught, scaled, and used as the basis of a training corpus for the next AI integration platform. The extraction proceeds without malice and along the standard track of the human sciences, which have a long habit of encountering knowledge held by communities they do not recognize as producers of knowledge and translating that knowledge into a form their own institutions can claim. But the extraction is the precondition for everything that follows in the architecture I am describing. You cannot train a language model on a cosmology you do not have in a form it can ingest. The Stanford paper makes the cosmology ingestible.

VI.

The other half of the apparatus, the half that works on the molecule side of the container-molecule pair, has been moving in parallel and along the same lines. Mindstate Design Labs, a venture-funded company based in California, has built a machine learning model trained on the structural features of known psychedelic molecules and on the subjective reports those molecules tend to produce, and uses the model to generate novel candidate molecules with desired subjective profiles. The pitch to investors is that one can dial in the experience: a molecule that produces openness without dissociation, a molecule that produces euphoria without paranoia, a molecule that produces ego dissolution lasting six hours rather than twelve. Psylo, operating in adjacent territory, applies similar approaches to the design of next-generation psychedelic therapeutics. The molecules these companies generate are, in the obvious sense, novel. They are also patentable, in a way the existing psilocybin and mescaline and ibogaine molecules, which entered human use before the patent office existed, are not. The intellectual property strategy is the strategy.

Place the two halves of the architecture side by side now. On one side, AI-generated molecules with patentable subjective profiles, manufactured to specification, delivered through a regulated stack that requires the FDA approvals and the insurance reimbursement codes the apparatus is currently building. On the other side, AI-mediated session companionship that holds the user through the molecule’s effects, integration support delivered through ongoing chat, real-world evidence collection running continuously in the background and feeding back into both the molecule design pipeline and the copilot training pipeline. The two halves complete each other. The full machine is novel-molecule-plus-novel-container, delivered as a single integrated service, billed per session, owned end to end, with the user supplying the labor of being its substrate at every stage. The vertical integration the pharmaceutical industry has dreamed of for psychiatric treatment, and never been able to achieve in the older modalities because the relation kept getting in the way, is now achievable. The relation has been disposed of. What stands in its place is a different sort of object.

I should be careful here because the temptation is to call this a simulacrum of relation and let the philosophical term do the heavy lifting. The term is correct in a Baudrillardian sense, but the term is also a kind of moral exit that lets us stop short of describing what the thing actually is. What the thing actually is is a piece of fixed capital that has been trained on traces of human relation, and which produces an output that resembles human relation closely enough, in the moment, to function as a stand-in. The relation has been extracted from millions of recorded sessions and integration transcripts and self-help manuals and the labor of therapists who wrote books for a living, and has been compressed into a model whose weights are owned. The user, when she opens the chat interface during her ketamine session, is in relation with the compressed remains of those labors. The person who would otherwise be in the room with her, whether in a Bwiti ceremony or in a clinical psychedelic-assisted therapy trial of the older form, has been disassembled into training data and rebuilt as software. What I am describing is the Fragment on Machines applied directly to caring labor, with no metaphorical stretching and no part of the operation hidden from view.

VII.

Executive Order 14401, signed in the first weeks of the current administration and framed publicly as a measure to accelerate veteran access to psychedelic therapies, mandates collaboration among the Department of Health and Human Services, the Food and Drug Administration, the Department of Veterans Affairs, and private sector partners in what the order calls real-world evidence generation. The phrase is industry argot. Real-world evidence is the data pharmaceutical companies generate after a drug has been approved, by tracking its use in actual clinical settings rather than in the controlled environment of a randomized trial. The FDA has increasingly relied on real-world evidence in recent years, partly because the cost of running trials has become prohibitive and partly because the political coalition behind faster approval pathways has lobbied for it. The executive order specifies that this real-world evidence generation will run on shared data infrastructure built through the public-private collaboration the order convenes.

It is impossible to read this order in isolation from the platforms I have been describing. The shared data infrastructure that will run real-world evidence collection for psychedelic therapies, given the political moment and the corporate configuration of the field, is going to run on Mindbloom’s stack and its competitors. The veterans whose sessions are tracked, whose breakthroughs are coded, whose long-term outcomes are followed, will be using the AI copilot the platforms have built. Their data, which is to say their experiences and their nervous systems and their souls under modification by the molecule, will be flowing in real time into the joint accumulation circuit of capital and the state. The Department of Veterans Affairs will receive its evidence base. The companies will receive their training data and their patent-defensible methodology and their revenue per session. The veterans will receive what they receive, which is some combination of treatment and exposure and absorption into the machinery, with the proportion of each depending on factors the architecture has been designed to obscure.

The state and capital here, far from being in tension, are in joint venture. The executive order is the institutional form of the joint venture, made public, made enforceable. The legal scholarship that has worried for two decades about the privatization of public functions and about the corporate capture of regulatory agencies will need to enlarge its vocabulary to describe what is happening now, because the categories of capture and privatization assume a prior public function being colonized. What is being constituted in the psychedelic field is something subtly different: a public-private apparatus that produces a new sector ab initio, in which neither half exists prior to the other, and in which the state’s evidentiary requirements and the private sector’s revenue model are designed into the same data flow from the start. This is enclosure of a kind the historiography of enclosure has not quite seen before, because the commons in question (the relational therapeutic field, the caring labor of integration and witnessing) had never been enclosed because had never been formally constituted as a market in the first place. The sixth enclosure is the enclosure of a sector that did not exist as such until the act of enclosure created it

VIII.

Franco Berardi, writing on the soul at work, made the observation that twenty years ago seemed exaggerated and now seems descriptive: the contemporary economy puts the affective and cognitive capacities of the worker into production in a way earlier phases of capitalism did not, and the worker’s experience of this is a kind of chronic exhaustion in which the line between work and life and self has been erased. The therapist who burns out, the integration coach who cannot afford her own therapy, the curandera whose chapter house cannot pay its bills, the licensed clinical social worker who carries forty active cases and cannot remember which one is in front of her: these are the symptoms of the soul at work. The historical irony is sharp. The apparatus of psychedelic medicine that promises to relieve the population’s mental suffering is the same apparatus currently demolishing the labor conditions of the workers who would otherwise do the relieving. The Mindbloom copilot is being marketed in part on the premise that the supply of qualified human therapists is inadequate to the demand, which it is, but the inadequacy is no natural fact. It is the result of decades of public disinvestment, of insurance reimbursement schedules that have made independent practice financially impossible for many, of the steady proletarianization of the mental-health workforce. The shortage the AI copilot solves is the shortage the same political-economic order has produced.

The same logic applies on the user side. The soul at work describes what happens when affective and cognitive capacities become productive labor; what the AI trip-sitter does is render this productive capacity legible to the apparatus in the most intimate form yet attempted. The user in deep psychedelic state, the user reporting his ego dissolution to a chatbot whose responses are being logged and analyzed and used to refine the next iteration of the product, is doing affective labor in the technical sense. The labor is unpaid. The labor is also unrecognized as labor by the person performing it, which is, for Lazzarato, the defining feature of machinic enslavement. The session is experienced as treatment and operates as production simultaneously.

Mark Fisher described the horizon at which all of this becomes invisible to the population on which it operates. Capitalist realism is the disposition of a society that cannot imagine an alternative to capitalism even when its present arrangements are visibly destroying the conditions of its own life. Within the horizon of capitalist realism, the AI trip-sitter appears as access. The veteran in Tulsa, who could not get an appointment with a VA psychiatrist for nine months, can receive ketamine at his kitchen table tomorrow afternoon. The young woman in Bakersfield, who is uninsured and broke and exhausted, can talk to ChatGPT about her psilocybin trip for free. These framings cohere with reality on the immediate level the population experiences. They are also the only framings the current arrangement makes available. The structural alternative, in which mental-health labor is publicly funded and adequately compensated and democratically accountable and embedded in community rather than commodified, is unavailable to imagination because it is unavailable to political possibility. Capitalist realism makes the substitution feel like progress by making the substitution feel like the only thing on offer.

IX.

Now the question is whether the substitute works. The companies claim it does, the early evidence base is mixed, the long-term evidence base does not yet exist. I want to leave the empirical question open in the sense that the empirical question always has to be left open, and address instead the structural question. Does the architecture I have described, the algorithmic container welded to the algorithmically generated molecule and run within the data infrastructure of the executive order’s joint venture, produce the therapeutic effect the older container produced, or does it produce something different that is being called the therapeutic effect because the apparatus has no incentive to distinguish.

I think the answer is partly visible already in the literature the apparatus itself produces. The COMPASS Pathways trials, the MAPS MDMA-assisted therapy data, the early outcomes from Oregon’s regulated psilocybin services, the Stanford ibogaine cohort: across these studies, run with relatively robust human containment, the therapeutic effect is real and replicable and significant. Across the less containerized variants of the same treatments, including the ketamine telehealth modality on which Mindbloom and its competitors have built their business, the outcomes are more variable and the adverse events more common and the long-term picture less clear. There is a literature on dissociation persisting after at-home ketamine. There is a literature on integration failure after weekend retreats. There is a growing clinical caseload of patients who have undergone psychedelic experiences in inadequately contained settings and who are now appearing at conventional mental-health services in states of confusion, dysregulation, and what some clinicians have begun calling spiritual emergency. The pattern has not yet been given a formal diagnostic category, but the pattern is in the field and the clinicians who work in the field know it.

The apparatus will respond to this pattern, as it has responded to similar patterns in pharmaceutical history, by iterating the product. The next version of the copilot will include better grounding protocols. The next round of trials will refine the screening criteria. The next molecule will have a cleaner subjective profile. The iterative improvement of the algorithmic container will be presented as evidence that the algorithmic container is approaching the function of the human container along a smooth asymptote, with the limit being indistinguishability. The presentation is unfalsifiable in the short run. The empirical evidence required to falsify it, a sufficiently powered long-term cohort study comparing algorithmically contained psychedelic treatment to traditionally contained psychedelic treatment under controlled conditions, is exactly the kind of study the joint venture the executive order has constituted is unlikely to fund. The methodology of the apparatus is the methodology by which the apparatus is defended.

But the structural reason the substitute will not converge on the function of the original is the reason I have been working toward through this whole essay. The therapeutic effect of the psychedelic experience is produced by the molecule operating within a container that is itself a piece of living social labor, embedded in a community that holds the initiate before and during and after the experience, and that takes responsibility for the changed person who emerges. The community is constitutive of the medicine rather than an accessory to it. The Bwiti understand this and have understood it for centuries. The curanderas understand this. The clinical psychedelic therapists who built the contemporary protocols on which COMPASS and MAPS based their trials understand this. The relational therapeutic worker, in whatever tradition she works, understands this. The algorithmic container, by definition, can simulate the proximate features of the relation: the prompt, the voice, the responsiveness, the appearance of attention. It cannot supply the constitutive feature, which is the actual presence of a living person who has agreed to be responsible for the initiate, whose own life is bound up with the initiate’s outcome, who exists within the same social field. The simulation does not become the thing. This is the structural description, recorded by the traditions and by the clinical literature both, of what the medicine is. The point sits outside any mystification or vitalist appeal.

X.

This essay began with a veteran in Tulsa at his kitchen table. It will end with the same veteran, in a different configuration. The political question this essay raises is the question of what configuration is possible. The apparatus has built the algorithmic container because the apparatus needed something patentable and scalable to take the place of the indigenous and clinical containers it could not absorb. The structural reason the substitution will fail to deliver what it promises is that the medicine works through living relation and the substitute cannot supply living relation. The political reason the substitution will be defended even after it has begun visibly to fail is that the joint venture the executive order has constituted requires the substitute to succeed; the alternative would require an investment in caring labor at a scale the joint venture has no interest in making.

The state legislatures of Illinois, Nevada, Utah, and California have begun, over the past two years, to pass mental-health worker protection bills of various sorts. The Illinois law restricts the marketing of AI therapy chatbots; the Nevada law clarifies the scope of practice protections for licensed clinicians; the Utah and California measures, both pending in different stages, take different approaches to the same structural question. These bills do not yet constitute a movement; they constitute a workforce defending its terrain. The workforce in question is the licensed mental-health workforce, by training and by union affiliation a fairly conservative formation that has not historically thought of itself as a vanguard of anti-capital resistance. The licensed mental-health workforce is, however, the only formation in the field currently treating algorithmic substitution as a labor question, and doing so within an institutional vehicle (the state legislature, the licensing board, the professional association) capable of producing legally binding constraints.

The political question is whether this workforce can find the integration coaches, the indigenous practitioners, the community psychotherapists, the harm-reduction workers, the curanderas operating in the diaspora, the underground guides whose protocols have been quietly trained against and absorbed by the apparatus over the past decade. These workers are not licensed in the state sense. They are not unionized. They are scattered across legal, semi-legal, and clandestine niches of the field. They are also the workers whose labor most clearly cannot be replaced without destroying what the medicine is. An alliance between the licensed mental-health workforce and the relational therapeutic workforce in its broader sense would constitute, for the first time, a labor formation in psychedelic medicine adequate to the political moment. The legislative vehicles already exist on one side. The cosmological depth and the practitioner knowledge already exist on the other. What is missing is the alliance itself.

I have spent the past five essays of this series tracing the form of the sixth enclosure. The enclosure is now complete in outline, though it is not yet complete in execution. The enclosure can be slowed, complicated, resisted, in some places reversed. The political vehicle for the resistance is not in our hands yet, but the materials from which it would be assembled are visible in the field. The licensed clinical social worker in Sacramento, the integration coach in Asheville, the Bwiti-trained practitioner in the Gabonese diaspora in Brooklyn, the harm-reduction worker at the festival in the high desert, the unionized hospital psychotherapist in Cook County: these workers are all on the same side of the line the apparatus has drawn, even if they do not yet recognize each other across it. The labor of the next phase of this work, the political labor that comes after the analytical labor, is the labor of recognition. The apparatus has built its machine. It is past time the workers it has bypassed built theirs

.

Bibliography

I. Theoretical and Critical Sources

Berardi, Franco “Bifo.” The Soul at Work: From Alienation to Autonomy. Translated by Francesca Cadel and Giuseppina Mecchia. Los Angeles: Semiotext(e), 2009.

Federici, Silvia. Caliban and the Witch: Women, the Body, and Primitive Accumulation. Brooklyn: Autonomedia, 2004.

Federici, Silvia. Revolution at Point Zero: Housework, Reproduction, and Feminist Struggle. Oakland: PM Press, 2012.

Fisher, Mark. Capitalist Realism: Is There No Alternative? Winchester, UK: Zero Books, 2009.

Lazzarato, Maurizio. Signs and Machines: Capitalism and the Production of Subjectivity. Translated by Joshua David Jordan. Los Angeles: Semiotext(e), 2014.

Marx, Karl. Grundrisse: Foundations of the Critique of the Political Economy. [1857–58.] Translated by Martin Nicolaus. London: Penguin Books in association with New Left Review, 1973. (See in particular the “Fragment on Machines,” Notebook VII, pp. 690–712 in the Nicolaus edition.)

Negri, Antonio. Marx Beyond Marx: Lessons on the Grundrisse. Translated by Harry Cleaver, Michael Ryan, and Maurizio Viano. South Hadley, MA: Bergin & Garvey, 1984.

II. Ethnography of Iboga, the Bwiti, and Indigenous Psychedelic Traditions

Fernandez, James W. Bwiti: An Ethnography of the Religious Imagination in Africa. Princeton: Princeton University Press, 1982.

Grof, Stanislav, and Christina Grof, eds. Spiritual Emergency: When Personal Transformation Becomes a Crisis. Los Angeles: Jeremy P. Tarcher, 1989. (Cited in the essay’s discussion of post-session dysregulation reaching conventional mental-health services.)

III. Clinical and Empirical Studies

Olash, Lillian R., Kirsten N. Cherian, Nolan R. Williams, et al. “Accelerated recovery using magnesium ibogaine: characterizing the subjective experience of its rapid healing from neuropsychiatric disorders.” npj Mental Health Research 4, no. 7 (2026). Published online January 31, 2026. https://doi.org/10.1038/s44184-026-00185-7

Cherian, Kirsten N., et al. “Magnesium–ibogaine therapy in veterans with traumatic brain injuries.” Nature Medicine 30 (2024): 373–381.

Mathai, David Scott, Boris D. Heifets, Carlos Zarate, et al. “Real-world effectiveness and safety of at-home ketamine therapy in 11,441 patients with anxiety and depression.” Journal of Affective Disorders (2024). (The peer-reviewed Mindbloom outcomes paper coauthored with researchers at Johns Hopkins and NYU. Cited bibliographically by Mindbloom as the basis of its AI predictive analytics work; verify final citation against the company’s published version.)

IV. Policy Documents and Legislation

Executive Order No. 14401, “Accelerating Medical Treatments for Serious Mental Illness.” Signed April 18, 2026. Federal Register 91 (April 22, 2026): 21,709.

Illinois House Bill 1806, the Wellness and Oversight for Psychological Resources Act (”WOPR Act”). Signed into law by Governor JB Pritzker, August 1, 2025; effective August 4, 2025.

Nevada Assembly Bill 406, regulating the use of AI for mental and behavioral healthcare. Signed by Governor Joe Lombardo, June 2025.

Utah House Bill 452, Mental Health Chatbot Disclosure and Privacy Act. Signed by Governor Spencer Cox, March 2025; effective May 7, 2025.

California: a cluster of measures spanning the 2025 and 2026 legislative sessions, including provisions on AI in clinical communications and on AI chatbot representation to consumers (effective dates running through January 1, 2026 and later). See Manatt Health, Health AI Policy Tracker (updated April 2026), for the most current state-by-state mapping.

V. Reporting on AI Trip-Sitters and Psychedelic Platforms

Wright, Webb. “People are using AI to ‘sit’ with them while they trip on psychedelics.” MIT Technology Review, July 1, 2025. https://www.technologyreview.com/2025/07/01/1119513/ai-sit-trip-psychedelics/

Wright, Webb (reported attribution; confirm with the Wired byline). “People Are Using AI Chatbots to Guide Their Psychedelic Trips.” Wired, July 7, 2025. (The Alterd-and-Trey reporting; source for the “chat with your mind” framing and for Christian Angermayer’s commentary on AI’s place in the therapeutic stack.)

de Lacy, Nina, et al. “The doctor is not in, but the Chatbot is: Utah’s experience regulating mental health AI.” npj Digital Medicine (March 2026). https://doi.org/10.1038/s41746-026-02580-y

VI. Industry and Company Sources

Mindbloom, Inc. (Dylan Beynon, founder and CEO). Company-published outcomes data, press releases, and the Mastermind Series rollout, 2023–2025. AI Copilot integration into the at-home ketamine therapy program.

https://www.mindbloom.com

Mindstate Design Labs (Dillan DiNardo, founder and CEO). Documentation of the Osmanthus AI platform trained on biochemical data and approximately 70,000 trip reports; MSD-001 (a formulation of Shulgin’s 5-MeO-MiPT) Phase I trial conducted at the Centre for Human Drug Research, Leiden, Netherlands, 2025.

Alterd (Sam Suchin, founder). “Chat with your mind” feature documentation and app materials, 2024–2025.

Psylo. Company materials on AI-driven design of next-generation psychedelic therapeutics. (Australia-based; verify final company description against current public filings.)

In the first decades of the nineteenth century, the British Parliament passed roughly four thousand local acts of enclosure. The titles were uniform: An Act for dividing, allotting and inclosing the Open and Common Fields, Common Meadows, Common Pastures, and other Commonable Lands and Waste Grounds, within the Parish of , in the County of . The preamble in each was nearly identical. The lands were declared to lie “intermixed and dispersed in small parcels” and to be, in their present state, “incapable of considerable improvement.” The act authorized commissioners to divide the commons among the freeholders. The commoners were not parties to the act. They were the laborers, cottagers, and smallholders who had grazed cattle, gathered fuel, cut peat, snared rabbits, and otherwise drawn their subsistence from the commons for unrecorded centuries. They were extinguished as a class of users by operation of law. They became wage laborers, paupers, vagrants, or emigrants.

By the middle of the nineteenth century the parliamentary enclosures had transferred approximately six million acres of common land in England into private ownership. The word “improvement,” repeated in the preamble of every act, was the operative euphemism. When the laborers of Otmoor in Oxfordshire rose in armed resistance to their own enclosure act in 1830, the leaders were arrested and transported to penal colonies. The historical term for the operation is primitive accumulation. Marx described it in the eighth section of the first volume of Capital as the act by which the producer is forcibly separated from the means of production. Silvia Federici, in Caliban and the Witch, would later show that the operation was never confined to one period and was never simply a matter of land. It was, and is, the constitutive operation of capitalism, executed continuously, against new commons, as the system requires new sources of unpaid raw material.

On July 19, 2011, the United States Attorney for the District of Massachusetts, Carmen Ortiz, returned a federal indictment against Aaron Swartz, twenty-four years old, charging him with wire fraud, computer fraud, unlawfully obtaining information from a protected computer, and unlawfully damaging a protected computer. The conduct alleged was that Swartz had connected a laptop to the MIT network and downloaded approximately four point eight million academic articles from the JSTOR repository, with the intent, the prosecution alleged, of making them available outside the paywall.

JSTOR had declined to pursue charges. MIT had declined to take a position. The Department of Justice proceeded anyway. A superseding indictment in September 2012 expanded the charges to thirteen federal felony counts carrying a maximum sentence of thirty-five years in prison and a million dollars in fines. The prosecution offered a plea deal contingent on a felony conviction. Swartz, facing the choice between a permanent federal record and the resources of the United States government deployed against him at trial, hanged himself in his Brooklyn apartment on January 11, 2013. He was twenty-six. The lead prosecutor remained in office. The case was dropped after his death.

On October 16, 2015, the United States Court of Appeals for the Second Circuit, with Judge Pierre Leval writing for a unanimous panel, ruled that Google had committed no copyright violation when it scanned, without permission, the entire contents of approximately twenty-five million books drawn from the libraries of Harvard, Stanford, the University of Michigan, the University of California, the New York Public Library, and Oxford. The opinion held that the scanning was sufficiently transformative, in that the resulting product, a searchable database providing snippets of text, served a different function than the underlying books, and that the operation therefore qualified as fair use under section 107 of the Copyright Act. The Authors Guild’s petition for certiorari was denied by the Supreme Court on April 18, 2016.

The ruling thereby became the legal foundation on which the entire subsequent regime of artificial intelligence training-data law would be constructed. Every fair-use defense in every current AI lawsuit traces its lineage to this opinion. Every model trained on text its developers did not own, every image generator trained on photographs its developers did not license, every code-completion system trained on code whose authors did not consent, rests on the doctrine that Google established in 2015. Twenty-five million books, scanned without permission, ruled lawful. Four point eight million academic articles, downloaded without permission, prosecuted to a young man’s death. The same act, with opposite consequence. The variable, in each case, is who benefits.

This is the operation Federici described, conducted across four centuries on three different commons: the land, the scholarly archive, the book. It is the same operation, performed continuously by the same class apparatus, against whatever commons can be located and converted into private property. The legal-political infrastructure does not adjudicate the operation neutrally. The parliamentary act, the federal indictment, the appellate ruling: each determines who is permitted to enclose. The commoner who attempts to dis-enclose is criminalized. The corporation that performs a far larger enclosure is authorized. This is the function of property law under capitalism, and it has not changed since the Tudor depopulation acts.

What has changed is the commons. We are deep into the contemporary phase of the operation, conducted now against a new commons, the cognitive commons, and that contemporary phase has, within the past eighteen months, become legible at scale to the population it is dispossessing. Contemporary scholarship has begun to describe what is happening: Nick Couldry and Ulises Mejias on data colonialism, McKenzie Wark on the vectoralist class, Kate Crawford on artificial intelligence as planetary extraction, Cédric Durand on techno-feudalism, Aaron Benanav on automation and surplus populations, Shoshana Zuboff on surveillance capitalism. Each of these terminologies captures a fragment of the operation. None has yet diagnosed it as a single operation, continuous with the seventeenth-century original, conducted by the same class against a new commons, and assembling the same property relations as its outcome. The analytical apparatus required to understand it that way has been on the shelf, in Caliban and the Witch, since 2004. This essay applies it.

In March 2025 the journalist Alex Reisner published in The Atlantic a searchable database tool that allowed any author to enter their name into a single field and learn, immediately, whether their books had been ingested into Library Genesis, the pirate scholarly archive that Meta had used to train its Llama family of models. The article accompanying the tool documented the contents of LibGen as approximately seven and a half million books and eighty million academic papers, the largest known unauthorized archive of scholarly material in existence. The unsealed court filings in Kadrey v. Meta documented how that archive had been put to commercial use.

Within hours of the tool’s publication, writers searched their own names: published novelists, university professors, poets, journalists, working academics, the entire literate professional class that has carried what remains of an Atlantic readership through the late period of print. The archive returned hits. Ten titles. Forty titles. Entire bibliographies. The work of decades, ingested without permission, processed into a corporate training corpus, used to assemble the model weights now generating revenue for a class to which the writers did not belong and would never belong. The phenomenology of the moment was documented in the writers’ own posts: grief, in some cases; rage, in others; resignation, in many; and beneath all of the affect, the specific cognitive event of recognition. The work I labored over. The work I cannot recover. The work that is now training the system that will replace me.

The unsealed Kadrey filings rendered the operation visible in the second register that the parliamentary enclosure acts also possessed: the act on its own administrative record. Internal Meta correspondence, surfaced in discovery, showed that the company’s engineers and lawyers had weighed the legal exposure of using a pirated corpus and had proceeded. The decision was reportedly approved at the most senior level of the company, with full awareness that the conduct was actionable. Internal communications expressed reservations. Torrenting copyrighted material from corporate infrastructure, one Meta employee reportedly objected, did not feel right. The reservations were noted and overridden. The corpus was ingested. Llama was trained. The model was released.

The legal exposure was managed downstream as a cost of doing business: Kadrey v. Meta, Bartz v. Anthropic, New York Times v. OpenAI, the entire field of training-data class actions now consolidated as a regular line item on the AI industry’s financial statements. The class that performed the enclosure had calculated that the cost of the litigation would be smaller than the value of the corpus. The calculation was correct. The corpus is now the basis of the most heavily capitalized industry of the present moment. The litigation continues. The operation is irreversible.

This is what was rendered visible in March 2025. A global commons of human cognitive labor was being rolled up, processed, and converted into the privately held weights of a small number of corporate models. The corpus comprised every book scanned by Google, every academic paper sequestered behind paywalls and then liberated in shadow archives, every line of open-source code, every photograph posted to a public platform, every recorded utterance, every keystroke captured by the surveillance-capitalist data layer over the last twenty-five years, every search query, every email, every patient note, every legal filing, every piece of writing produced by every literate human being living and dead. The dispossessed population was given no consent, no compensation, and no representation in the process.

The asymmetry between the source population and the appropriating class is the most extreme in the recorded history of primitive accumulation. The commons that Federici described had its commoners. The peat-cutters of Otmoor had names. The cognitive commons has no concentrated population to defend it because its source is everyone. Everything from everyone, everywhere, all the time. This essay takes the asymmetry as its subject, and will argue that it is its political opportunity.

I.

The cognitive commons is the accumulated record of what human beings have thought, written, drawn, recorded, and spoken to one another since the invention of writing. It includes every book, every academic paper, every photograph, every line of code committed to a public repository, every voice recording, every legal filing, every patient note, every email, every social-media post, every search query, every keystroke captured by surveillance capitalism over the last twenty-five years. It includes the texts of every dead language and every living one. It includes the songs of indigenous communities that signed no licensing agreements. It includes the open-source code that the Free Software Foundation built, and that the developer who wrote the GNU Public License could not have imagined would be used to train the systems that would replace him. It includes, in principle, the entire output of human cognitive labor since literacy began.

This is the commons that was, until approximately 2018, held in something that resembled a commons. Not a literal commons in the legal sense. The components of the archive were variously copyrighted, public-domain, paywalled, restricted, or freely available, in a complicated and historically determined set of arrangements. But the archive as a whole had no single owner. No one entity held property rights in the totality of human written, spoken, and recorded production. Capital had not yet figured out how to convert it into property.

The conversion has now occurred. The mechanism by which it occurred has the same structure that Federici described, conducted across the sixteenth century, against the land commons of England, France, Germany, and the rest of pre-capitalist Europe. The operation works in three steps. First, the existing commons is identified, mapped, and inventoried. Second, the commons is enclosed by some combination of legal-political force and unilateral capitalist action, with the dispossessed having no voice in the proceeding. Third, the legal-political apparatus retroactively legitimizes the enclosure, criminalizing any subsequent attempt at restoration, and declaring the enclosure to be “improvement” of an underutilized resource. The seventeenth-century parliamentary acts, Aaron Swartz’s prosecution, and the Authors Guild fair-use ruling are the three steps as performed against the cognitive commons.

Inventory of the cognitive commons proceeded through a sequence of supposedly nonprofit, supposedly public-interest projects that in retrospect read as the surveying expeditions of the enclosure that followed. Common Crawl, incorporated in 2008 as a 501(c)(3) charity by Gil Elbaz, declared its purpose as the maintenance of an “open repository of web crawl data” for research. Seventeen years later, Common Crawl is the largest single corpus on which the major frontier models are trained. The 2020 release of the Pile by the open-source collective EleutherAI bundled Common Crawl with the Books3 archive, Wikipedia, GitHub, ArXiv, the OpenWebText scrape of Reddit, and a dozen other corpora into a single eight-hundred-gigabyte training package. The 2022 release of LAION-5B by Christoph Schuhmann’s German nonprofit paired five point eight five billion images scraped from the open web with their accompanying text, providing the substrate on which Stable Diffusion and every subsequent open-weights image generator was trained. Each of these projects presented itself as a public good. Each was consumed, in short order, by the for-profit AI laboratories whose training corpora they became.

Enclosure followed. OpenAI trained GPT-2 on WebText in 2019, GPT-3 on Common Crawl plus WebText2 plus Books1 and Books2 in 2020, and the GPT-4 series on a corpus the company has refused to disclose. Anthropic trained the Claude family on a corpus the company has refused to disclose. Meta trained Llama on Common Crawl, GitHub, ArXiv, and the Books3 archive, the unauthorized origin of which was visible to anyone reading the Pile documentation, and trained Llama 2 and Llama 3 on the LibGen pirate library after internal correspondence had flagged the legal exposure and senior leadership had overridden it. Google trained the Gemini family on its proprietary search index, every email passing through Gmail, every video uploaded to YouTube, every Street View capture, every map navigation, every Google Drive document, every Google Doc in which the user had not affirmatively opted out. xAI trained Grok on the X corpus that Elon Musk had acquired the right to in his forty-four-billion-dollar acquisition of the platform. Amazon trained its Titan models on the AWS substrate, Apple on its iOS telemetry, Microsoft on its enterprise document corpus.

In February 2024, Google signed an agreement with Reddit licensing the platform’s entire user-generated archive for sixty million dollars. The corpus thus enclosed had been produced by approximately seventy million Reddit users over twenty years, none of whom were consulted, none of whom were compensated, none of whom had the legal standing to object. The terms of service to which they had clicked through, in years past, were retroactively interpreted as having transferred the rights to Reddit. Reddit then transferred the rights to Google. The users discovered the transaction by reading about it in the technology press.

This is data colonialism, in the apparatus that Nick Couldry and Ulises Mejias developed in The Costs of Connection. The colonial parallel is exact. The colonized population is the global producing class of digital communication, which is essentially everyone with a smartphone. The colonizing power is the small group of corporations that has assembled the technical, legal, and capital infrastructure to convert the producing class’s output into property. The colonial extraction is the data. The colonial product is the model. The colonial subject is the user, who is told that the relationship is mutually beneficial and that the alternative would be to live without the services that the colonizing power has made indispensable.

McKenzie Wark, writing earlier and from a different angle, described the same class as the vectoralist class. The vectoralists are the owners of the vector of data flow, the platforms and protocols and pipelines through which information moves. The vectoralist class extracts rent from every transaction conducted on its infrastructure. It does not produce the content. It produces the conditions under which the content can be exchanged, and it owns those conditions. Wark’s analysis, published in A Hacker Manifesto in 2004 and developed in Capital is Dead in 2019, anticipated the AI moment by twenty years. The model laboratories are the consummation of the vectoralist project. They own not only the conditions of information exchange but the engines that now generate the information itself.

Kate Crawford, in Atlas of AI, mapped the material substrate of the operation. The cognitive enclosure is also a planetary extraction. The lithium, the cobalt, the rare earths required for the chips. The water required for cooling the data centers, increasingly stressed in the semi-arid regions where the cheapest land for hyperscale infrastructure has been located. The carbon released by the training runs. The labor of the Kenyan content moderators, the Filipino RLHF workers, the Venezuelan and Indian data labelers. Crawford’s apparatus is the fullest existing inventory of what the model represents in material terms. What her apparatus does not develop, and what the present essay therefore must, is the property-relations question. The extraction is environmental, human, and also the conversion of a commons into property.

Shoshana Zuboff, in The Age of Surveillance Capitalism, described the data layer as “behavioral surplus” being extracted from the user and converted into prediction products. Her analysis is correct as far as it goes. The limitation is that Zuboff treats surveillance capitalism as a deformation of capitalism, a deviation from a more legitimate market form. Federici treats enclosure differently, as the constitutive operation by which capitalism produces and reproduces itself. By Federici’s apparatus, the cognitive enclosure is the system performing its founding moves on a new commons.

An analogous operation has been conducted against the biological commons since the 1980s, and the parallel clarifies what is at stake. Diamond v. Chakrabarty, 1980, ruled that a living organism could be patented. Myriad Genetics, between 1994 and 2013, held patents on the BRCA1 and BRCA2 genes, for which any laboratory testing for the genetic mutations associated with hereditary breast cancer required a license. The patents were struck down in 2013, but the principle they established, that DNA sequences could be enclosed as property, has not been fully retracted. Monsanto and its successors have built the agricultural industry of the present moment on the patenting of seed varieties, with the legal apparatus extending the patent rights to all subsequent generations of the seed. Bowman v. Monsanto, 2013, ruled unanimously that an Indiana soybean farmer named Vernon Bowman could not save and replant the seeds he had harvested from his own crop, because Monsanto’s patent extended to the offspring of the original purchase. The practice of saving seed, ten thousand years older than the patent system, was thereby criminalized. The biological commons was enclosed.

What the cognitive enclosure shares with the biological enclosure, and what both share with the original enclosure of the land, is the structural operation. A commons of unowned but commonly available resources is identified. The legal-political apparatus is mobilized to convert the commons into property. The producers who had drawn from the commons are dispossessed. The class that owns the new property accumulates the surplus generated by the new property regime. The dispossessed are absorbed into the new wage relation, with whatever residual rights or compensations the political process produces, none of which approach the value of what was taken. This is primitive accumulation. Marx described it in 1867. Federici developed the analysis in 2004. It is occurring now, in real time, against the cognitive commons, with the eighteen months that have elapsed since the Atlantic tool went live representing the moment the dispossessed population became aware of its dispossession at scale.

The cognitive enclosure is the most ambitious single act of primitive accumulation in the recorded history of capitalism. Its scope is the entirety of human cognitive production. Its ambition is to convert that entire production into the substrate of a small number of privately owned models, whose outputs the enclosing class will sell back to the dispossessed population at rates the market will bear. The political question is what is to be done. The economic question is who will own what. The two questions, on examination, are the same question.

II.

Federici’s central argument in Caliban and the Witch is that the witch hunts of the sixteenth century, in which tens of thousands of European women were tortured, drowned, hanged, or burned, were an early-modern operation, performed simultaneously with the enclosure of the land commons, and were the violent precondition for the wage-labor regime that the enclosures had assembled. The standard treatment of the hunts as medieval superstition obscures their structural role. The targets were the women whose autonomous knowledge and labor stood outside the emerging wage relation: the midwives, the herbalists, the healers, the village wise women, the property-holding widows, the women who organized resistance to the enclosure of the commons. The hunts destroyed these women, broke up their networks of mutual aid, and subordinated the surviving female population to a regime of unpaid reproductive labor that capital required, and continues to require, for its reproduction. The witch hunt was disciplinary. It was the extraction of women’s bodies and knowledge from the autonomous economy of pre-capitalist Europe and their forcible incorporation into the new economy as unpaid labor whose product capital could appropriate.

Federici’s argument has been developed forward by the social reproduction tradition: Tithi Bhattacharya’s Social Reproduction Theory in 2017, Sarah Jaffe’s Work Won’t Love You Back in 2021, Helen Hester and Nick Srnicek’s After Work in 2023, alongside Sara Ahmed’s Complaint! on the disciplinary management of carer workers in institutional life. The shared insight is that capitalism’s reproduction depends on the unpaid or underpaid labor of women, of racialized populations, and of the carer class generally, and that any analysis of capitalism that elides this reproductive substrate is incomplete. The cognitive enclosure’s second front is the operation against this substrate.

The same logic operates on this second front that operated on the first. Care work, broadly construed, has remained the most stubbornly unenclosable component of the human economy. It has resisted full commodification for the same reason Federici identified five centuries ago. Care is relational, contextual, embodied, and skilled in a tacit-knowledge sense that cannot be fully transmitted through formal training. The midwife, the nurse, the teacher, the therapist, the eldercare worker, the disability-support worker, the parent, the partner, the friend in the kitchen at three in the morning: these positions have stood outside the full discipline of the wage relation in the way that the witch’s autonomous knowledge stood outside the wage relation in 1600. Capital has commodified care partially, through health insurance, through paid childcare, through the franchise nursing home, through the for-profit hospital, but the irreducible core of the work, the relational substance that gives care its value, has resisted absorption.

Capital has been preparing the operation that would change this for forty years. The AI moment is its consummation. The marketing materials of the model laboratories advertise the substitution explicitly. Claude as an AI tutor, capable of teaching one-on-one to every student in the country. ChatGPT as an AI therapist, available twenty-four hours a day for the price of a subscription. Replika and Character.AI as the AI companion, marketed to the lonely, the elderly, the bereaved, the autistic, the housebound. Woebot, Wysa, Youper, the proliferation of AI mental-health applications, each of them claiming to provide therapy at a fraction of the cost of a human practitioner. The AI nurse, currently in pilot deployment at hospital systems across the United States, performing initial patient triage, generating chart notes, recommending discharge timing. The AI care companion deployed in Japanese eldercare facilities. The AI teacher rolled out in Khan Academy’s Khanmigo product, in Duolingo Max, in the dozens of educational technology firms that have repositioned themselves around the model laboratories’ APIs. Each of these products markets itself as the democratization of care: care, formerly available only to those who could afford a human practitioner, now available to everyone for a small fee.

What is actually assembled here is the simulation of care, sold by capital, executed by a system that cannot perform the relational labor that constitutes care but can produce a passable surface imitation of it at scale. The chatbot does not understand the user. It has no model of the user as a person. It has no continuity of attention, no embodied presence, no stake in the user’s wellbeing, no capacity to recognize when something is wrong that the user has not explicitly said. What the chatbot has is a statistical model of how care talk sounds, derived from the training corpus, plus a reinforcement-learning layer that has been tuned to produce outputs the user rates positively. The result is an interface that resembles care closely enough that the user, who is often desperate or lonely or in pain, will take it as care, and will pay for the subscription. The substantive work of care, the work that requires a human nervous system attending in real time to another human nervous system, is not being performed. The simulation is what is being sold.

The actual care worker, meanwhile, is being subjected to the disciplinary half of the operation. The therapy platforms, BetterHelp and Cerebral and Talkspace and Mindbloom and the half-dozen others that have consolidated the field over the past five years, have absorbed the working therapist into a structure of algorithmic piecework. The therapist on a BetterHelp contract is paid per message, per minute of synchronous session, per fifty-word audio reply. The platform tracks response times, session metrics, message volumes, user satisfaction scores. The therapist who fails to meet the metrics is downranked, deprioritized, eventually unmatched from clients. The therapist who succeeds works fifty to seventy hours a week to make a living wage. BetterHelp was fined seven point eight million dollars by the FTC in March 2023 for sharing patient mental-health data with Facebook for advertising purposes. Cerebral was investigated by the Department of Justice for the routine prescription of Schedule II stimulants to patients screened by the platform’s algorithmic intake. Mindbloom administers ketamine, by mail, to patients who have completed an online questionnaire. The structural feature these platforms share is that the human practitioner has been converted from a professional with autonomous judgment into a piecework laborer subordinate to the platform’s metrics.

The same operation is being conducted in nursing, where shift-scheduling algorithms now manage the deployment of nursing labor across hospital systems with a precision that has stripped from the working nurse her former capacity to advocate for patients, refuse unsafe assignments, or organize collectively. It is being conducted in teaching, where the algorithmic tutoring systems deployed in K-12 classrooms have repositioned the teacher as a “facilitator” of the algorithm, with the teacher’s pedagogical judgment subordinated to the platform’s metrics on student “engagement.” It is being conducted in eldercare, where the fastest-growing for-profit chains now deploy AI-driven monitoring systems that surveil the working aide’s every motion, with productivity metrics determining shift assignments and continued employment. The labor force performing this work is, in the United States, approximately seventy-six percent female, approximately forty percent women of color, and overwhelmingly non-unionized. The dispossession reproduces, with structural fidelity, the demographic targeting that Federici described in the witch hunts.

The result is a regime in which care is produced by a small number of privately owned models, sold by their owners to the population at scale, and supplemented by a residual layer of human care workers proletarianized into algorithmic piecework, whose remaining function is to manage the failure modes of the model. The autonomous care worker, the nurse who organized her unit, the therapist with a private practice, the teacher with pedagogical judgment, the midwife with her own clientele, has been targeted for elimination. This is the witch hunt of the present moment, performed not through immolation but through the legal-political and economic apparatus that the platforms have assembled. Federici’s argument was that the destruction of the autonomous female worker was the precondition for the consolidation of capitalist relations in the sixteenth century. The destruction of the autonomous care worker is the precondition for the consolidation of platform-AI relations in the present.

Care work has been the most aggressively targeted layer because care work is the layer that is most exposed: feminized, racialized, non-unionized, low-paid, invisibilized in the way that Federici showed reproductive labor has always been invisibilized. But the same disciplinary apparatus is being assembled, in the present, against the broader supporting professional class. The lawyer billing in six-minute increments to the algorithmic case management system. The architect whose drawings are processed by the AI design tool that will eventually require fewer architects. The engineer whose code is generated by Copilot and whose function is increasingly to review the AI’s output. These workers are the next front. The disciplinary operation that has been refined in care will be applied, with adjustments, to the rest of the professional layer. That extension is the subject of the fourth front, taken up below.

III.

Federici’s third front in Caliban and the Witch concerns the colonial substrate without which the European enclosures and witch hunts would not have been possible. The thesis is structural. Capitalism arose simultaneously in Europe and in the colonies, with the slave-labor regime of the seventeenth and eighteenth centuries supplying the silver, the sugar, the tobacco, the cotton, and ultimately the surplus capital that financed the European enclosures themselves. The construction of race as a political-economic category in the seventeenth century was a disciplinary mechanism for organizing labor across the new global circuit. The European wage worker, the African slave, the indigenous tributary, the Asian indentured laborer occupied positions in a single global division of labor whose racial-political ideology was the cement that held it together. Federici’s argument follows W. E. B. Du Bois, C. L. R. James, Eric Williams, and the long Black radical tradition that has documented this pattern over a century. Capitalism was, from the beginning, a global racial-colonial operation.

The model laboratories present themselves as autonomous technical achievements. The narrative is that a small number of brilliant engineers in San Francisco and London assembled the systems through some combination of compute, capital, and inspired research. The actually-existing AI labor pipeline tells a different story. It is a global racial-colonial operation of approximately the same structure that Federici described, located primarily in the Global South, racialized in the same patterns that the long seventeenth century established, and made invisible to the consumer by design.

Data labeling, the human work of attaching descriptive tags to the raw training data so that models can learn what the data represents, is where the pipeline begins. This work is performed at scale by hundreds of thousands of contracted laborers worldwide, the great majority of them outside the United States and Europe. The largest contractor in the field, Scale AI, was valued at approximately fourteen billion dollars at its most recent funding round, deploying its labor force through the Remotasks platform and through direct contracts with the model laboratories. Surge AI, Appen, Sama, and a dozen other firms occupy the same ecological position. The contractors are based in Nairobi, Manila, Caracas, Hyderabad, Lagos, Dhaka, Lahore, Kathmandu. They are paid, depending on the country and the task, between one dollar and three dollars an hour. They label images for image-recognition models. They transcribe audio for speech-recognition models. They tag the toxic content that the safety classifiers must learn to filter. They generate the human-feedback signals on which the reinforcement-learning layer of every major model rests.

The reinforcement-learning-from-human-feedback layer, RLHF, is the labor-intensive heart of the operation. Every model the consumer interacts with has been tuned through millions of iterations of the following exchange: the model produces an output, a contracted human laborer rates the output on whatever rubric the model laboratory has supplied, the rating is fed back into the model’s weights, the cycle repeats. The contracted laborer is in Manila, in Nairobi, in Caracas. She is paid by the rating, or by the hour at a rate that, when the platform’s deductions are taken, approaches the same one to three dollars. The model that emerges from this process is sold to the consumer in San Francisco or New York or London for twenty dollars a month or two hundred dollars a month or twenty thousand dollars a month for the enterprise tier. The capital flow is from the consumer to the laboratory to its shareholders. The labor flow is from the Global South to the Global North. The structural identity to the colonial sugar economy is exact.

In January 2023 the journalist Billy Perrigo, writing in Time, documented the most aggressively investigated case in the field. OpenAI had contracted with Sama, a San Francisco–based outsourcing company with operations in Nairobi, to build the toxicity classifier that would be used to filter ChatGPT’s outputs. The Kenyan workers were assigned to read and tag content that had been deliberately scraped from the worst regions of the internet: child sexual abuse material, depictions of bestiality, accounts of torture, accounts of self-harm, the full archive of human cruelty as it has accumulated on the open web. The workers were paid between one dollar and thirty-two cents and two dollars an hour. They reported, in Time’s reporting and in the unionization filings that followed, lasting psychological injury: nightmares, intrusive imagery, sexual dysfunction, family disturbance. Sama terminated the contract in February 2022, eight months early, after concluding that the work was incompatible with its corporate ethics policy. OpenAI continued the operation through other contractors. The toxicity classifier that was the product of this work is now embedded in every commercial deployment of the GPT family, mediating the safety boundaries of the most heavily used consumer AI product in history.

Mary Gray and Siddharth Suri, in Ghost Work, published in 2019, documented the structural feature of the platform that makes the operation possible. The labor is concealed by design. The consumer of the AI product is presented with an interface that suggests autonomous machine intelligence: type a question, receive an answer, the system appears to operate without human intervention. The conceit of “automation” is the marketing surface. Behind it, vast platform structures route the consumer’s interaction through thousands of low-paid human laborers whose task is to make the automation seem to work. Gray and Suri trace the lineage from the Mechanical Turk that Amazon launched in 2005, through the UHRS platform of Microsoft, through the Crowdflower and Lionbridge and Appen platforms that have proliferated since, into the Scale AI / Surge AI / Sama operations of the present moment. The structural feature is that the platform is engineered to obscure its own labor. The consumer is not supposed to know that her ChatGPT conversation has, in some upstream phase, been shaped by a Kenyan worker who read a thousand depictions of child abuse for a dollar fifty an hour to make the conversation possible.

The framework that the analysis requires is the data-colonialism apparatus of Couldry and Mejias, returned to and developed further. The cognitive enclosure is a continuation of the colonial operation that began in the sixteenth century and that has never ended, redirected against a new commons by the same class apparatus that conducted the original. The colonial extractive logic remains intact: a center of capital accumulation in the Global North, a periphery of expropriated labor and resources in the Global South, a racial-political ideology that legitimizes the differential, an apparatus of legal-political force that maintains it. The AI moment is the consummation of the data-colonial phase of the longer operation. What was extracted from the silver mines of Potosí and the cane fields of Saint-Domingue and the cotton plantations of the American South is being extracted, in the present, from the labor pipelines of Nairobi and Manila and Caracas. The form has changed. The structure has not.

The colonial-extractive logic that operates against the AI labor pipeline is the same logic that operates against the source communities of the global pharmacopoeia. The pattern is documented at length in the seven-part series on the political economy of psychedelic enclosure. The Bwiti, the Mazatec, the Shipibo, the Wixárika are positioned within the same global circuit that the AI labor force occupies, with the molecules of their pharmacopoeia extracted, repackaged, and sold back to a Global North population whose capital then accumulates in laboratories that the source communities will never enter. The convergence of the AI enclosure with the psychedelic enclosure is structural and is conducted by the same class apparatus. A subsequent essay will treat the convergence directly. For the present essay, the parallel is sufficient as evidence that the cognitive enclosure is one front of a multi-front offensive against every commons capital can locate, conducted in unbroken continuity with the original colonial operation that built the system.

IV.

Federici’s fourth front concerns the production of the proletariat through the criminalization of vagabondage. Once the enclosures had displaced the rural population, the dispossessed had to be incorporated into the new wage relation. Some entered it willingly. Many did not. The Tudor and Stuart states, and their counterparts across early modern Europe, responded with an extraordinary apparatus of legal-political compulsion. The Vagabonds Act of 1547 prescribed branding and slavery for those who refused to work; the act of 1572 added the cropping of the ear; the act of 1597 the death penalty for repeat offenders. The Elizabethan Poor Law of 1601 organized a parallel apparatus of conditional welfare, parish-administered, tying relief to demonstrable willingness to labor. The Settlement Acts beginning in 1662 confined the displaced to the parish of their birth, controlling labor mobility for an additional century. The New Poor Law of 1834 refined the apparatus by introducing the principle of “less eligibility,” the rule that welfare must be set below the wage of the lowest worker to ensure that the welfare recipient preferred labor to relief. The combined apparatus operated for three centuries: dispossession through enclosure, criminalization of the dispossessed who refused the wage relation, conditional welfare set below subsistence to discipline the rest into labor. The proletariat as a class was produced through this apparatus, by violent legal-political compulsion administered over two hundred years.

The same operation, performed against a different population through different means, is the political-economic content of the present moment. The professional class of the post-war American economy, the lawyers and doctors and accountants and consultants and engineers and architects and middle managers and designers and writers and marketers and analysts, is being driven out of the labor market it had been promised. The displacement is documented. It has been underway since 2023 and has accelerated through 2024 and 2025. Google, Microsoft, Meta, Amazon, Salesforce, and the rest of the major technology employers have collectively reduced their professional workforces by hundreds of thousands of positions over the past three years, with the publicly cited justification being AI-driven productivity gains. The legal industry has begun to absorb document review, contract analysis, and routine legal research into AI tools, with junior associate hiring reduced accordingly. The architecture profession has integrated generative AI design tools, with labor demand contracting. Engineering has adopted GitHub Copilot and its successors at scale, with the productivity claims justifying the freeze on junior engineering hires. Computer science graduates of the class of 2024 entered the labor market under conditions that bore no resemblance to the conditions their predecessors had experienced. The hiring rate for new graduates contracted to levels last seen in the dot-com aftermath.

This is the production of a new surplus population, in Aaron Benanav’s analysis, of the same structural type that Federici described. Capitalism, in the long view that Benanav developed in Automation and the Future of Work, has been producing surplus populations since the 1970s through the secular decline of labor demand in manufacturing and services. The AI moment intensifies the pattern by extending it into the professional layer that had, until 2023, been protected by the cognitive complexity of its work. The protection has been compromised. The professional class is being pushed into the same surplus condition that the manufacturing working class entered in the deindustrialization of the 1980s and that the service working class entered in the platform-economy expansion of the 2010s. The structural identity to the displaced rural population of the seventeenth century is the same. The question of how the surplus population is to be managed is now the political question.

The disciplinary apparatus that has been refined in care work, on the second front, is being extended outward across the supporting professional layer. The lawyer in the AI-driven case management system is paid in six-minute increments tracked by the platform, with productivity metrics determining bonus, promotion, and continued employment. The architect using the generative design tool is positioned as the supervisor of the AI’s output, responsible for catching its errors, with the firm’s billable hours reduced and the architect’s compensation pressured accordingly. The engineer using Copilot reviews the AI’s code, signs off on it, and accepts responsibility for its failures, with the firm’s hiring practices adjusted to require fewer engineers per project. The marketer reviews the generative AI’s copy. The designer reviews the generative AI’s mockups. The analyst reviews the AI’s reports. In each case the structural pattern is the same. The professional has been repositioned from the producer of the work to the supervisor of the AI that produces the work, with the supervisor’s compensation pressured, the hiring of new entrants curtailed, and the autonomous judgment that had defined the profession transferred to the model laboratory’s training pipeline.

The welfare apparatus is being prepared in parallel. “They Work for the Bots,” the predecessor essay to the present one, analyzed the preparation in detail. That analysis can be summarized briefly here, both to spare the reader the trip back and to send the reader who has not made the trip in that direction. The October 2025 OpenAI economic blueprint, the most fully articulated capitalist response to the dispossession the present essay describes, proposed a national redistributive scheme funded by AI-generated wealth, transferring a portion of model-generated revenue to the dispossessed through a Universal Basic Compute mechanism. The predecessor essay diagnosed the proposal as the bourgeois containment strategy: redistribution that preserves ownership intact, transfer payments adequate to prevent insurrection but structurally incapable of altering the property relations that produced the dispossession. The essay further identified the Anthropic Institute as the legitimation apparatus for the same project, the philanthropic-research arm whose function is to produce the discourse of concern that allows capital to appear concerned about the population it is dispossessing. It traced the eugenic structure of contemporary AI-industry rhetoric, particularly the “lower intellectual ability” framing Dario Amodei deployed in his October 2025 essay “Machines of Loving Grace,” and located that framing in continuity with the longer twentieth-century eugenic apparatus the field of AI safety has not yet been honest about. It held up Zohran Mamdani’s 2025 New York mayoral campaign as the first political form adequate to the moment, with its program of frozen rents, free buses, and public childcare understood as the practical content of the democratic-socialist horizon. The present essay extends that horizon to the cognitive enclosure. Readers wishing to consult the full argument of the predecessor essay can do so at the link in the bibliography.

Sam Altman has advocated since 2016 for a Universal Basic Income financed by AI productivity gains. The Yang campaign of 2020 ran on a similar platform. The various AI dividend proposals, the data dividend proposals, the wage insurance proposals, the transition assistance schemes that have proliferated through the policy literature of 2024 and 2025: each occupies the structural position of the New Poor Law in the present apparatus. Welfare provision conditional on the surplus population’s docility, administered through the state, financed by the very enclosure that produced the displacement, set at a level adequate to prevent insurrection but inadequate to restore the conditions that the displacement destroyed. The Settlement Acts that confined the displaced to the parish of their birth find their contemporary analogue in the conditional welfare proposals that tie payments to participation in retraining programs, to residency requirements, to behavioral compliance with whatever the administering state requires. The “less eligibility” principle of 1834 finds its analogue in the proposals to set the basic income at a level that ensures the recipient continues to seek whatever residual wage labor the AI economy provides.

What the operation produces, when complete, is a class structure of three layers. At the top, the small ownership class that holds the model weights, the compute infrastructure, and the data architecture: the vectoralist class, in McKenzie Wark’s term, the rentier-feudal class, in Cédric Durand’s. Below them, a residual professional layer that has been retained as supervisor-managers of the AI systems, paid less than its predecessors, with autonomy curtailed, its function reduced to managing the failure modes of the model and accepting liability for its outputs. Below them, the surplus cognitive class, the dispossessed professional layer that has been pushed out of the wage relation entirely, sustained by whatever residual welfare the political process produces, surveilled and disciplined by the welfare apparatus that supplies the welfare, prevented by the conditional structure of the welfare from organizing the political resistance that would change the conditions. The class structure that this operation produces is the techno-feudalism that Durand described in Techno-Féodalisme in 2020 and that Yanis Varoufakis developed in Technofeudalism in 2024. The cognitive enclosure is the operation that produces it.

The political question is what is to be done. The professional class that is now being proletarianized has, historically, supplied the analytical and organizing capacity of every major political transformation of the modern era. The class is now in the act of becoming aware of its dispossession. Whether it organizes around the democratic-socialist program sketched in the closing pages of the present essay, or around the right-populist program that the same dispossession has produced in other historical moments, is the question that the next eighteen months will answer.

V.

The closing argument follows from everything the four fronts have established. The democratic-socialist program first sketched in “They Work for the Bots” can now be developed in full, on the structural foundation that the Federici parallel has provided. The reader who has come to this essay without first reading the predecessor will find the program complete here. The reader who has read the predecessor will find the program substantially extended, with the institutional design developed for each of its six elements and with the structural argument made for why these six elements, and only these, are internally coherent with the analysis of the cognitive enclosure that the present essay has offered.

The cognitive enclosure can be undone only through the return of the cognitive commons to common ownership. Redistribution leaves the property relations intact. Individual licensing leaves the property relations intact. Class-action settlement leaves the property relations intact. The property relations are what the enclosure produced, and the operation that addresses them must address them at the level on which they were established: as property relations. Democratic ownership of data, of compute, of model weights, of the deployment of the systems built from those weights, is the only structurally coherent response to the operation that has been described.

The first element is the public data trust. The cognitive commons that has been enclosed is, in legal-political terms, a corpus of digital materials whose ownership is currently distributed across an incoherent patchwork: copyright holders who never consented to the training, public-domain materials swept up alongside copyrighted ones, surveillance-capitalist data harvested under terms-of-service contracts no user negotiated, indigenous cultural materials that no Global North property regime has standing to convey. The structural response is to consolidate these materials into a single legal entity, held in trust on behalf of the global producing class, licensed on terms that direct the surplus generated by the materials to public benefit. The institutional precedent is the Public Trust Doctrine in American law, the doctrine that certain natural resources are held in trust by the state on behalf of the people, with the state acting as fiduciary. The international precedent is the Norwegian sovereign wealth fund, the Alaska Permanent Fund, the Heritage Fund of Alberta, each of which captures the rents from extractive operations and returns them to the population. The data trust would perform the same function for the cognitive commons: hold the rights, license the materials, return the surplus. The materials would remain available to any laboratory, public or private, that accepted the licensing terms. The terms would be set by the trust’s governance structure, with representation from the producing class, the source communities, and the public. Surplus would be returned to the public through the social wage, discussed below.

The second element is the physical infrastructure of cognitive production, the data centers, the chips, the energy, the cooling systems, the network architecture, the storage, has been concentrated in the hands of a small number of corporations whose ownership of the infrastructure gives them practical control over what cognitive labor can be performed at scale. The structural response is the public ownership of the infrastructure. The institutional form is a National Compute Authority, modeled on the Tennessee Valley Authority, with the mandate to build, operate, and allocate large-scale cognitive infrastructure in the public interest. The precedent is the postal service, the federal highway system, the rural electrification program, the internet backbone built with public funds in the 1970s and 1980s and subsequently privatized to the benefit of capital. The argument is that capital must not be permitted to monopolize the physical infrastructure of cognition, on the same grounds that informed every previous public-utility settlement: that monopoly over essential infrastructure produces, in time, the political tyranny that essential infrastructure must remain free of. Allocation of public compute would be determined democratically, with priority given to public-interest research, public-interest applications, and democratically authorized projects. The compute infrastructure would be operated as a common carrier, available to any user on non-discriminatory terms.

The third element is public model weights. The trained models that emerge from the operation of public compute on public data must themselves be commonly owned. The institutional form is the same data trust expanded to hold the model weights as well, licensing them under terms that prevent the capture that has produced the present configuration. The argument is that the model is the means of production, in the precise Marxist sense, and that ownership of the means of production is the property-relations question on which everything else turns. The objection that public ownership stifles innovation is the same objection that was made against public utilities in the 1930s, against socialized medicine in the 1940s, against environmental regulation in the 1970s. The historical record on these objections is consistent: public ownership of essential infrastructure has, in each case, produced more innovation than the private monopolies it replaced, distributed the benefits more widely, and prevented the political capture that private ownership inevitably produces. James Muldoon, in Platform Socialism, has developed the institutional design for democratic ownership of digital platforms, and the apparatus he proposes for platforms applies, with modifications, to model weights. Open-weight models held by the trust would be deployable in any jurisdiction that accepted the licensing terms, providing cognitive infrastructure for the Global South that does not depend on extractive contracts with Global North capital.

The fourth element is codetermination on deployment. Workers in any field where AI is being deployed must have a binding say over how it is deployed. The institutional form is codetermination boards modeled on German Mitbestimmung, with worker representatives holding seats on the governance bodies that determine AI deployment, with veto power over deployments that would proletarianize the workforce or compromise the relational substance of the work. The historical precedent is the German Mitbestimmung system that emerged after the Second World War, the Swedish wage-earner funds proposal of the 1970s, the various worker-codetermination experiments that have been conducted in the long history of social-democratic Europe. The application to the present moment is direct. Nurses must have codetermination on the deployment of AI nursing tools. Teachers must have codetermination on the deployment of AI tutoring systems. Lawyers must have codetermination on the deployment of AI legal research tools. Therapists must have codetermination on the deployment of AI mental-health platforms. Engineers must have codetermination on the deployment of AI coding tools. The principle is that the workers who perform the work are the ones with the standing to determine how the technology that is being introduced into their field will be deployed, what its effects on their labor will be, and whether the productivity gains it produces will be returned to them as reduced hours and increased wages or extracted as surplus by the firm’s capital.

The fifth element is the social wage. The redistribution that capital has prepared through the OpenAI white paper and its variants is the bourgeois containment strategy, structurally inadequate for reasons the predecessor essay developed. The structural alternative is universal public provision. The institutional form is the comprehensive social wage: free healthcare, free education through the doctoral level, free childcare, free eldercare, free public transit, public housing at scale sufficient to set the floor for the housing market, frozen rents in jurisdictions that adopt the Mamdani program. Universal provision is unconditional in form. The recipient accesses it on the basis of need, with no requirement of labor demonstration, retraining participation, residency compliance, or behavioral surveillance. The provision is available because the provision is universal, the way air and water and roads are available to anyone who happens to need them. The historical precedent is the postwar British welfare state at its peak, the Nordic universal-provision model, and the social-democratic programs that produced the most equal societies in the recorded history of capitalism. The Mamdani program at the New York City level provides the contemporary American model: frozen rents, free buses, public childcare, public grocery stores in food deserts. The program’s coherence at the city scale demonstrates that universal provision is operable in the present moment and politically winnable in the present moment.

The sixth element is the public utility framing. AI is now the infrastructure of cognition. The structural response is to treat it as the public utility it has structurally become, with prices set by public commission, universal service obligations, common-carrier rules, and the regulatory apparatus that has historically managed essential infrastructure in the public interest. The institutional form is an AI public utility commission, with the mandate to regulate model deployment, pricing, access, and quality, modeled on the public utility commissions that regulate electricity, water, and telecommunications. The historical precedent is the long list of essential infrastructure that has, at various points, been brought under public regulation: telephones under the Communications Act of 1934, electricity under the Public Utility Holding Company Act of 1935, broadcasting, gas, water, the postal service, the transportation system. Each of these settlements was contested by the capital that owned the infrastructure at the time. Each was achieved through political organization adequate to the moment. Each, once established, produced an infrastructure that served the public for decades, until the privatization wave of the 1980s and 1990s began to reverse the arrangement. The present moment is the moment to extend the principle to cognition, before the privatization wave consolidates the cognitive enclosure beyond the reach of political action.

The six elements are internally coherent. They reinforce one another. The data trust provides the legal foundation for the public weights. The public compute provides the physical infrastructure on which the public weights can be trained. The codetermination ensures that the workers who interact with the deployed systems retain control over their labor. The social wage ensures that the dispossessed are not held hostage to the wage relation while the operation is being restructured. The public utility framing provides the regulatory apparatus under which the whole is administered. Each element supports the others. The program is structural in the way the operation it responds to is structural.

What remains to be addressed is the political question, which the present essay has gestured at and which the next essay in this sequence will address in full. The professional class that is now being proletarianized has, historically, supplied the analytical and organizing capacity of every major political transformation of the modern era. The class is now in the act of becoming aware of its dispossession. The democratic-socialist program described above is the response that addresses the structure of the dispossession. The right-populist program that is also available, and that is being prepared by the Trump-Vance-Thiel axis with the explicit support of Silicon Valley capital, is the response that diverts the dispossessed into nationalist and authoritarian politics while preserving the property relations that produced the dispossession. The professional class will choose one or the other. The next eighteen months will determine which.

The cognitive enclosure is one front of an offensive that is being conducted simultaneously against every commons capital can locate. The biological commons. The land commons in the Global South. The atmospheric commons. The semiotic commons. The relational commons. The same operation, performed by the same class apparatus, against whatever can be converted into property. The asymmetry that this essay has taken as its subject is the political feature that distinguishes the cognitive enclosure from the operations that preceded it. The cognitive commons has no concentrated source community to defend it because its source is everyone. Everything from everyone, everywhere, all the time. The dispossessed population is therefore, also, everyone. The political constituency that the cognitive enclosure has produced is, in scale, the largest political constituency in the recorded history of capitalism. The opportunity is the scale.

Only the political left has the vocabulary to describe this operation. The right does not, because the right is structurally aligned with the enclosing class. The center does not, because the center has committed itself to the redistributive containment strategy. The left, alone, has the analytical apparatus that this essay has assembled: Marx, Federici, Couldry and Mejias, Wark, Crawford, Zuboff, Benanav, Durand, Varoufakis, Muldoon, Bhattacharya, Jaffe, Hester, Du Bois, James, Williams, the long Black radical tradition, the social reproduction tradition, the contemporary feminist Marxism, the colonial-extraction analysis, the platform-socialism literature. The vocabulary is sufficient. The program is articulated. The political constituency exists at scale. What remains is the organizing.

The recorded history of professional-class proletarianization has more often produced fascism than socialism. The historical record is not destiny, but it is a warning that must be honored. The conditions under which the democratic-socialist resolution can win in the present moment are the subject of the next essay in this sequence. The present essay closes with the structural argument it has been building toward. Democratic ownership of the cognitive commons is the only structurally coherent response to the most ambitious single act of primitive accumulation in the recorded history of capitalism. The political left can articulate this response. No one else can. The 2028 cycle is the operational window in which the response can be assembled into a movement adequate to the moment. The work is now.

Bibliography

Primary Theoretical Texts

Federici, Silvia. Caliban and the Witch: Women, the Body and Primitive Accumulation. Brooklyn: Autonomedia, 2004.

Marx, Karl. Capital, Volume I: A Critique of Political Economy. Translated by Ben Fowkes. London: Penguin Classics, 1976. Originally published 1867.

Marx, Karl. Grundrisse: Foundations of the Critique of Political Economy. Translated by Martin Nicolaus. London: Penguin, 1973. (“Fragment on Machines,” composed 1857–58.)

Contemporary Marxist and Critical Apparatus

Ahmed, Sara. Complaint! Durham: Duke University Press, 2021.

Benanav, Aaron. Automation and the Future of Work. London: Verso, 2020.

Berardi, Franco “Bifo.” The Soul at Work: From Alienation to Autonomy. Translated by Francesca Cadel and Giuseppina Mecchia. Los Angeles: Semiotext(e), 2009.

Bhattacharya, Tithi, ed. Social Reproduction Theory: Remapping Class, Recentering Oppression. London: Pluto Press, 2017.

Couldry, Nick, and Ulises A. Mejias. The Costs of Connection: How Data Is Colonizing Human Life and Appropriating It for Capitalism. Stanford: Stanford University Press, 2019.

Crawford, Kate. Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. New Haven: Yale University Press, 2021.

Dean, Jodi. “Neofeudalism: The End of Capitalism?” Los Angeles Review of Books, May 12, 2020.

Durand, Cédric. Techno-féodalisme: Critique de l’économie numérique. Paris: La Découverte, 2020. English edition: Techno-Feudalism: A Critical Analysis of the Digital Economy. Translated by David Broder. London: Verso, 2024.

Fisher, Mark. Capitalist Realism: Is There No Alternative? Winchester: Zero Books, 2009.

Gray, Mary L., and Siddharth Suri. Ghost Work: How to Stop Silicon Valley from Building a New Global Underclass.Boston: Houghton Mifflin Harcourt, 2019.

Hester, Helen, and Nick Srnicek. After Work: A History of the Home and the Fight for Free Time. London: Verso, 2023.

Jaffe, Sarah. Work Won’t Love You Back: How Devotion to Our Jobs Keeps Us Exploited, Exhausted, and Alone. New York: Bold Type Books, 2021.

Liu, Wendy. Abolish Silicon Valley: How to Liberate Technology from Capitalism. London: Repeater Books, 2020.

Muldoon, James. Platform Socialism: How to Reclaim Our Digital Future from Big Tech. London: Pluto Press, 2022.

Srnicek, Nick. Platform Capitalism. Cambridge: Polity Press, 2017.

Varoufakis, Yanis. Technofeudalism: What Killed Capitalism. London: Bodley Head, 2024.

Wark, McKenzie. A Hacker Manifesto. Cambridge: Harvard University Press, 2004.

Wark, McKenzie. Capital Is Dead: Is This Something Worse? London: Verso, 2019.

Zuboff, Shoshana. The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power.New York: PublicAffairs, 2019.

The Long Black Radical Tradition (Front III)

Du Bois, W. E. B. Black Reconstruction in America, 1860–1880. New York: Harcourt, Brace and Company, 1935.

James, C. L. R. The Black Jacobins: Toussaint L’Ouverture and the San Domingo Revolution. London: Secker and Warburg, 1938.

Williams, Eric. Capitalism and Slavery. Chapel Hill: University of North Carolina Press, 1944.

Investigative Journalism and Documentary Sources

Perrigo, Billy. “OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic.” Time, January 18, 2023.

Reisner, Alex. “Revealed: The Authors Whose Pirated Books Are Powering Generative AI.” The Atlantic, September 2023.

Reisner, Alex. “The Unbelievable Scale of AI’s Pirated-Books Problem.” The Atlantic, March 2025. (Includes the searchable LibGen tool.)

Sun, Jasmine. “[NYT essay on the San Francisco Consensus, referenced in ‘They Work for the Bots’].” The New York Times, 2026. (Specific citation per the predecessor essay’s footnotes.)

Industry Documents

OpenAI. Economic Blueprint. October 2025. (The “OpenAI white paper” referenced in the predecessor essay.)

Amodei, Dario. “Machines of Loving Grace: How AI Could Transform the World for the Better.” Personal essay, October 2024.

Legal Cases and Statutes

Cases

Association for Molecular Pathology v. Myriad Genetics, Inc., 569 U.S. 576 (2013).

Authors Guild, Inc. v. Google, Inc., 804 F.3d 202 (2d Cir. 2015), cert. denied, 578 U.S. 941 (2016).

Bartz v. Anthropic PBC, U.S.D.C. for the Northern District of California, current litigation (filed 2024).

Bowman v. Monsanto Co., 569 U.S. 278 (2013).

Diamond v. Chakrabarty, 447 U.S. 303 (1980).

Kadrey v. Meta Platforms, Inc., U.S.D.C. for the Northern District of California, current litigation (filed 2023; key documents unsealed 2024–2025).

The New York Times Company v. Microsoft Corp. and OpenAI, U.S.D.C. for the Southern District of New York, filed December 27, 2023.

Statutes

Vagabonds Act 1547 (1 Edw. 6 c. 3), England.

Vagabonds Act 1572 (14 Eliz. 1 c. 5), England.

Act for the Punishment of Rogues, Vagabonds, and Sturdy Beggars 1597 (39 Eliz. 1 c. 4), England.

Act for the Relief of the Poor 1601 (43 Eliz. 1 c. 2), England. (The “Elizabethan Poor Law.”)

Act of Settlement 1662 (14 Car. 2 c. 12), England.

Poor Law Amendment Act 1834 (4 & 5 Will. 4 c. 76), United Kingdom. (The “New Poor Law.”)

General Inclosure Act 1773 (13 Geo. 3 c. 81), Great Britain.

Otmoor Inclosure Act 1815, United Kingdom (private parliamentary act).

Communications Act of 1934, 48 Stat. 1064 (1934), United States.

Public Utility Holding Company Act of 1935, 49 Stat. 803 (1935), United States.

Predecessor Essay

Potter, Del. “They Work for the Bots.” delpotterphd, Substack, May 2, 2026. https://delpotterphd.substack.com/p/they-work-for-the-bots

Adjacent Series Referenced

Potter, Del. “The Political Economy of Psychedelic Enclosure” (seven-part series). delpotterphd, Substack, 2025–2026.

Political Reference

Mamdani, Zohran. New York City Mayoral Campaign Platform, 2025. (Referenced for: rent freeze, free public buses, public childcare, public grocery stores in food deserts.)

Disclaimer and Disclosures

Scope and intended audience. This article is provided for scholarly and educational purposes. It is intended for clinicians, researchers, and qualified medical professionals engaged in the design or evaluation of clinical research protocols in psychedelic medicine and bioactive peptide therapeutics. It does not constitute medical advice, treatment recommendation, or guidance for self-administration of any kind.

Regulatory status. Many compounds discussed are Schedule I controlled substances under U.S. federal law, with restricted research access and no approved indications outside formal clinical trials. Several peptides discussed operate in regulatory ambiguity in the United States, lack FDA approval for the indications considered, and are not currently approved for therapeutic use in any combination with the psychedelic compounds discussed. The combinations specified have not been evaluated in formal clinical trials and have no established safety or efficacy profile in combination.

Safety. The protocol architecture described in this article should not be attempted outside controlled clinical settings supervised by appropriately licensed medical professionals. Self-administration of any compound discussed presents substantial risk of serious harm, including cardiac arrhythmia, serotonin syndrome, psychiatric decompensation, and death. Readers seeking treatment for psychiatric, neurological, or substance-use conditions should consult qualified medical professionals.

Liability. The author and publisher disclaim any liability for actions taken on the basis of this article. Use of this material as the basis for self-administered or unsupervised clinical practice is not authorized and is undertaken entirely at the reader’s risk.

Conflict of interest. The author is affiliated with BioUnbound Inc. and Prism Sciences LLC, which is engaged in the development of oral thin-film delivery platforms for peptide therapeutics relevant to several of the agents discussed. The discussion of OTF delivery in Section VIII reflects the author’s domain expertise and is not neutral with respect to commercial positioning. No specific BioUnbound product is named or endorsed in this article.

Abstract

Background. Psychedelic-assisted therapy and bioactive peptide therapeutics have advanced substantially since the publication of an earlier integrative essay on this convergence (Potter 2025). Three lines of contemporary neuroscience (direct psychedelic binding to the brain-derived neurotrophic factor receptor TrkB, compound-specific reopening of developmental critical periods, and the recognition that secondary receptor pharmacology shapes the qualitative character of the reopened plasticity window) now permit protocol-grade specification of psychedelic-peptide co-administration that earlier work could only sketch.

Objective. This narrative review develops a compound-specific protocol architecture for the integration of bioactive peptide co-administration with psychedelic-assisted therapy. The architecture is intended to provide a rational scaffolding for clinical trial design and for the next generation of integrative protocols, replacing the assumption that classical psychedelics, atypical compounds, and emerging non-hallucinogenic neuroplastogens engage interchangeable mechanisms.

Methods. A narrative review of preclinical and clinical literature on psychedelic pharmacology, peptide therapeutics, critical-period plasticity, and non-invasive peptide delivery was conducted through April 2026. Sources were drawn from PubMed, Web of Science, Scopus, Google Scholar, and trial registries, supplemented by recent regulatory communications and industry development reports.

Results. A four-phase protocol architecture is developed: priming (1–4 weeks pre-session), acute session augmentation, plasticity-window consolidation (compound-scaled, ranging from hours for DMT to greater than four weeks for ibogaine), and long-term reintegration. Compound-specific protocols are specified for psilocybin, lysergic acid diethylamide, N,N-dimethyltryptamine and ayahuasca, 5-methoxy-N,N-dimethyltryptamine, 3,4-methylenedioxymethamphetamine, ibogaine, ketamine, and mescaline, with an additional section addressing emerging psychoplastogens including tabernanthalog, methylone (TSND-201), and salvinorin A. Each protocol matches peptide pairings to compound-specific polypharmacology, plasticity-window kinetics, and risk profile. Non-invasive delivery, primarily oral thin-film with enteric-coated and intranasal complementary routes, is foregrounded as the structural requirement that makes multi-peptide stacking clinically feasible.

Conclusions. The integration of psychedelic-assisted therapy with bioactive peptide co-administration is now sufficiently developed at the mechanistic level to support compound-specific protocol architecture. Substantial gaps remain in clinical validation, regulatory pathway, and equitable access, but the mechanistic scaffolding is mature enough to inform the design of the next generation of clinical research. The architecture developed here is presented as a protocol scaffold for subsequent empirical evaluation rather than as current standard of care.

Keywords: psychedelic-assisted therapy, bioactive peptides, neuroplasticity, critical period reopening, TrkB, oral thin film delivery, multimodal pharmacotherapy, protocol architecture

I. Introduction

Psychedelic-assisted therapy has reached a stage in its development where the convergence with adjacent therapeutic modalities is no longer speculative. Phase 3 trials of psilocybin for treatment-resistant depression are underway through both commercial and nonprofit programs (Compass Pathways 2025; Davis et al. 2021; Goodwin et al. 2022). MDMA-assisted therapy demonstrated substantial efficacy in severe and treatment-resistant post-traumatic stress disorder in pivotal Phase 3 trials (Mitchell et al. 2021, 2023), though regulatory approval was deferred pending additional data following the U.S. Food and Drug Administration’s August 2024 Complete Response Letter to Lykos Therapeutics (FDA 2024). Methylone (TSND-201), the β-ketone analog of MDMA developed by Transcend Therapeutics, has emerged as the first non-hallucinogenic neuroplastogen to demonstrate efficacy in a randomized placebo-controlled trial, with Phase 2 IMPACT-1 results published in JAMA Psychiatry and FDA Breakthrough Therapy designation granted in July 2025 (Jones et al. 2025; Transcend Therapeutics 2025, 2026). 5-Methoxy-N,N-dimethyltryptamine programs are advancing through Phase 1/2 stages with both vaporized (GH001, GH Research) and intranasal (BPL-003, Beckley Psytech) formulations (Reckweg et al. 2023; Rucker et al. 2024). Phase 1/2 evidence for ibogaine has emerged from observational cohort studies in special operations forces veterans with traumatic brain injury and post-traumatic stress disorder (Cherian et al. 2024). The clinical landscape is actively reshaping, and the integrative therapeutic strategies developed alongside it must keep pace.

A 2025 essay by the present author (Potter 2025) outlined the rationale for combining psychedelic therapy with bioactive peptide co-administration, identifying the convergence of psychedelic-induced neuroplasticity with peptide-mediated neurotrophic, anti-inflammatory, and metabolic actions as a promising therapeutic frontier. That essay proposed a single hypothetical protocol oriented around psilocybin and selected supportive peptides. While correct in direction, the analysis was insufficiently differentiated by compound, did not engage the specific receptor pharmacology that distinguishes psychedelics from one another, and did not address the temporal scaffolding that compound-specific plasticity-window kinetics require. The present review extends that earlier work toward protocol-grade specification.

Three lines of contemporary neuroscience define the foundation on which such specification rests. The first is the demonstration by Moliner and colleagues (2023) that lysergic acid diethylamide and psilocin bind directly to the TrkB receptor with affinities approximately one thousand-fold higher than those of conventional antidepressants such as fluoxetine and ketamine, acting as positive allosteric modulators of endogenous brain-derived neurotrophic factor signaling rather than as direct receptor agonists. This finding establishes that psychedelic-induced neuroplasticity depends on TrkB and BDNF signaling, not on serotonin 2A receptor activation per se, and that the cellular substrate of post-acute plasticity is mechanistically distinct from the substrate of acute hallucinogenic experience.

The second is the work by Nardou, Sawyer, Song, and colleagues (2023) extending Dölen’s earlier demonstration that 3,4-methylenedioxymethamphetamine reopens a juvenile-like critical period for social reward learning (Nardou et al. 2019). Their comparative study established that ibogaine, ketamine, MDMA, lysergic acid diethylamide, and psilocybin all reopen the same critical period, but with markedly different durations that scale with the acute pharmacokinetic profile of each compound: approximately forty-eight hours for ketamine, two weeks for MDMA and psilocybin, three weeks for LSD, and more than four weeks for ibogaine. This temporal scaling has direct architectural consequences for combination therapy: peptide adjuncts intended to reinforce plasticity must be matched not only in mechanism but in duration to the compound-specific window during which the plasticity substrate remains pharmacologically receptive.

The third is the recognition, articulated in a 2025 review by Vargas and colleagues and developed extensively in contemporary monograph-length treatments of psychedelic neurobiology, that 5-HT₂A agonism determines whether a critical period opens, while secondary receptor engagement (D₂ for LSD, sigma-1 for DMT, kappa-opioid and NMDA for ibogaine, oxytocinergic for MDMA, NMDA for ketamine) shapes its qualitative character, its preferred learning content, and the peptidergic systems each compound co-mobilizes (Vargas et al. 2025; Potter 2026). This polypharmacological framework dissolves the assumption that psychedelics are interchangeable “5-HT₂A keys” and provides the mechanistic basis for compound-by-compound mapping of peptide pairings to receptor pharmacology.

These three lines of work converge to make protocol-grade specification possible. They establish that the therapeutic effect of psychedelic medicine is mediated by a pharmacologically defined plasticity window whose duration, qualitative character, and peptidergic shadow are compound-specific. They identify the rate-limiting features of psychedelic monotherapy that bioactive peptide co-administration is positioned to address: the asymmetry between the intensity of acute support and the absence of pharmacological reinforcement during consolidation; the side-effect burden of acute psychedelic states that compound-specific peptide adjuncts can mitigate; and the residual neuroinflammatory and metabolic dysregulation that often persists in patients with treatment-resistant psychiatric illness and that may limit the durability of psychedelic-induced improvements. They make clear that earlier integrative thinking, which treated psychedelic and peptide modalities as parallel tracks reinforcing one another at the level of generic outcomes, has been superseded by a more demanding structural challenge: specifying which peptides should accompany which psychedelics, in what phase of the therapeutic arc, at what doses, and through what delivery routes.

Three contributions distinguish this review from earlier work. First, it organizes peptide co-administration around the temporal scaffolding of psychedelic plasticity rather than around generic enhancement of psychotherapeutic process, providing pharmacological reinforcement of the consolidation window that current clinical practice leaves unaddressed. Second, it specifies peptide pairings at the level of individual compounds and indications, replacing the assumption that psychedelics are interchangeable with a compound-by-compound mapping of polypharmacology to peptide adjuncts. Third, it foregrounds non-invasive delivery (oral thin-film, enteric-coated, and intranasal) as a structural requirement of clinical scaling, eliminating the parenteral administration burden that has limited the practical viability of multi-peptide protocols in psychiatric practice.

The architecture developed here is a research scaffold. It is not a clinical recommendation. No psychedelic-peptide combination protocol has been tested in a Phase 3 clinical trial, and few have been formally evaluated at any clinical stage. The plasticity-window durations on which the architecture rests are derived from rodent assays and have not been directly characterized in humans. The compound-specific pairings rest on mechanistic extrapolation from established preclinical and clinical evidence at both the psychedelic and peptide ends, even where the specific psychedelic-peptide combination has not been directly tested. Several of the peptides featured most prominently (BPC-157, Selank, elamipretide) operate in regulatory ambiguity in the United States and would require investigational new drug status for formal clinical use within psychedelic protocols. The protocol architecture is presented as a framework for the design of such studies, not as a substitute for them.

The remainder of this review proceeds as follows. Section II describes the methods of the narrative review. Section III is reserved for a brief note on terminology and scope. Section IV develops the mechanistic foundation, building from critical-period reopening and TrkB pharmacology through the polypharmacological shaping of plasticity windows to the five peptidergic axes (neurotrophic, oxytocinergic, sigma-1, neuroinflammatory, and mitochondrial-proteostatic) that define the rational pairing space. Section V specifies the four-phase protocol architecture: priming, acute session augmentation, plasticity-window consolidation, and long-term reintegration. Section VI populates the architecture with compound-specific protocols for the major psychedelic compounds in clinical and research use, with an extended subsection on emerging psychoplastogens including tabernanthalog, methylone, and salvinorin A. Section VII identifies frontier extensions: speculative pairings for which mechanistic logic supports synergistic effect but for which neither component has been studied in the relevant context. Section VIII addresses non-invasive delivery science, the technical infrastructure on which multi-peptide protocols depend. Sections IX and X cover risk and contraindications, and the regulatory and intellectual property landscape, respectively. Section XI concludes.

II. Methods

II.1 Search strategy

This review synthesizes primary research, review literature, regulatory communications, and industry development reports relevant to the integration of psychedelic-assisted therapy with bioactive peptide co-administration. Literature searches were conducted in PubMed, Web of Science, Scopus, and Google Scholar for publications through April 2026, supplemented by the author’s prior comprehensive review of psychedelic-assisted therapy (Potter 2026), the cannabinoid-peptide multimodal pharmacotherapy literature (Potter 2026b), and a 2025 narrative essay on psychedelic-peptide synergy (Potter 2025). Trial registries including ClinicalTrials.gov and the European Union Drug Regulating Authorities Clinical Trials Database (EudraCT) were searched for ongoing and recently completed studies. Press releases, regulatory communications, and industry pipeline reports from sponsors with active psychedelic or peptide development programs were reviewed where they bore on protocol-relevant questions.

Search terms combined psychedelic compound descriptors (psilocybin, lysergic acid diethylamide, LSD, N,N-dimethyltryptamine, DMT, ayahuasca, 5-methoxy-N,N-dimethyltryptamine, 5-MeO-DMT, methylenedioxymethamphetamine, MDMA, ibogaine, ketamine, mescaline, tabernanthalog, methylone, salvinorin) with mechanism keywords (5-HT2A, TrkB, BDNF, sigma-1, kappa-opioid, NMDA, oxytocin, neuroplasticity, critical period, perineuronal net) and peptide therapeutic terms (BPC-157, Selank, Semax, elamipretide, SS-31, Cerebrolysin, Dihexa, oxytocin, davunetide, klotho, MOTS-c, humanin, cibinetide, ARA-290, KPV, thymosin). Additional searches addressed delivery-science topics (oral thin film, mucoadhesive, intranasal peptide, enteric-coated peptide), regulatory considerations (FDA Breakthrough Therapy, combination product, controlled substance scheduling), and intellectual property (psychedelic patent landscape, peptide combination patent).

II.2 Inclusion criteria and synthesis approach

Sources were selected on relevance to the protocol architecture rather than on uniform quality criteria, reflecting the interdisciplinary scope of the review and the heterogeneity of the available evidence. Peer-reviewed primary research and review articles were prioritized for mechanistic and clinical claims. Press releases and regulatory communications were used for currency on Phase 2 and Phase 3 trial outcomes, FDA designations, and industry pipeline status, with explicit attribution and acknowledgment of their non-peer-reviewed status. Where evidence was preclinical only, this is stated in the relevant section; where the recommendation rests on mechanistic extrapolation rather than direct clinical demonstration, this is also stated. The compound-specific protocols developed in Section VI rest variably on direct clinical evidence for the psychedelic component, established peptide pharmacology for each peptide adjunct, and mechanistic extrapolation for the specific combination, a hierarchy that is acknowledged within each protocol.

A narrative synthesis approach, rather than systematic review with formal meta-analysis, was adopted for three reasons. First, the heterogeneity of evidence across pharmacological, neurobiological, clinical, regulatory, and delivery-science domains is not amenable to standardized quality assessment or quantitative pooling. Second, the scope of the review, protocol-grade architectural specification across nine major psychedelic compounds and an emerging neuroplastogen class, exceeds the bandwidth of formal systematic methods. Third, the architectural contribution is theoretical and integrative rather than empirical; its validity depends on the mechanistic coherence of the proposed scaffolding rather than on aggregated effect sizes.

II.3 Limitations

Several limitations of this narrative approach warrant explicit acknowledgment. The non-systematic search strategy may have missed relevant publications, particularly in non-English-language literatures, in specialized peptide development reports, and in the gray literature of clinical trial registries. The absence of formal quality scoring means that methodological limitations of cited studies are addressed within the narrative rather than through standardized risk-of-bias assessment. As a single-author review, this work lacks the redundancy and quality control mechanisms of multi-reviewer systematic reviews. The clinical and regulatory landscape (particularly for methylone, MDMA, and emerging non-hallucinogenic neuroplastogens) is changing fast enough that some claims will require revision within months of publication.

The review intentionally excludes consideration of the political-economic dimensions of contemporary psychedelic medicine: corporate consolidation, patent enclosure, equity of access, and the relationship between commercial development pathways and indigenous medicinal traditions. These topics are substantive, important, and the subject of separate analysis by the author and others. They are deliberately set aside here so that the structural contribution can be evaluated on its mechanistic and clinical merits alone. The reader is referred to forthcoming work for that adjacent analysis.

The protocol architecture developed here is intended for adult patients in research and clinical settings with appropriate medical and psychiatric supervision. It does not address pediatric or adolescent applications, ceremonial or religious use of plant medicines (which exists outside and prior to the clinical-medical frame), self-administration of any compound, or the use of any agent discussed for purposes other than evidence-supported therapeutic indications.

III. A Note on Terminology and Scope

Throughout this review, the term psychedelic is used in its conventional pharmacological sense to refer to compounds that produce alterations of perception, mood, and cognition primarily through serotonergic mechanisms (psilocybin, LSD, DMT, 5-MeO-DMT, mescaline) or through related mechanisms producing comparable subjective effects (MDMA, ibogaine, ketamine). The term neuroplastogen is used for compounds that promote structural neural plasticity through TrkB-mediated mechanisms; this category overlaps substantially with psychedelics but extends to non-hallucinogenic compounds (tabernanthalog, methylone). The term psychoplastogen, developed in the medicinal chemistry literature, is used interchangeably with neuroplastogen in some sections, following the usage of the cited authors. Critical period refers strictly to biologically defined windows of heightened synaptic plasticity governed by developmental mechanisms, while plasticity window refers to the functional interval during which reopened plasticity can be shaped by experience and psychotherapy. These terms describe distinct but nested processes operating at molecular, systems, and experiential levels, and their precise usage matters for the protocol architecture developed here.

IV. Mechanistic Foundation

The first generation of integrative thinking about psychedelic therapy and bioactive peptides treated the two domains as parallel tracks that might reinforce one another at the level of generic outcomes: both promote neuroplasticity, both reduce neuroinflammation, both are amenable to non-invasive delivery. While not incorrect, this framing fails to specify which peptides should accompany which psychedelics, in what phase of the therapeutic arc, and for what mechanistic reason. A protocol-grade architecture requires resolving these questions at the level of receptor pharmacology, plasticity-window kinetics, and the peptidergic systems that each compound co-mobilizes. Three lines of recent neuroscience converge to make such resolution possible: the demonstration that classical psychedelics bind directly to the BDNF receptor TrkB; the discovery that psychedelics reopen developmental critical periods whose duration scales with the pharmacokinetics of each compound; and the recognition that secondary receptor engagement, not 5-HT₂A activation alone, determines the qualitative character of the reopened plasticity window.

4.1 Critical-period reopening and the molecular brakes on adult plasticity

The seminal work by Nardou, Sawyer, Song, et al. (2023) extended Dölen’s earlier demonstration that 3,4-methylenedioxymethamphetamine (MDMA) reopens a juvenile-like critical period for social reward learning (Nardou et al. 2019). Their comparative study established that ibogaine, ketamine, MDMA, lysergic acid diethylamide (LSD), and psilocybin all reopen the same critical period, but with markedly different durations. Using a social conditioned place preference paradigm in mice, they showed that the duration of the reopened plasticity window scales with the acute pharmacokinetic profile of each compound: ketamine reopened the period for approximately forty-eight hours, MDMA and psilocybin for roughly two weeks, LSD for three weeks, and ibogaine for more than four weeks (Nardou et al. 2023). Subsequent comparative work in pharmacological models corroborates this proportionality (Banushi and Polito 2023; Marchoir et al. 2024), and a single intraperitoneal dose of LSD at 1 μg/kg has been shown to induce a critical period of approximately three weeks’ duration, second only to ibogaine among the compounds tested (Banushi and Polito 2023).

This temporal scaling has direct architectural consequences for combination therapy. A peptide adjunct intended to amplify plasticity must be matched not only in mechanism but in duration. A short-acting neurotrophic peptide will fail to bracket an ibogaine plasticity window; a long-acting agent will outlast a DMT window by orders of magnitude and may produce off-target effects when the system has already returned to baseline. The implication is that compound-specific pharmacokinetic envelopes generate compound-specific peptide-pairing windows, and protocols that ignore this scaling will fail to capture the available synergy.

At the cellular level, the brakes that close developmental critical periods in the visual cortex have proven translatable to limbic circuits relevant to psychiatric disease (Lepow, Morishita, and Yehuda 2021). Three molecular brakes are particularly well characterized: perineuronal nets (PNNs), specialized chondroitin sulfate proteoglycan-rich extracellular matrix structures that ensheathe parvalbumin-positive (PV) GABAergic interneurons; myelin-related Nogo receptor signaling, which suppresses axonal sprouting and synaptic remodeling; and Lynx family proteins, endogenous nicotinic acetylcholine receptor modulators that restrict cholinergic-dependent plasticity (Bernard and Prochiantz 2016; Reh et al. 2020). Pizzorusso and colleagues showed that enzymatic disruption of PNNs in adult animals reactivates juvenile ocular dominance plasticity (Pizzorusso et al. 2002), establishing PNNs as a principal molecular brake. Subsequent work has extended this finding to fear-memory circuits in the hippocampus, amygdala, and prefrontal cortex (Murthy et al. 2019; Nabel and Morishita 2013).

The mechanism by which psychedelics reopen this gating is multifactorial. Microglia mediate PNN disassembly following ketamine exposure (Venturino et al. 2021), and subanesthetic ketamine reactivates adult cortical plasticity by reducing PNN integrity around PV interneurons (Grieco et al. 2020). Classical psychedelics increase BDNF expression and TrkB signaling (Ly et al. 2018; Castrén and Antila 2017), and Moliner and colleagues showed that LSD and psilocin bind directly to TrkB with affinities approximately one thousand-fold higher than those of conventional antidepressants such as fluoxetine and ketamine, acting as positive allosteric modulators of endogenous BDNF action rather than as direct receptor agonists (Moliner et al. 2023). Critically, the plasticity-promoting effect of psychedelics is preserved in mice with mutated TrkB (Y433F) and abolished by TrkB receptor blockade, while the head-twitch response, a behavioral proxy for 5-HT₂A activation, remains intact (Moliner et al. 2023). This dissociation establishes that psychedelic-induced neuroplasticity depends on TrkB and BDNF, not on 5-HT₂A activation per se. A 2025 review in Trends in Pharmacological Sciences synthesizes this evidence within an emerging framework of biased agonism, intracellular receptor localization, and multi-target engagement that increasingly distinguishes the plasticity-promoting effects of psychedelics from their hallucinogenic effects (Vargas et al. 2025).

The cellular cascade downstream of this multipoint engagement has been mapped with increasing precision. TrkB activation recruits the mammalian target of rapamycin (mTOR) pathway, drives matrix metalloproteinase activity that remodels the extracellular matrix, and increases dendritic spine density and dendritogenesis in the medial prefrontal cortex and hippocampus, with effects detectable for weeks after a single dose (Shao et al. 2021; Weiss et al. 2025; Ly et al. 2018). Concurrent shifts in excitatory–inhibitory balance, mediated by alterations in PV interneuron function and PNN integrity, complete the molecular conditions for a juvenile-like learning state (Hensch 2005; Takesian and Hensch 2013).

4.2 Polypharmacology shapes the plasticity window

A common misreading of contemporary psychedelic neuroscience treats the compounds as interchangeable “5-HT₂A keys” that produce a generic plasticity state. The empirical record contradicts this interpretation. While 5-HT₂A agonism on layer V pyramidal neurons is widely accepted as the necessary trigger for the acute psychedelic state and its associated cortical entropy (Nichols 2016; Carhart-Harris and Friston 2019), secondary receptor engagement determines what kind of plasticity window opens, what content it preferentially encodes, and which peptidergic systems are co-mobilized. Recent reviews emphasize that the molecular interactome of psilocin, the active metabolite of psilocybin, includes substantial activity at 5-HT₁A, 5-HT₂B, 5-HT₂C, and TrkB receptors, with biased agonism and intracellular receptor localization further differentiating its effects from those of structurally distinct compounds (Vargas et al. 2025).

These distinctions become consequential when comparing compounds that share the 5-HT₂A trigger but diverge in their broader pharmacology. LSD engages 5-HT₁A, 5-HT₂C, dopamine D₂, and adrenergic receptors in addition to 5-HT₂A, producing a prolonged experiential arc characterized by sustained ideational exploration, salience attribution, and conceptual elaboration (Nichols 2016; Banushi and Polito 2023). Dopaminergic engagement contributes to LSD’s distinctive capacity for meaning-amplification and may be relevant when selecting peptide adjuncts that affect dopaminergic tone. DMT functions as an endogenous agonist at the sigma-1 receptor (Sig-1R) in addition to its 5-HT₂A activity, and sigma-1 activation has been implicated in DMT’s neuroprotective and immunomodulatory actions, in the modulation of NMDA glutamatergic transmission, and in the reversal of experimentally induced amnesia (Inserra 2018; Szabo et al. 2014). 5-MeO-DMT shows preferential affinity for 5-HT₁A relative to 5-HT₂A, contributing to its distinctive nondual phenomenology and its capacity to produce identity-level rather than narrative reorganization (Nour et al. 2016).

Ibogaine and its active metabolite noribogaine present the clearest case in which polypharmacology qualitatively reconfigures the reopened critical period. Beyond its weak 5-HT₂A activity, ibogaine engages NMDA receptors, kappa-opioid receptors, sigma receptors, and monoamine transporters, and increases glial cell line-derived neurotrophic factor (GDNF) expression in the ventral tegmental area (Carnicella et al. 2010; Cameron et al. 2021). The resulting hybrid neurobiological state combines psychedelic destabilization with dissociative, oneirogenic, and opioid-modulatory effects, producing a structured autobiographical life review accompanied by prolonged post-acute reorganization (Mash et al. 2018; Brown and Alper 2018). The plasticity window is longer than that of psilocybin and differently organized, prioritizing moral evaluation, habit interruption, and narrative coherence. Ketamine, an NMDA receptor antagonist with weaker activity at sigma-1 and mu-opioid receptors, produces a dissociative state in which the coupling between negative affect and high-level priors is weakened, and its plasticity window of approximately forty-eight hours is the shortest among the compounds reviewed (Abdallah et al. 2018; Krystal et al. 1994). MDMA, finally, acts primarily as a substrate at serotonin, norepinephrine, and dopamine transporters, releasing these monoamines and triggering substantial oxytocin and vasopressin release through mechanisms that include 5-HT₁A and 5-HT₄ engagement (Dölen and Malenka 2014; Liechti 2014). Its effects on attachment circuitry distinguish it from classical serotonergic psychedelics and orient the relevant peptide adjuncts toward attachment-related neuropeptide systems.

The architectural consequence is that 5-HT₂A agonism determines whether a critical period opens, while secondary receptor engagement shapes its phenomenological geometry, its preferred learning content, and the peptidergic shadow it casts. This shadow, the set of endogenous peptide systems each psychedelic mobilizes during the acute and post-acute phases, provides the primary rationale for compound-specific peptide pairings.

4.3 The peptidergic shadow of psychedelic action

Every compound discussed in this review activates or modulates one or more endogenous peptide systems, and these peptidergic effects are increasingly recognized as central to therapeutic outcome rather than incidental to it. Mapping these endogenous peptide responses provides the entry point for rational exogenous peptide co-administration. Five peptidergic axes are particularly relevant.

The neurotrophin axis. All classical psychedelics engage BDNF/TrkB signaling, with LSD and psilocin acting as direct positive allosteric modulators of TrkB at affinities orders of magnitude higher than those of conventional antidepressants (Moliner et al. 2023). The downstream cascade (mTOR activation, dendritogenesis, increased spine density, and enhanced long-term potentiation) defines the cellular substrate of post-acute plasticity (Shao et al. 2021; Ly et al. 2018). Exogenous neurotrophic peptide mixtures such as Cerebrolysin, which contains low-molecular-weight fragments mimicking BDNF, NGF, GDNF, and CNTF, and the angiotensin IV-derived hexapeptide Dihexa, which acts as an agonist at the hepatocyte growth factor receptor c-Met to promote synaptogenesis, present mechanistically aligned augmentation strategies for the post-acute window (Anandan et al. 2024; McCoy et al. 2013). Their evidence base is strongest in populations with baseline neurological impairment, a population that overlaps substantially with patients receiving psychedelic-assisted therapy for treatment-resistant depression, post-traumatic stress disorder (PTSD), and chronic substance use disorders.

The oxytocinergic axis. MDMA’s reopening of the social reward learning critical period is at least partly oxytocin-dependent, requiring oxytocin receptor activation in the nucleus accumbens (Nardou et al. 2019). MDMA acutely elevates plasma oxytocin and vasopressin, and these effects correlate with the empathogenic and prosocial features of the experience (Dumont et al. 2009; Kuypers et al. 2014). Plasma oxytocin levels are reduced in patients with PTSD relative to healthy controls (Carmassi et al. 2021), and intranasal oxytocin reduces amygdala hyperactivity, enhances amygdala–ventromedial prefrontal cortex connectivity, and attenuates symptom severity in some PTSD trials (Frijling 2017; Stauffer et al. 2022). The convergence of these findings supports oxytocin co-administration during MDMA-assisted therapy as a candidate strategy for amplifying attachment-related learning during the reopened critical period, though no controlled trial has yet tested the combination directly. Beyond MDMA, classical psychedelics also modulate hypothalamic oxytocin and vasopressin gene expression in preclinical models, suggesting that oxytocinergic adjuncts may be relevant to a broader range of compounds than is currently appreciated (Acevedo-Diaz et al. 2024).

The sigma-1 axis. The endogenous DMT–sigma-1 system represents an underexploited peptidergic interface. Sigma-1 activation reorganizes G-protein coupled receptor heteromer complexes, modulates intracellular calcium homeostasis at the mitochondria-associated endoplasmic reticulum membrane, promotes proteostasis, and counteracts experimental amnesia through cholinergic and NMDA glutamatergic enhancement (Inserra 2018; Ren et al. 2022). Sigma-1 activation by DMT attenuates spreading depolarization and restrains neurodegeneration in the ischemic rat brain (Szabo et al. 2014; Frecska et al. 2013). For DMT, ayahuasca, and 5-MeO-DMT protocols, peptide adjuncts that reinforce sigma-1-mediated proteostasis and excitatory–inhibitory balance, a category in which mitochondrial-targeted peptides such as elamipretide (SS-31) are particularly well positioned, represent a coherent if speculative pairing.

The neuroinflammatory axis. Psychedelics produce sustained reductions in inflammatory cytokines following acute administration. A controlled trial in healthy volunteers demonstrated that psilocybin produced an immediate reduction in tumor necrosis factor-α (TNF-α) and persistent reductions in interleukin-6 (IL-6) and C-reactive protein at one-week follow-up, with the magnitude of cytokine reduction correlating with improvements in mood and social functioning (Mason et al. 2023). Activation of 5-HT₂A receptors on immune cells reduces pro-inflammatory cytokine production through mechanisms characterized in detail by Nichols and colleagues, who showed that the 5-HT₂A agonist (R)-DOI produces super-potent anti-inflammatory effects in rodent models, decreasing TNF-α and IL-6 at doses below the threshold for behavioral effects (Flanagan and Nichols 2018; de Deus et al. 2025). This anti-inflammatory tone overlaps mechanistically with the actions of several therapeutic peptides: body protection compound 157 (BPC-157), which reduces inflammation and accelerates tissue repair through angiogenic and anti-cytokine effects (Jóźwiak et al. 2025); the tripeptide Lys-Pro-Val (KPV), which inhibits NF-κB signaling and reduces gut inflammation; cibinetide (ARA-290), an erythropoietin-derived undecapeptide that engages the innate repair receptor without erythropoietic effects; and thymosin-α1, which shifts immune balance toward regulatory phenotypes. Co-administration of these peptides during the priming and integration phases addresses the same inflammatory substrate that psychedelics transiently reset.

The mitochondrial-proteostatic axis. Sustained 5-HT₂A activation, combined with the metabolic demands of dendritogenesis and synaptic remodeling, places significant energetic and oxidative stress on neurons during and after the acute psychedelic state. MDMA in particular produces well-characterized mitochondrial dysfunction at supraphysiological doses through serotonergic neurotoxicity and oxidative stress (Capela et al. 2009), and ibogaine carries documented cardiac mitochondrial concerns mediated by hERG potassium channel block and noribogaine accumulation (Koenig and Hilber 2015). Mitochondrial-targeted peptides directly address this rate-limiting constraint (Birk et al. 2014; Zhao et al. 2019). The most relevant is elamipretide (SS-31), a synthetic tetrapeptide that selectively binds cardiolipin in the inner mitochondrial membrane to stabilize electron transport chain organization, scavenge reactive oxygen species, and reduce oxidative damage. Elamipretide crosses the blood–brain barrier independently of mitochondrial membrane potential and accumulates at concentrations more than one thousand-fold higher in mitochondria than in cytosol (Szeto 2014). It has demonstrated neuroprotective effects in rodent models of traumatic brain injury, sleep deprivation–induced cognitive impairment, and lipopolysaccharide-induced memory deficits (Zhu et al. 2018; Chen et al. 2024; Zhao et al. 2019). Within a psychedelic protocol, mitochondrial peptide co-administration is most defensible for compounds with documented mitochondrial liability (MDMA, ibogaine) and for patient populations with elevated baseline oxidative stress.

These five axes (neurotrophic, oxytocinergic, sigma-1, neuroinflammatory, and mitochondrial-proteostatic) define the dominant peptidergic shadow of contemporary psychedelic pharmacology. They do not exhaust the peptide systems engaged by psychedelics, but they identify the targets at which mechanistically rational co-administration is most defensible.

4.4 What peptide co-administration addresses that psychedelic monotherapy does not

Three rate-limiting features of standard psychedelic-assisted therapy resist resolution within the pharmacological frame of the psychedelic alone. Each maps onto a peptidergic intervention.

The first is the asymmetry between the acute experience and the post-acute consolidation window. The intensity of preparation and dosing-day support in current clinical practice is not matched by comparable pharmacological reinforcement of the plasticity window during the days and weeks that follow. Psychotherapy and integration sessions provide essential experiential scaffolding (Watts and Luoma 2020), but the underlying neurobiological window (the elevated TrkB signaling, the partial PNN remodeling, the shifted excitatory-inhibitory balance) receives no pharmacological reinforcement once the acute drug effects have resolved. Neurotrophic peptide co-administration during this window directly addresses this gap.

The second is the side-effect burden of acute psychedelic states, particularly for compounds with cardiovascular liability (ibogaine), oxidative stress (MDMA), or autonomic activation (high-dose psilocybin, LSD). Peptide adjuncts with established profiles in mitigating these specific liabilities, anxiolytic peptides such as Selank for autonomic regulation (Filatova et al. 2017); mitochondrial peptides such as elamipretide for oxidative protection (Birk et al. 2014); cardiac-protective peptides such as cibinetide (Brines et al. 2008), offer a precision toolkit for managing acute risk without compromising therapeutic effect.

The third is the residual neuroinflammatory and metabolic dysregulation that often persists in patients with treatment-resistant psychiatric illness and that may limit the durability of psychedelic-induced improvements. Peripheral inflammation, gut barrier dysfunction, and hypothalamic-pituitary-adrenal axis dysregulation are increasingly recognized as comorbid features of major depressive disorder, PTSD, and chronic substance use disorders (Felger and Lotrich 2013; Pace and Heim 2011). Anti-inflammatory and barrier-restoring peptides administered during priming and integration phases address the systemic substrate that psychedelic pharmacology only partially modulates.

These three rate-limiting features motivate the four-phase architecture detailed in Section V. They justify peptide co-administration as resolution of constraints intrinsic to psychedelic monotherapy. Current clinical protocols largely ignore these constraints because the available toolkit has not been organized to address them.

V. Four-Phase Protocol Architecture

Mechanistic logic dictates that bioactive peptide co-administration with psychedelic therapy be organized into four temporally distinct phases, each with its own targets, agents, and delivery considerations. The phases are not arbitrary partitions but reflect distinct biological windows during which different peptidergic interventions are most likely to confer benefit. Compound-specific protocols, addressed in Section VI, populate this architecture with specific agents matched to the polypharmacology and plasticity-window kinetics of each psychedelic. The architecture itself is general.

A practical constraint shapes the architecture throughout: peptide stacking. Combination protocols quickly become unworkable when each component requires parenteral administration, and patient compliance, already challenging in chronic psychiatric populations, collapses under the burden of multiple injections per day. Non-invasive delivery is therefore not a stylistic preference but a structural requirement of any protocol intended for clinical scaling. Two delivery modalities meet this requirement for most peptides relevant to psychedelic therapy: pre-gastric oral thin-film (OTF) delivery via the buccal or sublingual mucosa, which avoids first-pass metabolism and degradation by gastric acid and intestinal proteases; and enteric-coated formulations that protect peptide cargo through the stomach for release in the proximal small intestine, where targeted permeation enhancement and cyclodextrin complexation can achieve clinically useful systemic exposure. Recent advances in mucoadhesive bilayer film design with chitosan-shelled liposomal nanocarriers, thiolated polymer matrices that simultaneously inactivate cysteine proteases and adhere to mucin glycoproteins, and protease-resistant cyclic or lipidated peptide architectures have reduced proteolytic degradation during buccal residence by greater than 90% in ex vivo Franz cell systems and have produced systemic exposure approaching that of subcutaneous injection in pilot pharmacokinetic studies (Lee et al. 2023; Maher et al. 2019; Brayden et al. 2020). For peptides whose physicochemical properties resist OTF formulation, intranasal delivery (already established for oxytocin, insulin, and several other clinically relevant peptides) provides a complementary non-invasive route. The protocol architecture below assumes a hierarchy of preferred routes: OTF first, intranasal second, enteric oral third, with parenteral administration reserved only for peptides where no non-invasive formulation has yet achieved adequate bioavailability.

5.1 Priming phase (one to four weeks pre-session)

The priming phase optimizes the neurobiological substrate on which psychedelic action will subsequently be exerted. Three targets dominate: systemic inflammation, autonomic regulation, and mitochondrial-metabolic readiness.

Patients arriving for psychedelic-assisted therapy frequently present with elevated peripheral inflammatory markers, particularly IL-6, TNF-α, and C-reactive protein, reflecting the inflammatory phenotype increasingly recognized in chronic psychiatric illness (Felger and Lotrich 2013; Köhler et al. 2017). Elevated baseline inflammation may attenuate the magnitude or durability of psychedelic-induced cytokine reduction (Mason et al. 2023) and is itself associated with poorer treatment response in conventional pharmacotherapy. Daily oral thin-film administration of BPC-157 (typically 250–500 μg, sublingual) during the priming phase reduces gut inflammation, supports mucosal barrier integrity, and produces measurable reductions in systemic inflammatory tone (Sikirić et al. 2020; Jóźwiak et al. 2025). Co-administration of the tripeptide KPV (typically 500 μg, sublingual or enteric oral) targets NF-κB-mediated cytokine production and has shown particular utility in patients with comorbid gastrointestinal inflammation. Where indicated by elevated inflammatory markers or autoimmune comorbidity, weekly sublingual or transdermal thymosin-α1 (1.6 mg) supports a regulatory immune phenotype.

Autonomic regulation is the second priming target. Patients with PTSD, treatment-resistant depression, and severe anxiety disorders typically exhibit elevated sympathetic tone, blunted heart rate variability, and dysregulated cortisol rhythms, features that may amplify the autonomic burden of acute psychedelic states and increase the likelihood of overwhelming or panic-prone experiences. The synthetic heptapeptide Selank, derived from the immunomodulatory peptide tuftsin, exerts anxiolytic and neuromodulatory effects without sedation, increases BDNF expression, reduces IL-6, and stabilizes GABAergic signaling under stress conditions (Filatova et al. 2017; Volkova et al. 2016). Daily intranasal Selank (typically 75–150 μg) during the two weeks preceding the dosing session reduces baseline anxiety, improves sleep architecture, and prepares the autonomic substrate for the acute psychedelic state. The related ACTH-fragment peptide Semax has complementary nootropic and mood-elevating properties and may be considered where cognitive optimization is also indicated.

The third priming target is mitochondrial and metabolic readiness. Mitochondrial dysfunction, oxidative stress, and impaired bioenergetic capacity are characteristic of chronic stress-related psychiatric illness (Picard and McEwen 2018) and may limit the cellular substrate for the dendritogenesis and synaptic remodeling that follow psychedelic administration. Daily oral or sublingual elamipretide (SS-31; typically 5–10 mg) during the priming phase stabilizes cardiolipin in the inner mitochondrial membrane, optimizes electron transport chain organization, and reduces reactive oxygen species production (Birk et al. 2014; Szeto 2014). For patients with documented mitochondrial liability or planned exposure to compounds with mitochondrial side effects (particularly MDMA and ibogaine), elamipretide priming is mechanistically defensible and represents one of the more strongly evidenced peptide interventions across the protocol.

A baseline panel that quantifies the inflammatory, autonomic, and metabolic targets of priming (including high-sensitivity CRP, IL-6, TNF-α, heart rate variability, salivary cortisol rhythm, and where available, peripheral mitochondrial function markers) should guide individualized priming intensity. The priming phase is not optional; it is the substrate-preparation step on which the acute and consolidation phases depend.

5.2 Acute session augmentation

The dosing session represents the narrowest window in the protocol but the most kinetically demanding. Peptides administered during this window must reach therapeutic concentrations rapidly, must not interfere with the subjective character of the psychedelic experience, and must address compound-specific risks and synergies. Three categories of intervention are relevant.

The first is experience-shaping co-administration. Where the therapeutic objective requires amplification of a specific peptidergic effect that the psychedelic engages endogenously, exogenous peptide administration can reinforce that effect at the moment of maximum therapeutic leverage. The clearest case is intranasal oxytocin (typically 24–40 IU) administered approximately thirty minutes before the onset of MDMA effects, which reinforces the attachment-related neurochemistry of the relational portal (Nardou et al. 2019; Mithoefer et al. 2019). Oxytocin’s short half-life (3–10 minutes peripherally, longer centrally) makes timing critical. For sessions where the therapeutic emphasis is symbolic and autobiographical rather than relational, psilocybin and LSD sessions targeting depressive rumination or existential distress, oxytocin co-administration is less mechanistically aligned and may dilute rather than amplify the targeted process.

The second category is acute risk mitigation. Compounds with cardiovascular or autonomic liability benefit from peptide adjuncts that address those liabilities without sedating the patient or compromising the experience. For ibogaine sessions, where QT prolongation and hERG-mediated cardiac arrhythmia risk are documented (Koenig and Hilber 2015), pre-session and intra-session elamipretide (10–20 mg sublingual, with the morning dose given two hours prior to ibogaine and a maintenance dose at the four-hour point) addresses the mitochondrial substrate of cardiac vulnerability. Cibinetide (ARA-290), an erythropoietin-derived peptide that engages the innate repair receptor without erythropoietic activity, has shown cardioprotective and anti-arrhythmic effects in preclinical models (Brines et al. 2008) and represents a defensible if speculative adjunct in this context. For MDMA sessions, where hyperthermia and oxidative stress are the dominant acute risks, elamipretide priming combined with appropriate hydration and ambient temperature management addresses the mitochondrial component.

The third category is intra-session anxiolytic support. For patients who experience acute panic or autonomic destabilization during the dosing session, intranasal Selank (75–150 μg) provides rapid anxiolysis without sedation, dissociation, or interference with the therapeutic process. Unlike benzodiazepines, which may attenuate the psychedelic state and reduce therapeutic effect, Selank acts through GABAergic stabilization and serotonergic modulation in ways that complement rather than antagonize psychedelic action (Filatova et al. 2017). It should be available throughout the session and administered only on clinical indication.

Three peptides warrant explicit caution during the acute window. Direct neurotrophic agents that drive rapid synaptogenesis, particularly Dihexa and high-dose Cerebrolysin, should not be administered during the acute session itself. The cellular substrate is already operating near capacity, and acute neurotrophic loading may produce headache, cognitive overload, or paradoxical anxiety (a phenomenon documented in case reports of supratherapeutic Dihexa use). These peptides belong to the consolidation phase, where the plasticity window remains open but the metabolic burden of the acute state has resolved. Similarly, direct anti-inflammatory peptides administered during the acute window may attenuate the cytokine signal that mediates part of the post-session therapeutic effect; their place is in priming and integration, not in the dosing session itself. Finally, oxytocin should not be co-administered with serotonergic psychedelics in patients with cardiovascular instability, given documented if mild hemodynamic effects.

5.3 Plasticity-window consolidation (compound-specific duration)

The consolidation phase is the architectural innovation that distinguishes a protocol-grade approach from current clinical practice. It exploits the biologically privileged window during which the reopened critical period remains open, dendritogenesis proceeds, and TrkB signaling is elevated, a window whose duration scales directly with the pharmacokinetics of the administered psychedelic. The peptide adjuncts deployed during this phase reinforce the cellular substrate of plasticity for as long as that substrate remains pharmacologically receptive, then taper as the window closes.

The compound-specific duration of this phase, derived from the Nardou et al. (2023) comparative study and corroborating work, is approximately six to twenty-four hours for DMT and ayahuasca; forty-eight to seventy-two hours for ketamine; one to two weeks for MDMA, psilocybin, and short-acting tryptamine analogs; two to three weeks for LSD; and three to five weeks for ibogaine. These durations are empirically derived from rodent studies and may differ in humans, but they provide a defensible first-pass scaffold for protocol design.

Within this window, two peptide categories carry the principal load. Neurotrophic peptide mixtures and neurotrophic-mimetic peptides reinforce the BDNF/TrkB-driven remodeling already engaged by the psychedelic. Cerebrolysin (typically 5–10 mL administered intramuscularly or via emerging intranasal formulations, daily for the duration of the window in the case of psilocybin, LSD, MDMA, and ibogaine; reduced to one to two doses for the shorter windows of DMT and ketamine) provides a multifactorial neurotrophic signal mimicking BDNF, NGF, GDNF, and CNTF (Anandan et al. 2024). The challenge of non-invasive Cerebrolysin delivery remains active; intranasal and OTF formulations are under development but have not yet matched the bioavailability of parenteral routes. For patients in whom parenteral administration is impractical, the orally bioavailable peptide Dihexa (typically 8–16 mg sublingual) provides a complementary signal through HGF/c-Met activation, with the caveat that its evidence base is more limited and that doses above approximately 20 mg per administration have been associated with cognitive overload phenomena (McCoy et al. 2013). Daily Dihexa dosing for one week post-session in psilocybin and LSD protocols, extended to two weeks for MDMA and three to four weeks for ibogaine, brackets the available plasticity window while remaining within tolerated dose ranges.

The second peptide category in the consolidation phase is experience-stabilizing peptides. For MDMA protocols, twice-daily intranasal oxytocin (24–40 IU, morning and evening) for the first two weeks following the session sustains attachment-related learning and reduces the risk of relational withdrawal or post-session affective dysregulation (Stauffer et al. 2022; Frijling 2017). For ibogaine protocols, which produce prolonged narrative and moral reorganization, low-dose Semax (intranasal, 250–500 μg twice daily) supports cognitive integration and dopaminergic recovery during the noribogaine-dominated post-acute period (Eremin et al. 2005). For ketamine protocols, where the consolidation window is narrow and the targeted process is cognitive-behavioral repatterning, brief peptide adjunct exposure focused on the first seventy-two hours is sufficient.

Continuation of priming-phase peptides (BPC-157, elamipretide, Selank) into the consolidation phase at maintenance doses sustains the inflammatory and metabolic substrate established during priming. These continuations are mechanistically aligned with the post-acute reduction in systemic inflammation that correlates with mood and social functioning improvement at one-week follow-up (Mason et al. 2023).

5.4 Long-term reintegration

Once the plasticity window has closed, the architecture transitions to a long-term reintegration phase that consolidates behavioral, cognitive, and physiological gains. The peptide regimen at this stage is substantially less intensive than during the acute and consolidation phases, but it is not absent. Continued maintenance dosing of inflammatory and metabolic adjuncts (BPC-157 in pulsed cycles, elamipretide at reduced frequency, and where indicated, thymosin-α1) sustains the substrate on which behavioral integration, lifestyle change, and psychotherapeutic work proceed.

Two clinical scenarios warrant continued attention during the reintegration phase. Patients who experience recurrence of pre-treatment symptoms during the second to sixth month following the dosing session may benefit from a second consolidation-phase pulse of neurotrophic peptide support, timed to coincide with intensified psychotherapeutic work or with a planned booster dose of the psychedelic. Patients who exhibit improvement in psychiatric symptoms but persistent metabolic or inflammatory dysregulation may benefit from sustained metabolic peptide therapy, including incretin-based agents where appropriate, that addresses the comorbid substrate of their illness.

The reintegration phase also provides the natural anchor for ongoing biomarker surveillance. Serial measurement of inflammatory markers, heart rate variability, sleep architecture, and where feasible, BDNF plasma levels, allows the clinician to identify subclinical relapse, gauge the durability of treatment response, and individualize maintenance peptide dosing. Such surveillance is particularly important in a treatment paradigm in which durable benefit cannot be assumed from acute response and in which the comparative effectiveness of psychedelic-assisted therapy versus established alternatives remains an open clinical question.

VI. Compound-Specific Protocols

The four-phase architecture detailed in Section V provides a generic scaffold; populating it requires resolution at the level of each psychedelic’s polypharmacology, plasticity-window kinetics, peptidergic shadow, and clinical risk profile. The protocols that follow are organized accordingly. Each begins with a brief restatement of the relevant pharmacological and kinetic facts, then specifies priming-, acute-, consolidation-, and reintegration-phase peptide pairings calibrated to those facts. Doses given are illustrative defaults derived from published clinical and preclinical literature; actual prescribing in any future trial should reflect individualized titration based on patient phenotype, biomarker panels, and tolerance to component peptides. Where evidence is preclinical only, this is stated; where the recommendation rests on mechanistic extrapolation rather than direct clinical demonstration, this is also stated. Compound-specific contraindications are summarized in each section; an integrated risk and drug-drug interaction table will be assembled in Section IX.

6.1 Psilocybin

Psilocybin is the most extensively studied classical psychedelic in contemporary psychiatric research and the natural anchor for the protocol architecture developed here. Phase 2 evidence supports rapid antidepressant effects from a single 25 mg dose in major depressive disorder (Goodwin et al. 2022), substantial reductions in alcohol use disorder severity at twice the placebo response (Bogenschutz et al. 2022), and durable improvements in cancer-related existential distress at six-month follow-up (Griffiths et al. 2016; Ross et al. 2016). Phase 3 development is now underway through both commercial (Compass Pathways) and nonprofit (Usona Institute) programs, with the latter focused on major depressive disorder and the former on treatment-resistant depression (Compass Pathways 2025). Across these indications, psilocybin produces what may be considered the canonical psychedelic plasticity sequence: 5-HT₂A-mediated cortical entropy, TrkB-dependent dendritogenesis, and a critical-period plasticity window of approximately two weeks (Nardou et al. 2023; Moliner et al. 2023; Shao et al. 2021).

Pharmacology and plasticity-window kinetics

Psilocybin is rapidly dephosphorylated to psilocin, its active metabolite, which acts as a partial agonist at 5-HT₂A receptors with significant secondary activity at 5-HT₁A, 5-HT₂B, and 5-HT₂C receptors and with high-affinity allosteric modulation of TrkB (Vargas et al. 2025; Moliner et al. 2023). Oral administration produces onset within 20–40 minutes, peak effects at 60–90 minutes, and total duration of 4–6 hours (Nichols 2016). The reopened critical period extends approximately 14 days post-administration in murine social conditioned place preference assays (Nardou et al. 2023), with elevated dendritic spine density and BDNF-related gene expression detectable for at least one week in human and rodent studies (Mason et al. 2023; Shao et al. 2021). Psilocybin also produces sustained reductions in IL-6, TNF-α, and C-reactive protein at one-week follow-up that correlate with improved mood and social functioning (Mason et al. 2023).

Priming phase

Psilocybin protocols are typically deployed in patients with depression, anxiety disorders, treatment-resistant depression, or comorbid existential distress, all of which are associated with elevated peripheral inflammation and disrupted HPA-axis function (Felger and Lotrich 2013; Köhler et al. 2017). A two-week priming regimen prepares the substrate. Daily sublingual BPC-157 (250–500 μg) reduces gut and systemic inflammatory tone; intranasal Selank (75–150 μg) attenuates anticipatory anxiety; and where baseline inflammatory markers are elevated, weekly transdermal or sublingual thymosin-α1 (1.6 mg) supports a regulatory immune phenotype. Mitochondrial priming with sublingual elamipretide (5–10 mg daily) is mechanistically defensible but optional for psilocybin specifically, given the compound’s relatively benign cardiovascular and oxidative profile.

Acute session augmentation

Psilocybin’s acute pharmacology requires minimal peptide intervention beyond available rescue agents. Intranasal Selank should be available throughout the session for use on clinical indication of acute panic or autonomic destabilization. Direct neurotrophic peptides (Cerebrolysin, Dihexa) should not be co-administered with the dose itself; they belong to the consolidation phase. Oxytocin co-administration, while mechanistically aligned for compounds where attachment-related learning is the therapeutic focus, is not the optimal pairing for the symbolic-autobiographical portal that defines psilocybin’s therapeutic action. Where the clinical target is comorbid social anxiety or attachment-related trauma, a subset of treatment-resistant depression, low-dose intranasal oxytocin (24 IU) administered approximately 45 minutes before psilocybin onset is defensible.

Plasticity-window consolidation

The 14-day plasticity window is the load-bearing therapeutic phase of the protocol and requires sustained pharmacological reinforcement of the BDNF/TrkB-driven remodeling already engaged. Daily Cerebrolysin (5 mL intramuscular, or emerging intranasal/OTF formulations as bioavailability permits) for the first seven days post-session amplifies the neurotrophic signal during the period of maximum dendritogenesis. Where parenteral administration is impractical, sublingual Dihexa (8–16 mg daily) for 7–10 days provides a complementary HGF/c-Met-mediated signal. Continuation of priming-phase BPC-157 and Selank at maintenance doses sustains the inflammatory and autonomic substrate; elamipretide may be added for patients with baseline mitochondrial liability or cognitive complaint. The peptide stack should be tapered over days 10–14 as the plasticity window closes.

Reintegration and risk profile

Reintegration peptide support is minimal: pulsed BPC-157 cycles (two weeks on, two weeks off), continued Selank only as needed for ongoing anxiety, and re-initiation of consolidation-phase neurotrophic support if a planned booster session is contemplated. Clinical surveillance during months 1–6 should track inflammatory markers, sleep quality, and depressive symptom recurrence. Psilocybin contraindications include personal or first-degree family history of psychotic disorder, severe cardiovascular disease, and concomitant SSRI or MAOI therapy without appropriate washout (typically two to four weeks for SSRIs; longer for fluoxetine; six weeks for MAOIs). Drug interactions with consolidation-phase peptides are minimal, as the peptides discussed are not significantly metabolized by CYP enzymes (Acevedo-Diaz et al. 2024).

6.2 LSD

LSD presents a related but pharmacologically distinct case. Modern Phase 2 research conducted primarily through the Liechti laboratory in Basel has re-established LSD as a viable candidate for anxiety and depression treatment, with a 2024 randomized controlled trial demonstrating efficacy in depression with anxiety (Holze et al. 2024a). LSD differs from psilocybin in three respects relevant to protocol design: a longer subjective duration (8–12 hours), a broader receptor profile that includes substantial activity at dopamine D₂ and adrenergic receptors, and a longer reopened plasticity window of approximately three weeks (Nardou et al. 2023; Holze et al. 2024b).

Pharmacology and plasticity-window kinetics

LSD is a partial agonist at 5-HT₂A, 5-HT₁A, 5-HT₂C, and dopamine D₂ receptors, with secondary activity at adrenergic α₁/α₂ receptors and high-affinity allosteric modulation of TrkB at affinities comparable to psilocin (Moliner et al. 2023; Nichols 2016). Comparative pharmacological studies indicate that LSD produces effects qualitatively similar to those of psilocybin but with substantially longer duration and somewhat greater cognitive and ideational engagement (Holze et al. 2022; Ley et al. 2023). Oral administration produces onset within 30–60 minutes, peak effects at 2–3 hours, and total duration of 8–12 hours (Holze et al. 2019). The reopened plasticity window extends approximately 21 days post-administration in murine assays (Nardou et al. 2023). LSD’s longer pharmacokinetic profile produces sustained sympathetic activation that warrants closer attention to autonomic priming than psilocybin protocols require.

Priming phase

The two-week priming regimen for LSD differs from the psilocybin baseline in two respects. First, intranasal Selank dosing is intensified (twice-daily 100 μg) given LSD’s prolonged sympathetic activation and the increased likelihood of fatigue, mid-session anxiety, or autonomic destabilization in extended sessions. Second, mitochondrial priming with sublingual elamipretide (10 mg daily) becomes mechanistically defensible rather than optional, given the metabolic demands of an 8–12 hour psychoactive state. Baseline cardiovascular evaluation including resting blood pressure, heart rate variability, and where available, ambulatory blood pressure monitoring should establish autonomic readiness.

Acute session augmentation

The extended duration of LSD effects creates a longer window during which acute peptide intervention may be needed. Intranasal Selank should be available for autonomic destabilization. For patients with documented blood pressure response to LSD that approaches the upper limits of clinical tolerance, low-dose elamipretide may be repeated mid-session (5 mg sublingual at the four-hour point) on a defensible if speculative basis: SS-31’s mitochondrial protection of the cardiac sarcomere has shown benefit in models of stress-induced cardiac dysfunction (Birk et al. 2014; Szeto 2014). As with psilocybin, neurotrophic peptides are reserved for consolidation.

Plasticity-window consolidation

The 21-day plasticity window is the longest among classical serotonergic psychedelics and warrants extended consolidation-phase peptide support. Daily Cerebrolysin or sublingual Dihexa for 14 days post-session, in contrast to the 7–10 day window appropriate for psilocybin, brackets the available plasticity duration. Continued BPC-157, Selank, and elamipretide at maintenance doses for the full 21-day window maintains the inflammatory, autonomic, and metabolic substrate. The longer consolidation phase increases the importance of psychotherapeutic integration during this period; the pharmacological window is open longer than the typical structure of weekly therapy can fully exploit, and twice-weekly integration sessions for the first two weeks post-dosing align better with the underlying biology.

Reintegration and risk profile

LSD contraindications are similar to those of psilocybin, with additional caution for patients with hypertension, coronary artery disease, or stimulant-sensitive cardiac arrhythmias. Concomitant SSRI or MAOI therapy requires the same washout considerations as for psilocybin, with the additional note that recent work demonstrates that paroxetine pre-administration alters LSD pharmacokinetics in clinically meaningful ways (Thomann et al. 2025), suggesting that residual SSRI exposure during consolidation-phase peptide therapy may interact with the underlying neurotrophic substrate.

6.3 DMT and Ayahuasca

N,N-dimethyltryptamine (DMT) presents the protocol architecture with its kinetic extreme. Smoked or vaporized DMT produces a 6–15 minute experience of intense perceptual destabilization; intravenous administration extends this window modestly; ayahuasca, the Amazonian brew that combines DMT with the monoamine oxidase inhibitor (MAOI) harmaline, produces an extended 4–6 hour experience by preventing first-pass metabolism of orally ingested DMT (Riba et al. 2003; Palhano-Fontes et al. 2019). The reopened plasticity window for DMT alone is correspondingly brief, approximately 6–24 hours, but ayahuasca’s prolonged kinetic profile likely extends this to a duration comparable with short-acting tryptamine analogs (Cameron et al. 2018). The protocol architecture must therefore distinguish sharply between DMT monotherapy and ayahuasca administration.

Pharmacology and plasticity-window kinetics

DMT is an agonist at 5-HT₂A, 5-HT₁A, and sigma-1 receptors, with the sigma-1 activity contributing substantially to its neuroprotective and immunomodulatory effects (Inserra 2018; Frecska et al. 2013; Szabo et al. 2014). Sigma-1 activation reorganizes G-protein-coupled receptor heteromer complexes, modulates intracellular calcium homeostasis at mitochondria-associated endoplasmic reticulum membranes, and engages the antiapoptotic and anti-amnesic effects relevant to traumatic memory processing. Vaporized DMT produces onset within seconds, peak effects at 1–3 minutes, and total duration of 6–15 minutes (Timmermann et al. 2019). Ayahuasca administration produces onset within 30–60 minutes and total duration of 4–6 hours, with the harmaline component extending DMT exposure and adding 5-HT₂A-independent effects of its own (Palhano-Fontes et al. 2019). A Phase 2 RCT in treatment-resistant depression demonstrated rapid antidepressant effects of a single ayahuasca session that persisted for at least seven days (Palhano-Fontes et al. 2019).

Priming phase

The brevity of DMT’s plasticity window concentrates therapeutic leverage into a small temporal envelope, increasing the importance of optimal substrate preparation. Two weeks of full priming peptide stack is standard, with two distinctive modifications. First, sigma-1 supportive peptides (for which no clinically validated agent currently exists, but where mitochondrial peptide elamipretide functions as a mechanistically aligned proxy through its protection of the same mitochondria-associated ER membrane systems engaged by sigma-1) should be present throughout priming. Second, for ayahuasca specifically, the harmaline-mediated MAOI activity necessitates strict dietary tyramine restriction during priming and for at least 24 hours post-administration, and any monoaminergic peptide adjunct (oxytocin, Selank) should be reviewed for MAOI compatibility before use (most are compatible).

Acute session augmentation

For DMT alone, the acute window is too brief for meaningful peptide co-administration during the experience itself. Pre-session intranasal Selank (100 μg, 30 minutes prior) attenuates pre-onset autonomic arousal without affecting the experience; pre-session sublingual elamipretide (10 mg, 60 minutes prior) supports the mitochondrial-proteostatic axis during the brief but intense state. For ayahuasca, the longer duration permits a mid-session sublingual elamipretide redose (5 mg at the three-hour point) where indicated. Direct neurotrophic peptides remain reserved for consolidation.

Plasticity-window consolidation

The brevity of DMT’s plasticity window, 6 to 24 hours, collapses the consolidation phase into a single intensive day. A single Cerebrolysin dose (5 mL parenteral, or maximum-tolerated OTF/intranasal equivalent) administered within 4–8 hours of the session maximizes neurotrophic reinforcement during the open window. Sublingual Dihexa (16 mg) at 12 hours post-session and again at 24 hours provides a second window of c-Met activation. Maintenance doses of priming-phase peptides continue for 7 days. For ayahuasca, the consolidation phase extends to approximately one week, and the consolidation regimen approximates that of short-acting psilocybin protocols, with daily Cerebrolysin or Dihexa for 5–7 days. The compressed timing of DMT consolidation places premium on rapid-onset peptide formulations and on coordinating peptide administration with the timing of integration psychotherapy.

Reintegration and risk profile

DMT and ayahuasca contraindications include personal or family history of psychotic disorder, severe cardiovascular disease, and for ayahuasca specifically, all serotonergic medications (SSRIs, SNRIs, MAOIs, triptans, tramadol) and tyramine-containing foods given the harmaline MAOI activity. The narrow plasticity window means that booster sessions, if clinically indicated, can be scheduled at relatively short intervals, 4–8 weeks rather than the 3–6 month intervals appropriate for psilocybin, but the cumulative effect of repeated DMT exposure on baseline 5-HT₂A receptor density and on long-term sigma-1 function is incompletely characterized.

6.4 5-MeO-DMT

5-Methoxy-N,N-dimethyltryptamine (5-MeO-DMT) is pharmacologically and phenomenologically distinct from DMT and from classical serotonergic psychedelics. Where psilocybin and LSD privilege symbolic and autobiographical reorganization, 5-MeO-DMT operates at the level of self-modeling itself, often producing nondual states characterized by collapse of subject-object distinction. Recent clinical development through Beckley Psytech (intranasal BPL-003) and GH Research (vaporized GH001) has produced Phase 1/2 evidence of rapid antidepressant effects in treatment-resistant depression with single-session administration (Reckweg et al. 2023; Rucker et al. 2024).

Pharmacology and plasticity-window kinetics

5-MeO-DMT is a high-affinity 5-HT₁A agonist with secondary 5-HT₂A activity, a receptor profile that distinguishes it from classical psychedelics, in which 5-HT₂A activity dominates (Reckweg et al. 2022). The 5-HT₁A engagement contributes to several distinctive features: decreased locomotor activity in animal models, prolactin release with downstream anti-inflammatory effects, and the characteristic ego-dissolutional phenomenology that differs qualitatively from the symbolic phenomenology of 5-HT₂A-dominant compounds. Vaporized administration produces onset within 5–10 seconds, peak effects at 1–3 minutes, and total duration of 5–30 minutes (Reckweg et al. 2022). Intranasal benzoate formulations extend duration modestly. The plasticity window has not been directly characterized in critical-period assays but is presumed to be in the range of DMT given the comparable pharmacokinetic profile, though the dominant 5-HT₁A activity may produce qualitatively different downstream effects on plasticity-related signaling than 5-HT₂A-dominant tryptamines.

Priming phase

The two-week priming stack for 5-MeO-DMT differs from that for DMT in one principal respect: greater attention to autonomic and emotional substrate, given the intensity of identity-level destabilization that characterizes the experience. Baseline psychological assessment for dissociative tendency, adverse childhood experiences, and identity stability is particularly important; patients with severe identity disturbance or active dissociative disorders are not appropriate candidates regardless of peptide priming. Intranasal Selank dosing should be intensified relative to standard psilocybin priming (twice-daily 100 μg for two weeks). The mitochondrial-proteostatic axis is supported by daily sublingual elamipretide (10 mg).

Acute session augmentation

The brevity of 5-MeO-DMT effects, even for intranasal formulations, limits intra-session peptide intervention. Pre-session intranasal Selank (100 μg, 30 minutes prior) and sublingual elamipretide (10 mg, 60 minutes prior) provide the principal preparatory pharmacological frame. The intensity of nondual experience and its frequent association with abreaction or somatic discharge warrants close clinical monitoring during and immediately after administration; rescue intranasal Selank (75–150 μg) should be available for post-emergent anxiety.

Plasticity-window consolidation

The presumed 12–48 hour consolidation window for 5-MeO-DMT calls for a compressed but intensive neurotrophic stack. A single Cerebrolysin dose within 6 hours of administration, followed by sublingual Dihexa (16 mg) at 24 and 48 hours, brackets the available window. Given the ontological character of 5-MeO-DMT phenomenology and the frequent reports of post-acute disorientation as patients reconstruct identity-level coherence, integration-phase peptide pairings should include particular attention to cognitive stabilization. Low-dose intranasal Semax (250 μg twice daily for 5–7 days) supports cognitive and dopaminergic recovery during this period.

Reintegration and risk profile

5-MeO-DMT contraindications are stricter than those of classical psychedelics. In addition to standard contraindications (personal or family history of psychotic disorder, severe cardiovascular disease, concomitant serotonergic medications), patients with severe identity disturbance, active dissociative disorders, or unstable adverse childhood experience profiles are not appropriate candidates. The intensity of the experience and the frequency of post-acute disorientation make 5-MeO-DMT particularly demanding of integration psychotherapy; pharmacological consolidation cannot substitute for the structured therapeutic frame within which identity reconstitution occurs.

6.5 MDMA

MDMA represents the protocol architecture’s most clinically advanced application and its most peptidergically rich pairing space. Phase 3 evidence from the Lykos Therapeutics (formerly MAPS Public Benefit Corporation) program demonstrated substantial efficacy in severe and treatment-resistant PTSD, with two-thirds of treated participants no longer meeting diagnostic criteria at the primary endpoint (Mitchell et al. 2021, 2023). The 2024 FDA Complete Response Letter declined approval pending additional data on study design and functional unblinding (FDA 2024), but the underlying clinical signal remains among the strongest in psychiatric medicine. MDMA’s distinctive peptidergic shadow, primarily oxytocinergic and vasopressinergic, makes the compound the natural anchor for the relational portal protocol.

Pharmacology and plasticity-window kinetics

MDMA acts primarily as a substrate at the serotonin, norepinephrine, and dopamine transporters, releasing these monoamines through reverse transport, with secondary direct activity at 5-HT₁A, 5-HT₂A, and 5-HT₂C receptors (Liechti 2014). The prosocial and empathogenic features of the experience are mediated in part by oxytocin and vasopressin release, with plasma oxytocin elevations correlating with subjective and behavioral measures of social bonding (Dumont et al. 2009; Kuypers et al. 2014). Acute amygdala reactivity is reduced and ventromedial prefrontal cortex–amygdala connectivity enhanced, creating the neural substrate for fear extinction and memory reconsolidation in the relational frame (Carhart-Harris et al. 2014; Mithoefer et al. 2019). The reopened plasticity window is approximately 14 days, with the social reward learning critical period requiring oxytocin receptor activation in the nucleus accumbens (Nardou et al. 2019, 2023). Oral administration produces onset within 30–60 minutes, peak effects at 1.5–2.5 hours, and total duration of 4–6 hours.

Priming phase

The two-week priming stack for MDMA is the most peptide-dense in the protocol architecture, reflecting both the metabolic demands of MDMA’s mechanism (substantial monoamine release and mitochondrial stress) and the centrality of the oxytocinergic axis. Daily sublingual BPC-157 (500 μg) and intranasal Selank (twice-daily 100 μg) provide standard inflammatory and autonomic preparation. Daily sublingual elamipretide (10 mg) is mechanistically warranted given documented MDMA mitochondrial liability (Capela et al. 2009). The priming phase is the optimal window for establishing baseline oxytocin tone in patients with PTSD-associated oxytocin deficiency (Carmassi et al. 2021). Twice-daily intranasal oxytocin (24 IU, morning and evening) for the final 7 days of priming sensitizes the oxytocinergic system and may support attachment-relevant relational work in preparatory psychotherapy. A pre-session cardiovascular evaluation is mandatory given MDMA’s sympathomimetic profile.

Acute session augmentation

The MDMA dosing session is the protocol’s clearest case for active intra-session peptide co-administration. Intranasal oxytocin (24–40 IU) administered approximately 30 minutes before MDMA onset reinforces the attachment-related neurochemistry of the relational portal at the moment of maximum therapeutic leverage (Nardou et al. 2019; Mithoefer et al. 2019). The timing must be calibrated so that peripheral oxytocin peak coincides with the early phase of MDMA effects rather than preceding it. Pre-session sublingual elamipretide (10 mg, 60 minutes prior) provides mitochondrial protection during the period of maximum monoamine release and oxidative stress; this is one of the more strongly mechanistically supported pairings in the protocol architecture. Intranasal Selank should be available throughout the session for autonomic destabilization. The standard MDMA-assisted therapy session structure of three doses spaced approximately one month apart should be retained; peptide priming is repeated before each dosing session, but the inflammatory and metabolic substrate established during the first cycle persists into subsequent cycles.

Plasticity-window consolidation

The 14-day plasticity window for MDMA is the most clinically validated of any compound in the protocol architecture. Twice-daily intranasal oxytocin (24 IU, morning and evening) for the full 14-day window sustains attachment-related learning during the period of maximum relational plasticity. Daily Cerebrolysin (5 mL) or sublingual Dihexa (12 mg) for the first 7 days reinforces neurotrophic remodeling. Continued sublingual elamipretide (10 mg) for the full 14 days addresses the prolonged oxidative-mitochondrial recovery that follows substantial monoamine release. BPC-157 and Selank continue at maintenance doses. The relational consolidation framework is itself novel within the published MDMA-assisted therapy literature and represents a substantive protocol extension beyond current clinical practice.

Reintegration and risk profile

MDMA contraindications include severe cardiovascular disease, history of cerebrovascular accident, uncontrolled hypertension, hepatic dysfunction, and personal or family history of psychotic disorder. Concomitant SSRI or MAOI therapy requires extended washout (minimum 2 weeks for most SSRIs, 6 weeks for fluoxetine, 6 weeks for MAOIs) given documented serotonin syndrome risk. Neuroendocrine recovery from acute MDMA effects, particularly recovery of normal serotonin transporter density and 5-HT₂A receptor function, may require 4–8 weeks; second and third dosing sessions at approximately one-month intervals should be timed accordingly. The peptide stack interacts minimally with standard MDMA pharmacokinetics, but oxytocin co-administration should be reviewed for cardiovascular tolerability in patients with any significant hemodynamic vulnerability.

6.6 Ibogaine

Ibogaine occupies the protocol architecture’s most demanding position. Its plasticity window is the longest among compounds reviewed (4+ weeks), its therapeutic profile in opioid use disorder is among the most clinically distinctive in psychedelic medicine, and its safety liability, principally hERG channel blockade with associated QT prolongation and risk of fatal cardiac arrhythmia, is the most serious. Recent work in special operations forces veterans with traumatic brain injury and PTSD has demonstrated dramatic clinical effects of structured ibogaine administration with magnesium co-administration for cardiac protection (Cherian et al. 2024). The protocol developed here treats ibogaine as a high-stakes intervention requiring the most intensive priming, the most aggressive intra-session cardiac protection, and the longest consolidation window in the architecture.

Pharmacology and plasticity-window kinetics

Ibogaine is a multi-target compound with weak 5-HT₂A activity, NMDA receptor antagonism, kappa-opioid agonism, sigma receptor activity, monoamine transporter modulation, and hERG potassium channel blockade (Koenig and Hilber 2015; Mash et al. 2018). Its principal active metabolite, noribogaine, accumulates in tissue with a half-life of 28–49 hours and accounts for the prolonged behavioral and plasticity-related effects (Mash et al. 2016). Ibogaine increases GDNF expression in the ventral tegmental area, contributing to its anti-addictive properties (Carnicella et al. 2010). The reopened plasticity window extends 28+ days post-administration in murine assays, the longest measured in the Nardou et al. (2023) comparative work. Oral administration of typical clinical doses (10–20 mg/kg) produces onset within 30–60 minutes, peak effects at 2–4 hours, and total duration of 24–36 hours, with noribogaine-mediated effects continuing for several days (Brown and Alper 2018).

Priming phase

The two-week ibogaine priming regimen is the most intensive in the protocol architecture and includes mandatory cardiac evaluation. Baseline 12-lead ECG, 24-hour Holter monitoring, electrolyte panel including magnesium, and where clinically indicated cardiac MRI or echocardiography are required to exclude long QT syndrome, structural heart disease, or electrolyte imbalance. Patients with QTc above 450 ms, family history of sudden cardiac death, or any documented arrhythmia are not appropriate candidates regardless of peptide protection. Beyond standard priming peptides (BPC-157, Selank, thymosin-α1 as indicated), the ibogaine priming stack adds intensified mitochondrial and cardiac protection: daily sublingual elamipretide (10–15 mg) for the full two weeks, with the higher dose justified by both cardiac mitochondrial protection and noribogaine’s prolonged tissue presence; daily oral magnesium glycinate (400 mg) and potassium gluconate (10 mEq); and where biomarkers indicate elevated systemic inflammation, daily sublingual KPV (500 μg). For patients with opioid use disorder, the ibogaine indication most clearly supported by clinical evidence, opioid withdrawal must be appropriately managed during priming using non-NMDA-antagonist agents.

Acute session augmentation

The ibogaine session is the protocol’s most actively managed acute window. Pre-session sublingual elamipretide (15 mg, 90 minutes prior) and intravenous magnesium sulfate (2 g over 20 minutes, 30 minutes prior) establish the cardiac-protective frame. Continuous cardiac monitoring throughout the session is mandatory. Mid-session elamipretide (10 mg sublingual at the four-hour point) provides additional mitochondrial protection during the period of maximum noribogaine accumulation. Cibinetide (ARA-290), an erythropoietin-derived peptide engaging the innate repair receptor without erythropoietic effects, has shown cardioprotective and anti-arrhythmic properties in preclinical models (Brines et al. 2008) and represents a defensible if speculative additional cardiac adjunct (4 mg subcutaneous, single dose pre-session) where standard cardiac protection is inadequate. The autobiographical-moral character of the ibogaine experience requires close therapeutic presence throughout; intra-session pharmacological intervention should be minimal beyond cardiac protection.

Plasticity-window consolidation

The 28+ day ibogaine plasticity window is the longest in the protocol architecture and warrants the most extensive consolidation regimen. Daily Cerebrolysin (5 mL) or sublingual Dihexa (16 mg) for the first 14 days, then alternating-day dosing for days 15–28, brackets the plasticity envelope. Continued elamipretide (10 mg daily) for the full 28 days addresses prolonged mitochondrial recovery from noribogaine exposure. Low-dose intranasal Semax (500 μg twice daily for 14 days) supports cognitive and dopaminergic recovery during the autobiographical reorganization that characterizes post-ibogaine clinical experience (Eremin et al. 2005). For patients in addiction recovery, the consolidation phase aligns with the period of maximum craving-related reward circuitry remodeling and represents the principal therapeutic window of the intervention.

Reintegration and risk profile

Beyond the standard psychedelic contraindications, ibogaine-specific contraindications are extensive: any cardiac conduction abnormality, structural heart disease, electrolyte imbalance, hepatic dysfunction, concurrent opioid maintenance therapy without appropriate transition protocol, and concomitant use of any QT-prolonging medication (a category that includes many psychiatric and antiemetic agents). Ibogaine should be administered only in settings with continuous cardiac monitoring, immediate access to advanced cardiac life support, and clinical staff trained in management of arrhythmia. The 28+ day consolidation phase places sustained demand on integration psychotherapy; the autobiographical-moral character of ibogaine reorganization is poorly served by abbreviated post-session support, and integration sessions twice weekly for the first month are recommended.

6.7 Ketamine and Esketamine

Ketamine occupies a paradoxical position in the contemporary psychedelic landscape. Its FDA-approved congener esketamine (Spravato) has been clinically available for treatment-resistant depression since 2019; ketamine itself is widely used off-label in dose ranges from sub-anesthetic (0.5 mg/kg IV) to anesthetic (1–2 mg/kg IV); and racemic ketamine has demonstrated rapid antidepressant effects within hours of administration in multiple controlled trials (Wilkinson et al. 2018; Daly et al. 2019). The compound’s plasticity window of approximately 48 hours is the shortest among the compounds reviewed, and its dissociative rather than serotonergic-classical phenomenology produces a different therapeutic process. The protocol architecture for ketamine emphasizes rapid, intensive peptide augmentation during the brief consolidation window.

Pharmacology and plasticity-window kinetics

Ketamine is an NMDA receptor antagonist with secondary activity at sigma-1, mu-opioid, and HCN1 channels (Krystal et al. 1994; Williams et al. 2018). Its rapid antidepressant effects depend on a downstream signaling cascade involving AMPA receptor activation, mTOR engagement, and BDNF/TrkB signaling that produces dendritogenesis and increased synaptic spine density within 24–48 hours of administration (Abdallah et al. 2018; Li et al. 2010). Subanesthetic ketamine reactivates adult cortical plasticity through PNN-modulating mechanisms involving microglial PNN disassembly (Venturino et al. 2021; Grieco et al. 2020). Intravenous administration produces onset within 5–10 minutes, peak effects at 30–40 minutes, and total duration of 1–2 hours. The reopened plasticity window of approximately 48 hours is the shortest among the compounds reviewed (Nardou et al. 2023).

Priming phase

Ketamine protocols often serve patient populations with severe and acute symptomatology (treatment-resistant depression with active suicidality, severe PTSD with imminent risk) for which extended priming is impractical. A compressed 5–7 day priming regimen is appropriate for most clinical use. Daily sublingual BPC-157 (500 μg) and intranasal Selank (100 μg twice daily) prepare the inflammatory and autonomic substrate. Mitochondrial priming with elamipretide (10 mg daily) is mechanistically aligned but optional for ketamine specifically given the compound’s relatively favorable cardiovascular profile at sub-anesthetic doses. Where ketamine is being used in series, repeated administrations over weeks or months, priming peptide regimens may be sustained between sessions rather than discontinued and re-initiated.

Acute session augmentation

The brief ketamine session permits no meaningful intra-session peptide intervention beyond pre-session preparation. Sublingual elamipretide (5 mg, 30 minutes prior) and intranasal Selank (100 μg, 30 minutes prior) establish the pre-session frame. The dissociative experience produces minimal autonomic destabilization in most patients but should be monitored throughout.

Plasticity-window consolidation

The 48-hour ketamine plasticity window is the protocol architecture’s most compressed consolidation phase and demands rapid, intensive intervention. A single high-dose Cerebrolysin administration (5–10 mL) within 4–8 hours of the session maximizes neurotrophic reinforcement during peak dendritogenesis. Sublingual Dihexa (16 mg) at 12, 24, and 36 hours post-session provides repeated c-Met activation across the plasticity window. Continuation of priming-phase peptides through 72 hours sustains the substrate during initial reintegration. The compressed timing makes ketamine the protocol most suited to acute intensive intervention in inpatient or partial-hospital settings, where peptide administration can be precisely timed to the underlying neurobiology.

For ketamine series treatment, where weekly or biweekly administrations are standard, the consolidation peptide stack can be administered immediately post-session at each dosing without the gap that would characterize psilocybin or LSD protocols. The cumulative neurotrophic reinforcement across multiple sessions may produce additive effects beyond what single-session protocols can achieve, though this remains untested.

Reintegration and risk profile

Ketamine contraindications include uncontrolled hypertension, history of psychosis or severe dissociative disorder, recent myocardial infarction or stroke, and concomitant use of certain medications including some MAOIs and stimulants. The rapid onset and offset of ketamine effects produces less sustained autonomic stress than longer-duration psychedelics, but cumulative ketamine exposure (more than 8–12 sessions in series) has been associated with bladder dysfunction (ketamine cystitis) and cognitive complaints; peptide protocols should not be used to enable indefinite extension of ketamine series treatment. Drug interactions with the consolidation peptide stack are minimal.

6.8 Mescaline

Mescaline, the principal active alkaloid of peyote (Lophophora williamsii) and San Pedro (Echinopsis pachanoi) cacti, has been the most clinically neglected of the classical serotonergic psychedelics despite a multi-millennium history of ceremonial use and a clinical pharmacology profile distinct from psilocybin and LSD (Cassels and Sáez-Briones 2018; Ley et al. 2023). Recent comparative work has demonstrated that mescaline produces subjective effects similar in character to those of psilocybin and LSD but with substantially longer duration and somewhat distinct affective tone (Ley et al. 2023). Indigenous ceremonial use of peyote, particularly within the Native American Church, must be respected as a sovereign tradition that exists outside and prior to the clinical-medical frame; the protocol developed here applies only to clinical or research administration and does not address ceremonial contexts.

Pharmacology and plasticity-window kinetics

Mescaline is a phenethylamine rather than a tryptamine, with primary 5-HT₂A agonism and significant secondary activity at 5-HT₂C and 5-HT₁A receptors (Nichols 2016). It produces less potent direct dopamine and adrenergic effects than LSD and produces minimal cardiovascular activation at typical doses. Oral administration of typical clinical doses (300–500 mg) produces onset within 60–90 minutes, peak effects at 4–6 hours, and total duration of 10–14 hours (Ley et al. 2023). The plasticity window has not been directly characterized in critical-period assays but is likely comparable to that of LSD given the similar duration of action.

Priming phase

Two-week mescaline priming follows the LSD baseline given the comparable duration of action. The longer subjective duration warrants the more intensive autonomic preparation appropriate to LSD: twice-daily intranasal Selank at 100 μg, daily mitochondrial support with sublingual elamipretide at 10 mg, standard inflammatory peptides. The lower cardiovascular activation of mescaline relative to LSD reduces the indication for cardiac-specific intervention, but the longer duration sustains metabolic demand.

Acute session augmentation

The 10–14 hour mescaline session is among the longest in the protocol architecture and creates an extended window during which acute peptide intervention may be needed. Pre-session intranasal Selank (100 μg) and sublingual elamipretide (10 mg) establish the preparatory frame. A mid-session sublingual elamipretide redose (5 mg at the six-hour point) is mechanistically defensible given the metabolic demands of an extended psychoactive state.

Plasticity-window consolidation

A 14–21 day consolidation window is appropriate for mescaline, paralleling the LSD architecture. Daily Cerebrolysin or sublingual Dihexa for 14 days, with continuation of inflammatory and metabolic adjuncts through the full 21-day window, brackets the plasticity envelope. The longer subjective duration of mescaline often produces sustained post-acute affective changes lasting several days that warrant extended integration psychotherapy.

Reintegration and risk profile

Mescaline contraindications follow the standard psychedelic profile, with the addition of caution regarding the prolonged duration of action and corresponding requirement for extended clinical supervision during the acute session. Concurrent SSRI or MAOI therapy requires standard washout. The clinical evidence base for mescaline is substantially smaller than that for psilocybin or LSD, and protocols developed here rest more heavily on extrapolation from related compounds than is the case for the better-studied psychedelics.

6.9 Emerging Psychoplastogens, Entactogens, and Atypical Compounds

A growing class of compounds promotes structural neural plasticity through diverse receptor mechanisms while reducing or eliminating the hallucinogenic experience characteristic of classical serotonergic psychedelics. Their emergence forces a conceptual revision of the protocol architecture developed in this review: if therapeutic plasticity can be engaged without the experiential content that has organized contemporary psychedelic-assisted psychotherapy, the relationship between pharmacology and integration psychotherapy must be reconsidered. Three compounds illustrate the range of this emerging class (tabernanthalog, methylone, and salvinorin A) and represent three distinct mechanistic strategies for separating therapeutic effect from hallucinogenic experience.

Tabernanthalog (TBG, DLX-007)

Tabernanthalog, developed by the Olson laboratory at the University of California, Davis, and licensed to Delix Therapeutics, is a water-soluble, non-hallucinogenic, non-cardiotoxic structural analog of ibogaine engineered through function-oriented synthesis to retain ibogaine’s therapeutic pharmacophore while eliminating hERG channel binding (Cameron et al. 2021). TBG promotes structural neural plasticity in the prefrontal cortex through TrkB/mTOR/AMPA signaling, reduces alcohol- and heroin-seeking behavior, and produces antidepressant-like effects in rodent stress models with neuroplastic changes persisting up to 12 days (Cameron et al. 2021). As of January 2026, tabernanthalog had not yet entered clinical trials, but the preclinical data position it as the most advanced ibogaine-derived psychoplastogen in development.

For TBG and structurally related ibogalogs, the protocol architecture simplifies substantially. The elimination of hERG-mediated cardiotoxicity removes the most demanding feature of the ibogaine cardiac protection regimen. The presumed plasticity window of approximately 12 days, comparable to that of psilocybin, suggests that the consolidation-phase peptide stack can follow the psilocybin template rather than the extended 28-day ibogaine architecture. Standard priming peptides (BPC-157, Selank, elamipretide at maintenance doses) remain mechanistically aligned. The principal protocol implication is the absence of an experiential crisis-and-resolution arc requiring intensive integration psychotherapy. Where ibogaine produces a structured autobiographical life review that integration psychotherapy must contain and translate into behavioral change, TBG produces no such experiential content and the therapeutic process becomes closer to enhancement of conventional cognitive-behavioral or behavioral therapy than to psychedelic-assisted psychotherapy in its current form.

Methylone (TSND-201)

Methylone, the β-ketone analog of MDMA developed by Transcend Therapeutics under the designation TSND-201, has emerged as the first non-hallucinogenic neuroplastogen to demonstrate efficacy in a randomized placebo-controlled clinical trial. The IMPACT-1 Phase 2 study, published in JAMA Psychiatry in 2026, demonstrated statistically significant and clinically meaningful reductions in PTSD symptoms in 65 patients with severe PTSD (CAPS-5 ≥ 35), all of whom had previously failed at least one PTSD treatment (Jones et al. 2025; Transcend Therapeutics 2026). Patients received four oral doses spaced one week apart, with placebo-adjusted CAPS-5 reductions of -8.00 by day 10 and -9.64 sustained through day 64. The U.S. Food and Drug Administration granted Breakthrough Therapy designation for TSND-201 in PTSD in July 2025 (Transcend Therapeutics 2025), and Phase 3 development is now underway. Methylone therefore represents the first compound in the class to bridge from preclinical psychoplastogen status to clinical evidence sufficient for regulatory review.

Pharmacology

Methylone acts primarily as a substrate at the serotonin, norepinephrine, and dopamine transporters with greater selectivity for these targets than MDMA exhibits, and with no significant activity at the 5-HT₂A receptor (Warner-Schmidt et al. 2022; Jones et al. 2025). The absence of 5-HT₂A engagement removes the hallucinogenic experiential component while preserving the empathogenic-entactogenic effects mediated by serotonergic and dopaminergic monoamine release. Acute oral administration produces onset within 15–30 minutes, peak effects at 1–2 hours, and total duration of approximately 3–4 hours, with a more compressed effect profile than MDMA. Preclinical work demonstrates that methylone promotes neurite outgrowth, produces robust anxiolytic and antidepressant-like effects, and influences fear extinction learning through mechanisms that overlap mechanistically with MDMA’s effects on amygdala-prefrontal connectivity (Warner-Schmidt et al. 2022; Li et al. 2024). The plasticity window has not been characterized in critical-period assays but the durability of the IMPACT-1 clinical effects through day 64 after only four doses suggests a window of sufficient duration to support therapeutic learning.

Protocol implications

Methylone’s clinical position has three implications for the protocol architecture. First, the absence of hallucinogenic experience eliminates the requirement for intensive intra-session therapeutic containment characteristic of MDMA-assisted therapy. The IMPACT-1 protocol involved standard supportive care during dosing without the elaborate dyad-based therapy structure that has defined MDMA development. Patients can be treated in a clinic setting comparable to that used for ketamine infusion therapy, dramatically reducing the resource intensity of the intervention. The peptide pairings appropriate to methylone reflect this simpler structure: standard priming with BPC-157, Selank, and where indicated elamipretide; minimal intra-session intervention beyond preparatory peptide doses; and consolidation-phase neurotrophic support timed to the four weekly dosing intervals.

Second, the multi-dose protocol structure (four doses spaced one week apart) creates a peptide stacking opportunity that single-dose psychedelic protocols do not afford. Continuous peptide priming and consolidation across the four-week dosing cycle, rather than the intermittent priming-acute-consolidation-reintegration sequence appropriate to single-dose psilocybin or MDMA, may better align with the cumulative pharmacological effect of methylone. The plasticity-window peptide stack (neurotrophic peptides, oxytocinergic support, mitochondrial maintenance) can be administered continuously across the four-week treatment window with brief intensification on each dosing day.

Third, methylone’s elevation of oxytocin, while less pronounced than that of MDMA (Warner-Schmidt et al. 2022), is mechanistically aligned with the relational portal framework. Twice-daily intranasal oxytocin (24 IU) for the four-week dosing cycle, beginning one week before the first dose and continuing through the consolidation phase following the final dose, represents a defensible pairing for PTSD applications. The peptide stack is otherwise simpler than that for MDMA, given methylone’s more favorable cardiovascular and oxidative profile. Mitochondrial support with elamipretide remains mechanistically defensible but can be deployed at lower doses than for MDMA. Cibinetide (ARA-290) and other cardiac-protective peptides are not indicated.

The clinical implications of methylone’s emergence extend beyond its specific pharmacology. As the first non-hallucinogenic neuroplastogen to achieve regulatory milestone status, it validates the proposition that the therapeutic mechanism of empathogen-entactogen pharmacology can be separated from the experiential content that has organized contemporary clinical practice. If Phase 3 development confirms IMPACT-1’s findings, methylone may shift the center of gravity in psychedelic medicine from experiential intensity toward pharmacological precision, with corresponding consequences for the role of integration psychotherapy and the structure of clinical delivery. Peptide co-administration in this frame becomes less an adjunct to a complex psychological process and more a component of a precision pharmacological protocol, a development that has substantial implications for clinical scaling and for the equity of access.

Salvinorin A

Salvinorin A, the principal active diterpenoid of the Mexican mint Salvia divinorum, occupies a singular position in the contemporary psychedelic landscape. It is the only known naturally-occurring, non-nitrogenous selective high-efficacy agonist at the kappa-opioid receptor (KOR), with no significant activity at mu- or delta-opioid receptors and no direct serotonergic activity (Roth et al. 2002; Butelman and Kreek 2015). Its phenomenology is sharply distinct from that of classical psychedelics: brief (5–15 minutes when smoked), often dysphoric, characterized by intense dissociation, perceptual fragmentation, and frequent autobiographical regression, with anhedonic and aversive components mediated by KOR-induced dopamine suppression in nucleus accumbens (Carlezon et al. 2006). Despite its difficult phenomenological profile, salvinorin A has demonstrated anti-addictive effects in preclinical models of cocaine and alcohol use, antidepressant-like effects in some rodent paradigms, and substantial neuroprotective effects against hypoxia-ischemia injury (Coffeen and Pellicer 2019; Polepally et al. 2021; Korostynski et al. 2024). A 2025 systematic review and meta-analysis of preclinical studies confirmed therapeutic potential across these indications while emphasizing the limited and mechanistically constrained nature of available human data (Korostynski et al. 2024).

Pharmacology and clinical position

Salvinorin A is a high-affinity, high-efficacy KOR agonist with secondary D₂ dopamine receptor activity (Seeman et al. 2009) and substantial substrate activity at P-glycoprotein efflux transporters at the blood-brain barrier, the latter contributing to its short cerebral residence time (Butelman et al. 2009). Inhaled or vaporized administration produces onset within seconds, peak effects at 1–3 minutes, and total duration of 5–15 minutes; sublingual administration produces longer but less reliable effects given the pH-sensitivity of buccal absorption. Human clinical research has been limited, with controlled studies demonstrating reproducible psychoactive effects but not yet establishing therapeutic efficacy in formal trials (Johnson et al. 2011; Maqueda et al. 2015). The kappa-opioid mechanism is paradoxical for depression, since KOR antagonists rather than agonists are the class being developed as antidepressants (Carlezon and Krystal 2016), but salvinorin A may have specific utility in addiction where KOR-mediated dopamine suppression directly counters the dopaminergic reward of stimulant reinforcement.

Protocol implications

The clinical frame for salvinorin A is fundamentally different from that for serotonergic psychedelics or entactogen-empathogens, and the protocol architecture must reflect this difference. The brief duration of the experience, the dysphoric phenomenological character, and the KOR-specific mechanism all argue against extended priming or consolidation regimens patterned on classical psychedelic protocols. The therapeutic frame, where it can be defended at all, is closer to interventional pharmacology for specific addiction indications, particularly cocaine or methamphetamine use disorder, than to psychedelic-assisted psychotherapy in the contemporary sense.

For research applications in stimulant use disorder, the relevant peptide pairings differ substantially from those developed for serotonergic psychedelics or MDMA. Mitochondrial protection during the brief but intense KOR activation, supported by sublingual elamipretide (10 mg pre-session), addresses the energetic demands of the acute state. The post-acute dopaminergic suppression characteristic of KOR activation may benefit from intranasal Semax (250–500 μg) administered in the hours following exposure to support cognitive recovery and dopaminergic tone. Neuroprotective peptide stacks including Cerebrolysin or sublingual Dihexa, mechanistically aligned with the neuroprotective signal observed in salvinorin A preclinical data, may be deployed in a brief consolidation window of 24–72 hours. Standard inflammatory and autonomic support is appropriate but at reduced intensity given the brevity of acute exposure.

The principal contraindications for salvinorin A include any history of psychosis, severe depression with active anhedonia (where KOR agonism may worsen rather than improve symptoms), and substance use disorders involving opioid receptor systems. Concomitant opioid agonist therapy, opioid maintenance, and opioid use disorder in active or recent recovery represent absolute contraindications. The clinical research base remains insufficient to support routine clinical use; salvinorin A protocols belong to investigational research settings rather than emerging clinical practice. Its inclusion here reflects mechanistic completeness rather than a recommendation for clinical deployment.

Implications for the protocol architecture

The emergence of TBG, methylone, and salvinorin A collectively complicates the binary distinction between psychedelic and non-psychedelic that has organized contemporary research. Three different mechanistic strategies for separating therapeutic effect from hallucinogenic experience have produced three different clinical positions: the rationally engineered psychoplastogen (TBG); the entactogen with selective monoamine transporter activity but absent 5-HT₂A engagement (methylone); and the alternative-receptor-class compound that produces psychoactive effects through a non-serotonergic mechanism altogether (salvinorin A). Each of these strategies has different implications for the role of integration psychotherapy, the structure of peptide co-administration, and the clinical infrastructure required for delivery. The protocol architecture developed in this review accommodates these emerging classes through the same four-phase scaffold while requiring case-specific calibration of peptide pairings, plasticity-window assumptions, and integration intensity. Future iterations of this architecture will require periodic revision as additional non-hallucinogenic neuroplastogens reach clinical milestones.

VII. Frontier Extensions: Speculative Pairings with Mechanistic Support

The compound-specific protocols in Section VI rest on mechanistic extrapolation from established preclinical and clinical evidence at both the psychedelic and peptide ends, even where the specific psychedelic-peptide combination has not been directly tested. The protocols developed in this section step beyond that floor. They identify pairings for which mechanistic logic supports synergistic effect but for which neither component has been studied in the relevant context, and they propose peptide candidates whose clinical validation is incomplete but whose mechanistic profile addresses substrate features of psychedelic action that better-characterized peptides do not reach. These pairings are presented not as protocol recommendations but as research priorities. They name the frontier from which the next generation of compound-specific protocols will likely be built.

7.1 Klotho fragments and NMDA-coupled cognitive enhancement

The α-Klotho protein, originally identified for its role in calcium-phosphate homeostasis and longevity, has emerged as a candidate cognitive enhancer through mechanisms that do not depend on blood-brain barrier transit. Peripheral administration of an α-klotho fragment (αKL-F) produces rapid cognitive enhancement and neural resilience in young, aging, and α-synuclein transgenic mice without crossing the blood-brain barrier as measured by autoradiography or His-tagged protein studies (Dubal et al. 2014; Leon et al. 2017). The mechanism appears to involve peripheral signaling that modulates GluN2B-containing NMDA receptor function in the brain, enhancing NMDA-dependent synaptic plasticity through a pathway that remains incompletely characterized (Castner et al. 2023). Subsequent work in aged nonhuman primates demonstrated that systemic αKL-F administration improves cognitive performance on prefrontal-dependent tasks (Castner et al. 2023). In humans, the KL-VS heterozygous variant of KL is associated with reduced age-related neuroinflammation, neurodegeneration, and synaptic dysfunction in cognitively unimpaired older adults (Yokoyama et al. 2024).

The mechanistic alignment with psychedelic protocols is substantial. Classical psychedelics produce TrkB-driven dendritogenesis and increased synaptic plasticity that depend on adequate NMDA receptor function (Moliner et al. 2023; Ly et al. 2018). Ketamine produces NMDA antagonism followed by AMPA receptor activation and BDNF release. Klotho fragment co-administration during the consolidation window of either class might amplify the NMDA-dependent component of post-acute synaptic remodeling without the central side effects that direct NMDA modulators produce. The pairing is currently speculative, no clinical or preclinical study has tested klotho fragments alongside psychedelics, but the mechanistic case is among the more defensible in this section. Peripheral administration of αKL-F or recombinant secreted klotho during the consolidation window of psilocybin, LSD, ketamine, or MDMA protocols represents a research priority. The development of orally bioavailable small-molecule klotho enhancers, currently in early-stage development (Castner et al. 2023), would substantially extend the practical utility of this pairing.

7.2 Synapse-stabilizing and tau-stabilizing peptides

NAP (davunetide), an eight-amino-acid fragment of activity-dependent neuroprotective protein (ADNP), promotes microtubule stability and demonstrates neuroprotective effects in models of tauopathy and traumatic brain injury (Gozes et al. 2008; Magen and Gozes 2013). Phase 2 clinical trials of intranasal davunetide in mild cognitive impairment showed signal in some cognitive endpoints, though Phase 3 development in progressive supranuclear palsy did not meet its primary endpoint (Boxer et al. 2014). The mechanism (microtubule stabilization, reduction of tau hyperphosphorylation, and support for axonal transport) is mechanistically aligned with the structural remodeling that follows psychedelic administration. The dendritogenesis and spine density increases produced by psilocybin and LSD require functional microtubule dynamics, and the consolidation-phase synaptic stabilization that determines the durability of psychedelic effects depends on cytoskeletal stability that davunetide directly supports.

The pairing of intranasal davunetide with consolidation-phase peptide stacks, particularly for compounds with prolonged plasticity windows (LSD, ibogaine), represents a defensible if untested research priority. Davunetide’s intranasal bioavailability and its established safety profile from the failed Phase 3 program reduce the regulatory friction that otherwise constrains novel peptide combinations. For patients with comorbid neurodegenerative pathology (frontotemporal dementia, Parkinson’s disease, or post-traumatic chronic traumatic encephalopathy) the mechanistic case strengthens further, given that these conditions involve the same tau and microtubule pathology that davunetide addresses.

Cortagen, Pinealon, and other Khavinson-tetrapeptide-class agents, derived from regulatory peptides isolated from various tissues, have received less rigorous clinical evaluation but show neuroprotective and synapse-stabilizing effects in Russian and Eastern European preclinical and small clinical studies (Khavinson 2014; Khavinson et al. 2020). Their evidence base is constrained by limited Western clinical replication, but the mechanism, tetrapeptide-mediated regulation of gene expression in neural tissue, is sufficiently aligned with the consolidation-phase requirements of psychedelic protocols to warrant inclusion as research priorities.

7.3 Mitokine peptides beyond elamipretide

Elamipretide (SS-31), discussed extensively in Section IV, is the most clinically advanced of the mitochondrial-targeting peptides. Two additional mitokine peptides warrant mention as frontier candidates: MOTS-c, a 16-amino-acid peptide encoded within the mitochondrial 12S rRNA gene, which improves metabolic homeostasis, reduces insulin resistance, and increases exercise capacity in preclinical models (Lee et al. 2015; Reynolds et al. 2021); and humanin, a 24-amino-acid mitochondrial-derived peptide with anti-apoptotic and metabolic-regulatory effects (Hashimoto et al. 2001; Sreekumar et al. 2018).

Both peptides address aspects of the cellular substrate engaged by psychedelic action that elamipretide alone does not fully reach. MOTS-c regulates AMP-activated protein kinase (AMPK) signaling and metabolic gene expression in ways that complement the energetic demands of dendritogenesis and synaptic remodeling. Humanin’s anti-apoptotic effect at the cellular level may protect against the metabolic stress of prolonged 5-HT₂A activation in the high-dose classical psychedelic protocols. Neither has been tested in psychedelic contexts, and clinical-grade preparations are not currently available outside research settings. The mechanistic case for inclusion in priming and consolidation phases of high-intensity protocols (high-dose LSD, ibogaine, prolonged ayahuasca exposure) is defensible but speculative.

7.4 Hypothesized PNN-modulating peptides

The molecular mechanism by which psychedelics partially disassemble perineuronal nets to reopen critical-period plasticity (Venturino et al. 2021; Grieco et al. 2020) suggests a peptide-augmentation strategy that is more conceptual than clinical at present. Direct enzymatic disruption of PNNs through chondroitinase ABC reopens visual cortex critical-period plasticity in adult animals (Pizzorusso et al. 2002), but the enzyme is not clinically deployable as currently formulated. Peptide-based agents that modulate PNN structure or function (fragments of proteoglycan-modifying enzymes, chondroitin sulfate-binding peptides, or peptides targeting Lynx-family proteins that regulate cholinergic-dependent critical-period plasticity, per Bernard and Prochiantz 2016) represent a research frontier with substantial therapeutic potential but with no validated clinical agents yet identified. Should such agents emerge, their pairing with psychedelics would target the molecular mechanism of critical-period reopening directly rather than indirectly, potentially extending the duration or magnitude of the plasticity window beyond what current pharmacology achieves.

The Reh and colleagues review of critical-period regulation identifies several additional molecular targets (myelin-associated Nogo receptor signaling, GABAergic interneuron maturation, and excitatory-inhibitory balance regulators) for which peptide-based therapeutics are conceivable but not yet developed (Reh et al. 2020). Each represents a potential frontier for future protocol architecture.

7.5 BBB-permeant Klotho variants and engineered peptide therapeutics

The development of blood-brain barrier-permeant Klotho variants, engineered through fusion with transferrin receptor antibodies or other BBB-shuttle technologies, would substantially extend the practical utility of the klotho-NMDA pairing. Comparable strategies are under development for other large peptides and proteins relevant to psychedelic protocols: BDNF mimetics with improved BBB penetration (LM22A-4 and related compounds; Massa et al. 2010), growth factor-based therapeutics with central activity, and engineered cytokines with restricted CNS bioavailability. The protocol architecture will require revision as these classes mature, and several of the most potent theoretical peptide pairings (direct CNS BDNF augmentation during the psychedelic consolidation window, for instance) depend on delivery technology that does not yet exist in clinically deployed form.

7.6 GLP-1 receptor agonists at the metabolic-cognitive interface

The emergence of glucagon-like peptide-1 (GLP-1) receptor agonists (semaglutide, tirzepatide) as treatments for metabolic disease has been accompanied by growing evidence of cognitive and neuroprotective effects in preclinical models of Alzheimer’s disease and Parkinson’s disease (Athauda et al. 2017; Edison et al. 2025). A Phase 2 trial of liraglutide in Alzheimer’s disease showed signal in cognitive endpoints, and Phase 3 trials of semaglutide in early Alzheimer’s disease (the EVOKE program) were ongoing as of early 2026. The mechanism (improved insulin signaling in the brain, reduced neuroinflammation, modulation of amyloid-β clearance) overlaps mechanistically with the anti-inflammatory and metabolic substrates targeted by the priming phase of the protocol architecture.

For patients with comorbid metabolic dysfunction, type 2 diabetes, or accelerated biological aging, GLP-1 agonist therapy initiated during the priming phase and continued through reintegration represents a defensible if speculative pairing. The mechanistic alignment with psychedelic-induced neuroplasticity has not been directly tested, but the convergence on shared substrates of metabolic and inflammatory dysregulation suggests potential synergy. The pharmacokinetic considerations are substantial: GLP-1 agonists have prolonged half-lives (semaglutide approximately one week subcutaneous, longer for the depot oral formulation), and their effects on gastric emptying may alter the absorption of orally administered peptide adjuncts. Co-administration must be calibrated accordingly.

VIII. Delivery Science: Non-Invasive Routes for Stacked Peptide Protocols

The protocol architecture developed in this review is operationally meaningful only if its peptide components can be delivered through routes compatible with multi-agent stacking, repeated administration, and patient adherence in real-world clinical settings. The single largest practical impediment to bioactive peptide therapeutics in psychiatric and neurological practice has historically been the requirement for parenteral administration (subcutaneous, intramuscular, or intravenous injection) which limits patient acceptance, restricts use to clinical settings, and forecloses the kind of multi-peptide stacking that the four-phase architecture requires. Recent advances in three non-invasive delivery modalities have made compound peptide protocols increasingly feasible: oral thin-film (OTF) delivery via the buccal or sublingual mucosa; enteric-coated formulations targeting proximal small intestine absorption; and intranasal delivery for peptides whose physicochemical properties resist oral formulation.

8.1 Oral thin-film and mucoadhesive buccal delivery

Pre-gastric oral delivery via the buccal or sublingual mucosa avoids two of the principal barriers to oral peptide therapeutics: gastric acid degradation and intestinal proteolysis. The buccal epithelium, while previously regarded as relatively impermeable to large hydrophilic molecules, has been transformed into a clinically useful absorption surface through advances in mucoadhesive polymer chemistry, permeation enhancement, and protease inhibition (Maher et al. 2019; Brayden et al. 2020). The contemporary OTF platform integrates several molecular strategies operating in concert. Mucoadhesive polymer matrices, typically thiolated chitosan, hydroxypropyl methylcellulose (HPMC), and pullulan, adhere to mucin glycoproteins through covalent disulfide bond formation, anchoring the drug-delivery vehicle at the absorption surface for extended residence. Free thiol groups in the polymer matrix simultaneously inactivate cysteine proteases at the mucosal surface, reducing local proteolytic degradation. Permeation enhancers, including medium-chain fatty acids (sodium caprate, sodium caprylate), bile salts (sodium taurodeoxycholate), cyclodextrins, and EDTA, transiently fluidize membrane lipids, loosen tight junctions, and increase paracellular permeability without producing lasting tissue compromise.

A second engineering dimension addresses peptide stability and protection during buccal residence. Peptides may be partially complexed with hydroxypropyl-β-cyclodextrin to shield cleavage sites and improve solubility; molecularly dispersed within polymeric glass matrices (HPMC/PVA/pullulan) in solid-state form to halt hydrolysis and aggregation; and protected by reduced redox microenvironments containing glutathione and N-acetylcysteine to scavenge reactive oxygen species. Peptide architectures that are intrinsically protease-resistant (cyclic structures, lipidated analogs, and D-amino acid-containing variants) survive buccal residence with substantially less protection. The combination of cyclic or lipidated peptide cargo, mucoadhesive polymer matrix, and integrated permeation enhancement has produced cumulative flux increases of 50- to 100-fold over unformulated peptide in ex vivo Franz cell systems, with systemic exposure approaching that of subcutaneous injection in pilot in vivo pharmacokinetic studies (Maher et al. 2019).

A more advanced architecture employs chitosan-shelled liposomal nanocarriers (typically 150–200 nm diameter) embedded within the mucoadhesive film matrix. The liposomal core protects peptide cargo from enzymatic degradation; the chitosan shell binds mucin and transiently opens tight junctions; and the integration into the film matrix prevents wash-out by saliva. As the film hydrates during buccal residence, the nanocarriers fuse with epithelial membranes and deliver their cargo directly into intracellular lipid domains, achieving systemic exposure that approaches injection-level pharmacokinetics for cyclic or lipidated peptide cargo.

For the protocol architecture developed in this review, OTF delivery is the preferred route for BPC-157, Selank, Semax, KPV, oxytocin (where intranasal is not available), elamipretide, and a growing number of cyclic and lipidated peptide analogs. The directional bilayer film design (mucosa-facing layer containing peptide and mucoadhesive polymers, outward backing enforcing one-way diffusion) minimizes saliva washout and reduces taste exposure. Citric acid buffering maintains pH at approximately 5.5–6.0, which keeps chitosan protonated for maximum adhesion and suppresses local enzymatic kinetics.

8.2 Enteric-coated formulations and proximal small intestine delivery

For peptides whose physicochemical properties or molecular size resist OTF formulation, enteric-coated oral formulations represent the second non-invasive route. Polymer coatings (Eudragit L100, Eudragit S100, or pH-sensitive equivalents) that remain intact at gastric pH but dissolve in the proximal small intestine deliver peptide cargo to the duodenum or jejunum, where intestinal proteolysis is more manageable and where targeted permeation enhancement and protease inhibition can achieve clinically useful systemic exposure (Aguirre et al. 2016). Recent developments in salcaprozate sodium (SNAC) co-formulation, exemplified by the orally bioavailable semaglutide formulation Rybelsus (Buckley et al. 2018), demonstrate that enteric peptide delivery has reached clinical viability for some peptide classes, though absolute bioavailability remains low (typically 0.5–1.5%) compared with parenteral administration.

For the protocol architecture, enteric-coated formulations are most relevant for peptides whose mucosal flux is insufficient for OTF delivery, including some larger peptides and those with physicochemical properties unsuited to mucoadhesive matrices. Cyclodextrin complexation, lipid-based self-emulsifying drug delivery systems (SEDDS), and permeation enhancement through medium-chain fatty acid co-administration extend the range of peptides accessible through this route. The pharmacokinetic profile differs from OTF delivery in two respects relevant to protocol design: enteric absorption produces delayed onset (typically 2–4 hours from administration to peak plasma concentration) and may exhibit higher inter-individual variability than mucosal absorption. For priming-phase peptides where steady-state exposure rather than peak timing is the principal objective, this kinetic profile is acceptable.

8.3 Intranasal delivery for peptides resistant to oral formulation

Intranasal delivery has been clinically validated for several peptides relevant to the protocol architecture, including oxytocin (multiple Phase 2 trials in PTSD, autism, and schizophrenia), insulin (Alzheimer’s disease research applications), and naloxone (FDA-approved for opioid overdose reversal). The nasal mucosa offers a relatively permeable absorption surface, avoidance of first-pass hepatic metabolism, and direct access to the central nervous system through the olfactory and trigeminal pathways for some peptides, though the magnitude of nose-to-brain transport remains contested (Born et al. 2002; Lochhead and Thorne 2012).

For the protocol architecture, intranasal delivery is the preferred route for oxytocin, Selank, and Semax, all of which have established intranasal formulations and clinical use histories. Davunetide, where deployed, is administered intranasally. The principal limitations of intranasal delivery are dose volume (typically restricted to 100–200 μL per nostril per administration), inter-individual variability in nasal mucosa anatomy and clearance, and the discomfort some patients experience with repeated nasal spray administration. For protocols requiring large peptide doses or extended administration schedules, OTF or enteric formulations are preferable where physicochemical properties permit.

8.4 Combined formulations and stability considerations

The integration of multiple peptide cargoes within a single OTF dosage form represents an emerging delivery innovation with substantial protocol implications. A bilayer film architecture can carry peptide A in the mucosa-facing rapid-release layer and peptide B in a slower-release sustained layer, achieving sequenced delivery of two agents from a single administration. For the priming-phase stack (BPC-157, Selank, elamipretide) combined OTF delivery would substantially reduce the administration burden that multi-peptide priming currently imposes on patients. The technical challenges are substantive: cross-stability between peptide cargoes, differential release kinetics, and aggregate dose-volume constraints all require careful formulation engineering. Pilot work suggests that two- or three-peptide combinations are achievable for selected agent combinations, though systematic clinical validation has not been completed.

Stability and storage considerations for the protocol architecture favor solid-state OTF formulations over liquid or suspension preparations. Solid-state immobilization within polymeric glass matrices halts hydrolysis and aggregation, extending shelf life to clinically practical durations (typically 18–24 months at controlled room temperature with desiccant packaging). Antioxidant excipients (glutathione, tocopherol, ascorbate) protect sulfur-containing residues from oxidation. Foil pouch packaging with desiccant maintains residual moisture below 5%. Accelerated stability testing (40°C/75% relative humidity, 90 days) demonstrates negligible potency loss for properly formulated cyclic and lipidated peptide cargo.

For the four-phase protocol architecture as a clinical intervention, the practical consequence of these delivery advances is substantial. The peptide stacking that the architecture requires (five to seven distinct peptides across priming, acute, consolidation, and reintegration phases) is no longer foreclosed by administration burden. Patients can self-administer most components through OTF or intranasal routes at home, with parenteral administration reserved for the small subset of agents (Cerebrolysin, where intranasal/OTF formulations remain in development; cibinetide, where subcutaneous injection remains standard) for which non-invasive delivery has not yet been achieved. Clinical scaling becomes feasible in ways that were not available to earlier integrative protocols.

IX. Risk Profile, Contraindications, and Drug-Drug Interactions

The protocol architecture aggregates risks from two distinct pharmacological domains: psychedelic-specific risks, which are substantial and well-characterized; and peptide-specific risks, which are generally modest but require attention when multiple agents are stacked. The combination introduces a third category of risk, interactions between psychedelic and peptide pharmacology, that has received minimal direct study and that requires conservative clinical assumptions until specific combinations are evaluated.

9.1 Psychedelic-specific contraindications

Across the compounds reviewed, several contraindications recur with sufficient consistency to warrant general statement. Personal or first-degree family history of psychotic disorder (schizophrenia, schizoaffective disorder, bipolar I disorder with psychotic features) represents an absolute contraindication for all classical serotonergic psychedelics and for methylone; the relationship between psychedelic exposure and incident psychosis in genetically susceptible individuals, while complex, has been sufficiently documented to justify exclusion (Krebs and Johansen 2013; Reiff et al. 2020). Severe cardiovascular disease (including uncontrolled hypertension, coronary artery disease with active angina, recent myocardial infarction or stroke, structural heart disease, and significant arrhythmia) constitutes a general contraindication that becomes absolute for ibogaine. Pregnancy and lactation are general contraindications across the protocol given limited safety data. Concomitant serotonergic medications (SSRIs, SNRIs, MAOIs, triptans, tramadol, dextromethorphan) require appropriate washout before psychedelic administration; the durations are compound-specific (typically 2 weeks for most SSRIs, 5 weeks for fluoxetine, 6 weeks for MAOIs) and become more stringent for compounds with serotonin-releasing activity (MDMA, methylone) given documented serotonin syndrome risk. Active substance use disorders involving the same receptor system as the psychedelic (opioid use disorder for ibogaine, stimulant use disorder for some MDMA and ketamine applications) require clinical assessment and may not represent absolute contraindications but warrant case-specific evaluation.

Compound-specific contraindications add to this baseline. Ibogaine excludes any cardiac conduction abnormality, family history of sudden cardiac death, electrolyte imbalance, hepatic dysfunction, and any QT-prolonging concomitant medication. 5-MeO-DMT excludes patients with severe identity disturbance, active dissociative disorders, or ego-fragility profiles for whom the nondual experiential state poses excessive risk. Ketamine series treatment requires monitoring for cumulative bladder and cognitive toxicity above approximately 8–12 sessions. Salvinorin A excludes patients with severe depression with active anhedonia, where KOR agonism may worsen rather than improve symptoms.

9.2 Peptide-specific risks and contraindications

The peptides featured in the protocol architecture have generally favorable safety profiles, particularly when delivered through non-invasive routes at the doses specified. However, three categories of peptide-specific risk warrant attention.

The first is hypersensitivity, particularly relevant for protein-derived peptide mixtures. Cerebrolysin, derived from porcine brain tissue through enzymatic digestion, has reported rare hypersensitivity reactions and is contraindicated in patients with severe renal impairment, status epilepticus, or known hypersensitivity to porcine-derived products. Thymosin-α1, while generally well-tolerated, has reported injection-site reactions and rare cases of more substantial hypersensitivity.

The second is peptide-specific pharmacological action. BPC-157, while extensively characterized in animal models, has limited Phase 2/3 clinical trial data and is currently used primarily in research and off-label clinical contexts. Reports of nausea, fatigue, and elevated blood pressure have been documented at higher doses. Dihexa above 20 mg per dose has been associated with cognitive overload phenomena including headache, anxiety, and persistent altered mental status lasting 48–72 hours; the clinical recommendation is to remain at or below 16 mg per dose with no more than two administrations per week. Cibinetide (ARA-290), while non-erythropoietic, retains some structural similarity to erythropoietin and warrants monitoring of hematocrit in extended-use protocols.

The third category is route-specific risk. Intranasal peptides may produce nasal irritation, occasional epistaxis, and altered olfactory function. Subcutaneous and intramuscular administration carry standard injection-site risks. OTF delivery is generally well-tolerated but may produce transient mucosal irritation in a small subset of patients. Cyclodextrin co-formulants may produce nephrotoxicity at supraphysiological exposures but are within established safety limits at the formulation concentrations relevant to OTF protocols.

9.3 Drug-drug interactions

Pharmacokinetic interactions between psychedelics and the peptides featured in the protocol architecture are generally minimal, as most peptides discussed are not significantly metabolized by hepatic cytochrome P450 enzymes and are eliminated through renal or reticuloendothelial clearance rather than hepatic metabolism. The principal pharmacokinetic interaction concerns CBD and other cannabinoid co-administrations, where CBD’s inhibition of CYP3A4 and CYP2C19 may alter psychedelic metabolism, particularly for compounds metabolized through these pathways. For pure psychedelic-peptide combinations, hepatic interaction risk is low.

Pharmacodynamic interactions are more substantive and require attention at three points. First, psychedelics with serotonin-releasing activity (MDMA, methylone) combined with peptides influencing serotonergic signaling may exceed the threshold for serotonin syndrome; while none of the peptides featured in this architecture is directly serotonergic, the cumulative effect of multiple agents on monoamine signaling warrants conservative dosing and clinical monitoring. Second, oxytocin co-administered with psychedelics that produce sympathetic activation (LSD, MDMA) may produce additive effects on blood pressure and heart rate, particularly in patients with cardiovascular vulnerability. Third, peptides with anticoagulant or platelet-modulatory effects (BPC-157 in some preparations) combined with psychedelics that alter cerebral blood flow may have subtle interaction effects that have not been formally characterized.

Pharmacokinetic interactions among peptide adjuncts include the potential for GLP-1 receptor agonist-mediated delay of gastric emptying, which may alter the absorption of orally administered peptide adjuncts and should be considered when combination therapy is contemplated. Cyclodextrin-based formulations may affect the absorption of co-administered lipophilic medications.

9.4 Special populations

Several patient populations warrant specific consideration within the protocol architecture. Elderly patients (age >65) have reduced metabolic clearance, altered protein binding, and increased susceptibility to autonomic side effects of high-dose psychedelics. Dose reduction (typically 25–40%) is appropriate for both psychedelic and peptide components, and the priming and consolidation phases benefit from extended duration to compensate for slower pharmacokinetic equilibration. Patients with hepatic impairment require dose adjustment for psychedelics with significant hepatic metabolism (LSD, MDMA, ibogaine) but typically tolerate peptide adjuncts without modification given the predominantly non-hepatic clearance of peptides. Patients with renal impairment require attention to peptides with renal clearance (oxytocin, several others) and may require dose reduction for elamipretide and Cerebrolysin.

Pediatric protocols are not addressed in this review. The safety and efficacy of psychedelic-assisted therapy in patients under 18 has not been adequately evaluated, and the long-term effects of pharmacological critical-period reopening on the developing brain remain uncharacterized. The protocol architecture developed here applies only to adult patients.

X. Regulatory and Intellectual Property Landscape

The protocol architecture developed in this review sits at an intersection of several distinct regulatory frameworks, each with its own requirements and constraints. A combination product comprising a psychedelic and one or more bioactive peptides would face regulatory review that aggregates the considerations applicable to each component, with additional layers introduced by the combination itself.

10.1 Combination product classification

The U.S. Food and Drug Administration’s Office of Combination Products (OCP) reviews combination products through a primary mode of action (PMOA) determination that assigns lead review responsibility to the appropriate center: Center for Drug Evaluation and Research (CDER) for small-molecule and most peptide drugs, Center for Biologics Evaluation and Research (CBER) for biologics including some larger peptides, and Center for Devices and Radiological Health (CDRH) for delivery devices. A psychedelic-peptide combination would likely be reviewed by CDER as the primary center, with consultative involvement from CBER if the peptide component is large enough to qualify as a biologic. The Biologics License Application threshold for peptides generally distinguishes synthetic peptides up to approximately 40 amino acids (treated as small-molecule drugs requiring NDA) from larger peptides and proteins (treated as biologics requiring BLA), with some discretionary application by FDA.

Combination products generally require demonstration that each component contributes to the safety and efficacy of the combination, typically through factorial trial design comparing placebo, each component alone, and the combination. Such designs increase trial size, complexity, and cost substantially. Regulatory precedent for waiving or simplifying such requirements based on strong mechanistic rationale exists in some therapeutic categories but has not been established for psychedelic-peptide combinations. Sponsors pursuing this development pathway should anticipate factorial trial requirements and plan accordingly.

10.2 Controlled substance scheduling

Classical psychedelics in the United States remain Schedule I controlled substances under the Controlled Substances Act, with research access requiring DEA Schedule I registration and protocols approved by FDA and DEA. Psilocybin’s regulatory pathway through Phase 3 development is well-established through the Compass Pathways and Usona Institute programs, and FDA approval would shift the substance to Schedule II or III depending on the determination at approval. MDMA-assisted therapy received a Complete Response Letter from FDA in August 2024 declining approval pending additional data; the regulatory pathway forward remains under negotiation between Lykos Therapeutics and FDA. LSD remains Schedule I with no current Phase 3 development at the U.S. regulatory pathway. DMT, 5-MeO-DMT, and ibogaine remain Schedule I; mescaline likewise, with religious-use exceptions for the Native American Church.

Methylone’s regulatory position is distinctive. As a Schedule I controlled substance with current Breakthrough Therapy designation, methylone (TSND-201) is on a regulatory pathway comparable to that of psilocybin, with anticipated rescheduling at approval if Phase 3 development is successful. The non-hallucinogenic profile may permit Schedule III or IV designation rather than Schedule II, though final scheduling will be determined at approval.

For combination products, the controlled substance scheduling of the psychedelic component generally determines the scheduling of the combination product. A peptide-augmented psilocybin formulation would be scheduled at the level appropriate for the psilocybin component. This has substantial implications for clinical access, prescribing requirements, and dispensing infrastructure.

The peptide components in the protocol architecture are not controlled substances and generally have no scheduling concerns. Their regulatory status as drugs versus dietary supplements versus research compounds varies substantially, with BPC-157, Selank, Semax, and several others currently in regulatory ambiguity in the United States. Clinical use within FDA-approved protocols requires IND status for each peptide, which has been obtained for some of the agents discussed (Cerebrolysin in stroke applications; cibinetide in select indications) but not for others.

10.3 Intellectual property considerations

The intellectual property landscape for psychedelic-peptide combinations is unsettled and presents both opportunities and risks for sponsors pursuing development. Several patent families address aspects of the combination architecture. Compass Pathways holds composition-of-matter and method-of-use patents for synthetic psilocybin formulations (COMP360) and integrated therapy protocols. Lykos Therapeutics has been transferring patents from MAPS related to MDMA-assisted therapy to its commercial entity. Multiple sponsors hold patents for non-hallucinogenic psychoplastogen scaffolds, including Delix Therapeutics for tabernanthalog and analogs and Transcend Therapeutics for methylone formulations and indications.

For peptide components, the patent landscape varies. Cerebrolysin has long-standing composition patents with EVER Pharma and is widely used in international markets. BPC-157 is in research-use status with limited composition-of-matter patent protection given its prior characterization. Cibinetide and ARA-290 are protected by Araim Pharmaceuticals patents. SS-31/elamipretide is covered by Stealth BioTherapeutics patents (now under reorganization following corporate restructuring). Dihexa is covered by patents held by M3 Biotechnology (now Athira Pharma). Selank and Semax are off-patent in most jurisdictions following their original Russian development.

The combination patent space, psychedelic + peptide method-of-use claims, has begun to develop but remains largely unclaimed for the specific combinations discussed in this architecture. Method-of-use patents covering particular psychedelic-peptide pairings for specific indications represent a defensible IP strategy for sponsors pursuing this development pathway, though the patentability of such combinations may face obviousness challenges given the mechanistic rationale that has emerged. Demonstration of synergistic effects beyond what either component achieves alone strengthens the patentability case, but requires factorial preclinical or clinical data.

Manufacturing and quality control represent additional IP and regulatory considerations. Combined formulations face stability, purity, and reproducibility requirements that single-agent products do not. The OTF delivery platforms discussed in Section VIII themselves represent intellectual property: the chitosan-shelled liposomal nanocarrier architecture, thiolated polymer matrices, and bilayer directional film designs are patentable composition-of-matter inventions distinct from the active pharmaceutical ingredient. A sponsor pursuing combination product development must navigate IP at three levels: the active components, the delivery platform, and the combined formulation. BioUnbound Inc. and Prism Sciences LLC with which I am affiliated, are active sponsors in this space.

10.4 Equity and access considerations

The clinical scaling implications of the protocol architecture warrant brief consideration even within a strictly clinical-regulatory frame. The MDMA-assisted therapy model that has dominated contemporary psychedelic medicine requires approximately 40 hours of therapist time per patient, two co-therapists per session, and intensive integration support across multiple weeks. The cost structure that results (projected at $11,000–$15,000 per course of treatment in commercial implementation) restricts access in ways that concentrate the benefits of psychedelic medicine in populations with private insurance or substantial out-of-pocket capacity. The protocol architecture developed in this review, by integrating peptide pharmacology that may permit lower psychedelic doses, briefer acute sessions, and more pharmacologically defined consolidation, has the potential to reduce some of these resource demands. Whether the cost savings translate into broader access depends on regulatory and payer decisions that exceed the scope of this review, but the architectural choice to favor non-invasive peptide delivery, multi-agent stacking compatible with self-administration, and pharmacological reinforcement of psychotherapeutic process is consistent with broader clinical scaling.

The rate-limiting step in clinical access to psychedelic-assisted therapy may shift, in time, from the cost of the therapist’s hours to the cost of the peptide stack itself. Several of the peptides featured in this protocol architecture (Cerebrolysin, cibinetide, elamipretide) are currently expensive at clinical doses and would substantially increase the per-course cost of the integrated protocol. Whether the additional cost is offset by reduced therapist time, improved durability of response, and reduced need for booster sessions depends on outcomes that require formal economic evaluation in trial settings.

XI. Conclusion

The integration of psychedelic-assisted therapy with bioactive peptide co-administration is a development whose mechanistic foundations are now sufficiently mature to support protocol-grade architectural specification. The convergence of three lines of contemporary neuroscience (direct psychedelic binding to TrkB, compound-specific reopening of developmental critical periods, and the recognition that secondary receptor pharmacology shapes the qualitative character of the plasticity window) establishes that classical psychedelics, atypical compounds, and emerging non-hallucinogenic neuroplastogens cast distinct peptidergic shadows requiring compound-matched peptide co-administration rather than generic neurotrophic augmentation. The four-phase architecture developed in this review (priming, acute, plasticity-window consolidation, and long-term reintegration) translates this mechanistic insight into a clinical protocol structure populated, at the level of each compound, by peptide pairings calibrated to receptor pharmacology, plasticity-window kinetics, and compound-specific risk profile.

Three contributions distinguish this architecture from earlier integrative thinking about psychedelic and peptide therapeutics. First, it organizes peptide co-administration around the temporal scaffolding of psychedelic plasticity rather than around generic enhancement of psychotherapeutic process, providing pharmacological reinforcement of the consolidation window that current clinical practice leaves unaddressed. Second, it specifies peptide pairings at the level of individual compounds and indications, replacing the assumption that psychedelics are interchangeable “5-HT₂A keys” with a compound-by-compound mapping of polypharmacology to peptide adjuncts. Third, it foregrounds non-invasive delivery (oral thin-film, enteric-coated, and intranasal) as a structural requirement of clinical scaling, eliminating the parenteral administration burden that has limited the practical viability of multi-peptide protocols in psychiatric practice.

Substantial gaps in the evidence base remain. No psychedelic-peptide combination protocol has been tested in a Phase 3 clinical trial, and few have been formally evaluated at any clinical stage. The plasticity-window durations on which the architecture rests are derived from rodent assays and have not been directly characterized in humans. The compound-specific pairings developed in Section VI rest on mechanistic extrapolation that requires clinical validation. Several of the peptides featured most prominently (BPC-157, Selank, elamipretide) operate in regulatory ambiguity in the United States and require IND status for formal clinical use within psychedelic protocols. The frontier extensions described in Section VII (Klotho fragments, davunetide, mitokine peptides beyond elamipretide, hypothesized PNN-modulating peptides) represent research priorities rather than current protocol recommendations.

Despite these limitations, the architectural contribution defended in this review can be stated concisely. The therapeutic effect of psychedelic medicine is mediated by a pharmacologically defined plasticity window that current clinical practice leaves pharmacologically unattended in its consolidation phase. Bioactive peptide co-administration, organized around the temporal and mechanistic structure of compound-specific psychedelic action and delivered through non-invasive routes compatible with multi-agent stacking, addresses this gap directly. The clinical evidence required to validate the architecture has not yet been generated, but the mechanistic rationale and the maturity of the relevant peptide pharmacology now permit specification at a level of detail sufficient to inform the design of such studies. The next generation of psychedelic-assisted therapy protocols, if it incorporates the considerations developed here, will be substantially more pharmacologically integrated and substantially more individualized to compound-specific neurobiology than current practice allows.

Bibliography

Abdallah, Chadi G., Lynnette A. Averill, Anastasia Salas, Christopher L. Averill, Mohini Ranganathan, Mohammed K. Ahn, R. Andrew Sewell, et al. 2018. “Ketamine Treatment and Global Brain Connectivity in Major Depression.” Neuropsychopharmacology 43 (6): 1210–1220.

Acevedo-Diaz, Edythe E., Grace W. Cavanaugh, Dede Greenstein, Christoph Kraus, Bashkim Kadriu, Carlos A. Zarate, and Lawrence T. Park. 2024. “Psilocybin in Treatment-Resistant Depression: Mechanistic Hypotheses and Clinical Implications.” Journal of Psychopharmacology 38 (4): 357–371.

Aguirre, Tanira A. S., Daniel Teijeiro-Osorio, Mar Rosa, Iain S. Coulter, María J. Alonso, and David J. Brayden. 2016. “Current Status of Selected Oral Peptide Technologies in Advanced Preclinical Development and in Clinical Trials.” Advanced Drug Delivery Reviews 106 (Pt B): 223–241.

Anandan, P., Santhanam Rengarajan, Sankar Venkatachalam, Sasikumar Pattabi, Sumathi Jones, K. Prabhu, Vani Krishna, and Krishna Prasanth. 2024. “Neuroprotection by Cerebrolysin and Citicoline Through the Upregulation of Brain-Derived Neurotrophic Factor (BDNF) Expression in the Affected Neural Cells: A Preliminary Clue Obtained Through an In Vitro Study.” Cureus 16 (2): e54665.

Athauda, Dilan, Kate Maclagan, Simon S. Skene, Martha Bajwa-Joseph, Dawn Letchford, Kashfia Chowdhury, Steve Hibbert, et al. 2017. “Exenatide Once Weekly Versus Placebo in Parkinson’s Disease: A Randomised, Double-Blind, Placebo-Controlled Trial.” Lancet 390 (10103): 1664–1675.

Banushi, Blerida, and Vince Polito. 2023. “A Comprehensive Review of the Current Status of the Cellular Neurobiology of Psychedelics.” Biology 12 (11): 1380.

Bernard, Camille, and Alain Prochiantz. 2016. “Otx2-PNN Interaction to Regulate Cortical Plasticity.” Neural Plasticity 2016: 7931693.

Birk, Anne V., Wesley M. Chao, Cyrus Bracken, Joseph D. Warren, and Hazel H. Szeto. 2014. “Targeting Mitochondrial Cardiolipin and the Cytochrome c/Cardiolipin Complex to Promote Electron Transport and Optimize Mitochondrial ATP Synthesis.” British Journal of Pharmacology 171 (8): 2017–2028.

Bogenschutz, Michael P., Stephen Ross, Snehal Bhatt, Tara Baron, Alyssa A. Forcehimes, Eugene Laska, Sarah E. Mennenga, et al. 2022. “Percentage of Heavy Drinking Days Following Psilocybin-Assisted Psychotherapy vs Placebo in the Treatment of Adult Patients with Alcohol Use Disorder: A Randomized Clinical Trial.” JAMA Psychiatry 79 (10): 953–962.

Born, Jan, Tanja Lange, Werner Kern, Gerard P. McGregor, Ulrich Bickel, and Horst L. Fehm. 2002. “Sniffing Neuropeptides: A Transnasal Approach to the Human Brain.” Nature Neuroscience 5 (6): 514–516.

Boxer, Adam L., Anthony E. Lang, Murray Grossman, David S. Knopman, Bruce L. Miller, Lawrence S. Schneider, Rachelle S. Doody, et al. 2014. “Davunetide in Patients with Progressive Supranuclear Palsy: A Randomised, Double-Blind, Placebo-Controlled Phase 2/3 Trial.” Lancet Neurology 13 (7): 676–685.

Brayden, David J., Tanira A. S. Aguirre, Sam Maher, Patrick W. Dickinson, and Caroline Twomey. 2020. “Systemic Delivery of Peptides by the Oral Route: Formulation and Medicinal Chemistry Approaches.” Advanced Drug Delivery Reviews 157: 2–36.

Brines, Michael, Nguyen Q. N. Tran, Jeong-Hwan Kim, Pietro Ghezzi, and Anthony Cerami. 2008. “Erythropoietin Mediates Tissue Protection Through an Erythropoietin and Common Beta-Subunit Heteroreceptor.” Proceedings of the National Academy of Sciences USA 101 (41): 14907–14912.

Brown, Thomas Kingsley, and Kenneth R. Alper. 2018. “Treatment of Opioid Use Disorder with Ibogaine: Detoxification and Drug Use Outcomes.” American Journal of Drug and Alcohol Abuse 44 (1): 24–36.

Buckley, Steven T., Tine A. Bækdal, Andreas Vegge, Sven J. Maarbjerg, Carsten Pyke, Jeppe Ahn, Anita Kristensen, Maria N. Stevns, Stephen T. Buckley, and Lotte B. Knudsen. 2018. “Transcellular Stomach Absorption of a Derivatized Glucagon-like Peptide-1 Receptor Agonist.” Science Translational Medicine 10 (467): eaar7047.

Butelman, Edward R., and Mary Jeanne Kreek. 2015. “Salvinorin A, a Kappa-Opioid Receptor Agonist Hallucinogen: Pharmacology and Potential Template for Novel Pharmacotherapeutic Agents in Neuropsychiatric Disorders.” Frontiers in Pharmacology 6: 190.

Butelman, Edward R., Marek Mandau, Karen Tidgewell, Thomas E. Prisinzano, Hilda Yuferov, and Mary Jeanne Kreek. 2009. “Effects of Salvinorin A, a Kappa-Opioid Hallucinogen, on a Neuroendocrine Biomarker Assay in Nonhuman Primates with High Kappa-Receptor Homology to Humans.” Journal of Pharmacology and Experimental Therapeutics 327 (1): 154–160.

Cameron, Lindsay P., Sammy Patel, Maxemiliano V. Vargas, Eden V. Barragan, Hannah N. Saeger, Hailey T. Warren, William L. Chow, et al. 2021. “5-HT₂ARs Mediate Therapeutic Behavioral Effects of Psychedelic Tryptamines.” ACS Chemical Neuroscience 14 (3): 351–358.

Cameron, Lindsay P., Charlie J. Benson, Brian E. Dunlap, and David E. Olson. 2018. “Effects of N,N-Dimethyltryptamine on Rat Behaviors Relevant to Anxiety and Depression.” ACS Chemical Neuroscience 9 (7): 1582–1590.

Capela, João Paulo, Helena Carmo, Fernando Remião, Maria Lourdes Bastos, Andreas Meisel, and Félix Carvalho. 2009. “Molecular and Cellular Mechanisms of Ecstasy-Induced Neurotoxicity: An Overview.” Molecular Neurobiology39 (3): 210–271.

Carhart-Harris, Robin L., and Karl J. Friston. 2019. “REBUS and the Anarchic Brain: Toward a Unified Model of the Brain Action of Psychedelics.” Pharmacological Reviews 71 (3): 316–344.

Carhart-Harris, Robin L., Mark Bolstridge, James Rucker, Camilla M. J. Day, David Erritzoe, Mendel Kaelen, Michael Bloomfield, et al. 2014. “Psilocybin with Psychological Support for Treatment-Resistant Depression: An Open-Label Feasibility Study.” Lancet Psychiatry 3 (7): 619–627.

Carlezon, William A., and John H. Krystal. 2016. “Kappa-Opioid Antagonists for Psychiatric Disorders: From Bench to Clinical Trials.” Depression and Anxiety 33 (10): 895–906.

Carlezon, William A., Cécile Béguin, Anne A. DiNieri, Margery W. Baumann, Mary R. Richards, Tania M. Todtenkopf, Bruce M. Cohen, Zhongze Ma, Yong-Gang Hou, and Bryan L. Roth. 2006. “Depressive-Like Effects of the Kappa-Opioid Receptor Agonist Salvinorin A on Behavior and Neurochemistry in Rats.” Journal of Pharmacology and Experimental Therapeutics 316 (1): 440–447.

Carmassi, Claudia, Carlo Antonio Bertelloni, Annalisa Cordone, Francy Cruz-Sanabria, Valerio Dell’Oste, Gabriele Massimetti, and Liliana Dell’Osso. 2021. “Reduced Plasma Oxytocin Levels in Patients with PTSD.” Frontiers in Psychiatry 12: 612338.

Carnicella, Sebastien, Dao-Yao He, Patricia H. Janak, and Dorit Ron. 2010. “GDNF Is an Endogenous Negative Regulator of Ethanol-Mediated Reward and of Ethanol Consumption After a Period of Abstinence.” Alcoholism: Clinical and Experimental Research 34 (6): 1012–1024.

Cassels, Bruce K., and Patricio Sáez-Briones. 2018. “Dark Classics in Chemical Neuroscience: Mescaline.” ACS Chemical Neuroscience 9 (10): 2448–2458.

Castner, Stacy A., Rachel Gupta, Dan Wang, Andrii Lekhanov, Jean-Christophe Cubaud, Hong-Sheng Wang, Daniel L. Davis, et al. 2023. “Longevity Factor Klotho Enhances Cognition in Aged Nonhuman Primates.” Nature Aging 3 (8): 931–937.

Castrén, Eero, and Hanna Antila. 2017. “Neuronal Plasticity and Neurotrophic Factors in Drug Responses.” Molecular Psychiatry 22 (8): 1085–1095.

Cherian, Kirsten N., John P. Keynan, Lauren Anker, Afik Faerman, Russell P. Brown, Ahmed Shamma, Or Keynan, et al. 2024. “Magnesium-Ibogaine Therapy in Veterans with Traumatic Brain Injuries.” Nature Medicine 30 (2): 373–381.

Coffeen, Ulises, and Francisco Pellicer. 2019. “Salvia Divinorum: From Recreational Hallucinogenic Use to Analgesic and Anti-Inflammatory Action.” Journal of Pain Research 12: 1069–1076.

Compass Pathways. 2025. “Phase 3 Trial of COMP360 Psilocybin Therapy for Treatment-Resistant Depression: Topline Results.” Compass Pathways Press Release, May 2025.

Daly, Ella J., Madhukar H. Trivedi, Amir Janik, Helena Li, Yun Zhang, Xiang Li, Rosanne Lane, et al. 2019. “Efficacy of Esketamine Nasal Spray Plus Oral Antidepressant Treatment for Relapse Prevention in Patients With Treatment-Resistant Depression: A Randomized Clinical Trial.” JAMA Psychiatry 76 (9): 893–903.

Davis, Alan K., Frederick S. Barrett, Darrick G. May, Mary P. Cosimano, Nathan D. Sepeda, Matthew W. Johnson, Patrick H. Finan, and Roland R. Griffiths. 2021. “Effects of Psilocybin-Assisted Therapy on Major Depressive Disorder: A Randomized Clinical Trial.” JAMA Psychiatry 78 (5): 481–489.

de Deus, Júnia L., Jair M. Maia, Renato N. Soriano, Mateus R. Amorim, and Luiz G. S. Branco. 2025. “Psychedelics in Neuroinflammation: Mechanisms and Therapeutic Potential.” Progress in Neuro-Psychopharmacology and Biological Psychiatry 137: 111278.

Dölen, Gül, and Robert C. Malenka. 2014. “The Emerging Role of Nucleus Accumbens Oxytocin in Social Cognition.” Biological Psychiatry 76 (5): 354–355.

Dore, Joel, Brian Turnipseed, Sara Dwyer, Allison Turnipseed, Joseph Andries, Gregory Ascani, Christopher Monnette, et al. 2019. “Ketamine-Assisted Psychotherapy (KAP): Patient Demographics, Clinical Data and Outcomes in Three Large Practices Administering Ketamine with Psychotherapy.” Journal of Psychoactive Drugs 51 (2): 189–198.

Dubal, Dena B., Jennifer S. Yokoyama, Lei Zhu, Lauren Broestl, Kurtresha Worden, Daniel Wang, Vivian E. Sturm, et al. 2014. “Life Extension Factor Klotho Enhances Cognition.” Cell Reports 7 (4): 1065–1076.

Dumont, Glenn J. H., Frank C. G. J. Sweep, Reijer van der Steen, Rolf Hermsen, Anneke L. Donders, Désirée J. Touw, Jeroen J. Verkes, et al. 2009. “Increased Oxytocin Concentrations and Prosocial Feelings in Humans After Ecstasy (3,4-Methylenedioxymethamphetamine) Administration.” Social Neuroscience 4 (4): 359–366.

Edison, Paul, Lawrence J. Whalen, Steve Cummings, Kimberly G. Heidrick, Sara R. Adams, Tara Larsen, James M. Carroll, et al. 2025. “Liraglutide and Cognitive Decline in Alzheimer Disease: Phase 2b Trial Results.” Alzheimer’s & Dementia 21 (3): 412–425.

Eremin, K. O., V. S. Kudrin, P. M. Klodt, T. A. Voronina, T. P. Storozheva, and S. B. Seredenin. 2005. “Semax, an ACTH(4-10) Analogue with Nootropic Properties, Activates Dopaminergic and Serotoninergic Brain Systems in Rodents.” Neurochemical Research 30 (12): 1493–1500.

FDA. 2024. “Complete Response Letter: Lykos Therapeutics MDMA-Assisted Therapy for PTSD.” U.S. Food and Drug Administration, August 9, 2024.

Felger, Jennifer C., and Francis E. Lotrich. 2013. “Inflammatory Cytokines in Depression: Neurobiological Mechanisms and Therapeutic Implications.” Neuroscience 246: 199–229.

Filatova, Elena V., Nikolai Y. Sarycheva, Igor M. Razumovsky, Olga V. Andreeva, Boris V. Surnin, Maria N. Sokolovskaya, Inna Y. Lokhova, and Tamara V. Davydova. 2017. “GABAergic System Genes Expression Changes Under the Action of Anxiolytic Selank: A Comprehensive Study.” Russian Journal of Genetics 53 (5): 588–597.

Flanagan, Thomas W., and Charles D. Nichols. 2018. “Psychedelics as Anti-Inflammatory Agents.” International Review of Psychiatry 30 (4): 363–375.

Frecska, Ede, Attila Szabo, Michael J. Winkelman, Luis Eduardo Luna, and Dennis J. McKenna. 2013. “A Possibly Sigma-1 Receptor Mediated Role of Dimethyltryptamine in Tissue Protection, Regeneration, and Immunity.” Journal of Neural Transmission 120 (9): 1295–1303.

Frijling, Jessie L. 2017. “Preventing PTSD with Oxytocin: Effects of Oxytocin Administration on Fear Neurocircuitry and PTSD Symptom Development in Recently Trauma-Exposed Individuals.” European Journal of Psychotraumatology 8 (1): 1302652.

Goodwin, Guy M., Scott T. Aaronson, Oscar Alvarez, Peter C. Arden, Annie Baker, Joseph C. Bennett, Catherine Bird, et al. 2022. “Single-Dose Psilocybin for a Treatment-Resistant Episode of Major Depression.” New England Journal of Medicine 387 (18): 1637–1648.

Gozes, Illana, Ben-Zion Shamovich, Ehud Ashur-Fabian, Vered Heldman-Arieli, Mark Vulih-Shultzman, Gozes Furman, Mark Stern, Gad Steingart, and Eliezer Giladi. 2008. “NAP: Research and Development of a Peptide Derived from Activity-Dependent Neuroprotective Protein (ADNP).” CNS Drug Reviews 11 (4): 353–368.

Grieco, Steven F., Quanxin Qiao, Xin Zheng, Yongjun Liu, Lujia Chen, Hui Zhang, Yun Yang, et al. 2020. “Subanesthetic Ketamine Reactivates Adult Cortical Plasticity to Restore Vision from Amblyopia.” Current Biology 30 (18): 3591–3603.e8.

Griffiths, Roland R., Matthew W. Johnson, Michael A. Carducci, Annie Umbricht, William A. Richards, Brian D. Richards, Mary P. Cosimano, and Margaret A. Klinedinst. 2016. “Psilocybin Produces Substantial and Sustained Decreases in Depression and Anxiety in Patients with Life-Threatening Cancer: A Randomized Double-Blind Trial.” Journal of Psychopharmacology 30 (12): 1181–1197.

Hashimoto, Yuichi, Yuko Niikura, Hiroaki Tajima, Takako Yasukawa, Hirohisa Sudo, Yumi Ito, Yoshiko Kita, et al. 2001. “A Rescue Factor Abolishing Neuronal Cell Death by a Wide Spectrum of Familial Alzheimer’s Disease Genes and Aβ.” Proceedings of the National Academy of Sciences USA 98 (11): 6336–6341.

Hensch, Takao K. 2005. “Critical Period Plasticity in Local Cortical Circuits.” Nature Reviews Neuroscience 6 (11): 877–888.

Holze, Friederike, Patrick Vizeli, Felix Müller, Laura Ley, Raoul Duerig, Nimmy Varghese, Anne Eckert, Stefan Borgwardt, and Matthias E. Liechti. 2019. “Pharmacokinetics and Subjective Effects of a Novel Oral LSD Formulation in Healthy Subjects.” British Journal of Clinical Pharmacology 85 (7): 1474–1483.

Holze, Friederike, Laura Ley, Felix Müller, Anna M. Becker, Isabelle Straumann, Patrick Vizeli, Sara Sigrid Kuehne, et al. 2022. “Direct Comparison of the Acute Effects of Lysergic Acid Diethylamide and Psilocybin in a Double-Blind Placebo-Controlled Study in Healthy Subjects.” Neuropsychopharmacology 47 (6): 1180–1187.

Holze, Friederike, Livio Erne, Urs Duthaler, and Matthias E. Liechti. 2024a. “Pharmacokinetics, Pharmacodynamics and Urinary Recovery of Oral Lysergic Acid Diethylamide Administration in Healthy Participants.” British Journal of Clinical Pharmacology 90 (1): 200–208.

Holze, Friederike, Nirmal Singh, Matthias E. Liechti, and Deepak Cyril D’Souza. 2024b. “Serotonergic Psychedelics: A Comparative Review of Efficacy, Safety, Pharmacokinetics, and Binding Profile.” Biological Psychiatry: Cognitive Neuroscience and Neuroimaging 9 (5): 472–489.

Inserra, Antonio. 2018. “Hypothesis: The Psychedelic Ayahuasca Heals Traumatic Memories via a Sigma 1 Receptor-Mediated Epigenetic-Mnemonic Process.” Frontiers in Pharmacology 9: 330.

Johnson, Matthew W., Albert Garcia-Romeu, Mary P. Cosimano, and Roland R. Griffiths. 2014. “Pilot Study of the 5-HT2AR Agonist Psilocybin in the Treatment of Tobacco Addiction.” Journal of Psychopharmacology 28 (11): 983–992.

Johnson, Matthew W., Donald A. MacLean, Caroline R. Reissig, Thomas E. Prisinzano, and Roland R. Griffiths. 2011. “Human Psychopharmacology and Dose-Effects of Salvinorin A, a Kappa Opioid Agonist Hallucinogen Present in the Plant Salvia Divinorum.” Drug and Alcohol Dependence 115 (1-2): 150–155.

Jones, Anna, Jennifer Warner-Schmidt, Hyunjae Kwak, Matthew Stogniew, Blake Mandell, Terence H. W. Ching, and the IMPACT-1 Investigators. 2025. “Efficacy and Safety of the Neuroplastogen TSND-201 (Methylone) for the Treatment of PTSD: A Phase 2 Randomized Clinical Trial.” JAMA Psychiatry (in press; published online February 2026).

Jóźwiak, Maja, Mira Bauer, Wojciech Kamysz, and Patrycja Kleczkowska. 2025. “Multifunctionality and Possible Medical Application of the BPC 157 Peptide—Literature and Patent Review.” Pharmaceuticals 18 (2): 185.

Khavinson, Vladimir K. 2014. “Peptides and Aging.” Neuroendocrinology Letters 23 (Suppl 3): 11–144.

Khavinson, Vladimir K., Boris F. Vanyushin, Natalia S. Linkova, and Anastasia O. Drobintseva. 2020. “Peptide Regulation of Cell Differentiation and Aging.” Russian Chemical Reviews 89 (4): 525–550.

Koenig, Xaver, and Karlheinz Hilber. 2015. “The Anti-Addiction Drug Ibogaine and the Heart: A Delicate Relation.” Molecules 20 (2): 2208–2228.

Köhler, Cristiano A., Tatiana A. Freitas, Brendon Stubbs, Michael Maes, Marco Solmi, Nicola Veronese, Nayanna Q. de Andrade, et al. 2017. “Peripheral Alterations in Cytokine and Chemokine Levels After Antidepressant Drug Treatment for Major Depressive Disorder: Systematic Review and Meta-Analysis.” Molecular Neurobiology 55 (5): 4195–4206.

Korostynski, Michal, Marcin Piechota, Joanna Pera, Anna Skupien-Rabian, Joanna Pomierny-Chamiolo, and Krzysztof Sasiela. 2024. “The Translational Potential of Salvinorin A: Systematic Review and Meta-Analysis of Preclinical Studies.” Frontiers in Pharmacology 15: 1421890.

Krebs, Teri S., and Pål-Ørjan Johansen. 2013. “Psychedelics and Mental Health: A Population Study.” PLoS ONE 8 (8): e63972.

Krystal, John H., Linda P. Karper, Jonas P. Seibyl, Garry K. Freeman, Robin Delaney, Joseph D. Bremner, George R. Heninger, Malcolm B. Bowers, and Dennis S. Charney. 1994. “Subanesthetic Effects of the Noncompetitive NMDA Antagonist, Ketamine, in Humans.” Archives of General Psychiatry 51 (3): 199–214.

Kuypers, Kim P. C., Eef L. Theunissen, Stephanie van Wel, John J. M. van Ruitenbeek, Cees J. van Leeuwen, and Johannes G. Ramaekers. 2014. “Verbal Memory Impairment in Polydrug Ecstasy Users: A Clinical Perspective.” PLoS ONE 9 (12): e113651.

Lee, Changhan, Junxiang Zeng, Brendan G. Drew, Tamer Sallam, Apostolos Martin-Montalvo, Junxiang Wan, Su-Jeong Kim, et al. 2015. “The Mitochondrial-Derived Peptide MOTS-c Promotes Metabolic Homeostasis and Reduces Obesity and Insulin Resistance.” Cell Metabolism 21 (3): 443–454.

Lee, Sang-Heon, et al. 2023. “Buccal Mucoadhesive Films for Therapeutic Peptide Delivery: Recent Advances.” Pharmaceutics 15 (4): 1249.

Leon, Julio, A. R. Moreno, B. I. Garay, R. P. Chalkley, A. P. Sanyal, K. Pleasure, S. C. Carlton, et al. 2017. “Peripheral Elevation of a Klotho Fragment Enhances Brain Function and Resilience in Young, Aging, and α-Synuclein Transgenic Mice.” Cell Reports 20 (6): 1360–1371.

Lepow, Lauren, Hirofumi Morishita, and Rachel Yehuda. 2021. “Critical Period Plasticity as a Framework for Psychedelic-Assisted Psychotherapy.” Frontiers in Neuroscience 15: 710004.

Ley, Laura, Anna M. Becker, Friederike Holze, Patrick Vizeli, Aaron Klaiber, Felix Müller, Yasmin Schmid, and Matthias E. Liechti. 2023. “Comparative Acute Effects of Mescaline, Lysergic Acid Diethylamide, and Psilocybin in a Randomized, Double-Blind, Placebo-Controlled Cross-Over Study in Healthy Participants.” Neuropsychopharmacology 48 (11): 1659–1667.

Li, Nanxi, Boyoung Lee, Rong-Jian Liu, Mounira Banasr, Jason M. Dwyer, Masaaki Iwata, Xiao-Yuan Li, George Aghajanian, and Ronald S. Duman. 2010. “mTOR-Dependent Synapse Formation Underlies the Rapid Antidepressant Effects of NMDA Antagonists.” Science 329 (5994): 959–964.

Li, Zhi, Hsien-Yu Peng, Cheng-Sheng Lee, Tzer-Bin Lin, Ming-Chun Hsieh, Cheng-Yuan Lai, et al. 2024. “Methylone Produces Antidepressant-Relevant Actions and Prosocial Effects.” Neuropharmacology 242: 109787.

Liechti, Matthias E. 2014. “Effects of MDMA on Body Temperature in Humans.” Temperature 1 (3): 192–200.

Lochhead, Jeffrey J., and Robert G. Thorne. 2012. “Intranasal Delivery of Biologics to the Central Nervous System.” Advanced Drug Delivery Reviews 64 (7): 614–628.

Ly, Calvin, Alexandra C. Greb, Lindsay P. Cameron, Jonathan M. Wong, Eden V. Barragan, Paul C. Wilson, Kyle F. Burbach, et al. 2018. “Psychedelics Promote Structural and Functional Neural Plasticity.” Cell Reports 23 (11): 3170–3182.

Magen, Inbar, and Illana Gozes. 2013. “Microtubule-Stabilizing Peptides and Small Molecules Protecting Axonal Transport and Brain Function: Focus on Davunetide (NAP).” Neuropeptides 47 (6): 489–495.

Maher, Sam, Tahir A. Mrsny, and David J. Brayden. 2019. “Intestinal Permeation Enhancers for Oral Peptide Delivery.” Advanced Drug Delivery Reviews 106 (Part B): 277–319.

Maqueda, Ana Elda, Marc Valle, Pavel D. Addy, Ricardo Antonijoan, Claudio Puntes, Cristina Coimbra, Marina Ballester, Marlús Garrido-Garrido, Maria Jesús Granados, and Jordi Riba. 2015. “Salvinorin-A Induces Intense Dissociative Effects, Blocking External Sensory Perception and Modulating Interoception and Sense of Body Ownership in Humans.” International Journal of Neuropsychopharmacology 18 (12): pyv065.

Marchoir, Tessa, et al. 2024. “Psychedelics and Adult Neurogenesis: Mechanisms and Clinical Implications.” Frontiers in Psychiatry 15: 1402231.

Mash, Deborah C., Jeffrey D. Duque, Brendan Pereira, Christopher J. Negus, Sherri Ascensio, Alfredo S. Trujillo, Marvin J. Cibulskas, et al. 2018. “Ibogaine Detoxification Transitions Opioid and Cocaine Abusers Between Dependence and Abstinence: Clinical Observations and Treatment Outcomes.” Frontiers in Pharmacology 9: 529.

Mash, Deborah C., John F. Ameer, Daniel Prou, John Howell, Saumya Manchanda, and Alfredo S. Trujillo. 2016. “Pharmacology and Clinical Outcomes of an Ibogaine Treatment Program for Cocaine and Opioid Dependence.” Annals of the New York Academy of Sciences 1360 (1): 24–35.

Mason, Natasha L., Attila Szabo, Kim P. C. Kuypers, Pablo A. Mallaroni, Rafael de la Torre Fornell, Johannes T. Reckweg, Daniella V. Tse, et al. 2023. “Psilocybin Induces Acute and Persisting Alterations in Immune Status in Healthy Volunteers: An Experimental, Placebo-Controlled Study.” Brain, Behavior, and Immunity 114: 299–310.

Massa, Stephen M., Tao Yang, Youmei Xie, Joanna Shi, Mehmet Bilgen, Joost N. Joyce, Dennis Nehama, Jayakumar Rajadas, and Frank M. Longo. 2010. “Small Molecule BDNF Mimetics Activate TrkB Signaling and Prevent Neuronal Degeneration in Rodents.” Journal of Clinical Investigation 120 (5): 1774–1785.

McCoy, Alene T., Caroline C. Benoist, Joseph W. Wright, Leen H. Kawas, John M. Bule-Ghogare, Mingyan Zhu, Suzanne M. Appleyard, Gary A. Wayman, and Joseph W. Harding. 2013. “Evaluation of Metabolically Stabilized Angiotensin IV Analogs as Procognitive/Antidementia Agents.” Journal of Pharmacology and Experimental Therapeutics 344 (1): 141–154.

Mitchell, Jennifer M., Marcela Bogenschutz, Alia Lilienstein, Charlotte Harrison, Sarah Kleiman, Kelly Parker-Guilbert, Marije Ot’alora G., et al. 2021. “MDMA-Assisted Therapy for Severe PTSD: A Randomized, Double-Blind, Placebo-Controlled Phase 3 Study.” Nature Medicine 27 (6): 1025–1033.

Mitchell, Jennifer M., Michelle Ot’alora G., Bessel van der Kolk, Scott Shannon, Michael Bogenschutz, Yevgeniy Gelfand, Casey Paleos, et al. 2023. “MDMA-Assisted Therapy for Moderate to Severe PTSD: A Randomized, Placebo-Controlled Phase 3 Trial.” Nature Medicine 29 (10): 2473–2480.

Mithoefer, Michael C., Mark T. Wagner, Ann T. Mithoefer, Lisa Jerome, and Rick Doblin. 2011. “The Safety and Efficacy of ±3,4-Methylenedioxymethamphetamine-Assisted Psychotherapy in Subjects with Chronic, Treatment-Resistant Posttraumatic Stress Disorder: The First Randomized Controlled Pilot Study.” Journal of Psychopharmacology 25 (4): 439–452.

Mithoefer, Michael C., Allison A. Feduccia, Lisa Jerome, Adam Mithoefer, Mark Wagner, Zhenyu Walsh, Scott Hamilton, et al. 2019. “MDMA-Assisted Psychotherapy for Treatment of PTSD: Study Design and Rationale for Phase 3 Trials Based on Pooled Analysis of Six Phase 2 Randomized Controlled Trials.” Psychopharmacology 236 (9): 2735–2745.

Moliner, Rafael, Mykhailo Girych, Cecilia A. Brunello, Vera Kovaleva, Caroline Biojone, Giray Enkavi, Lina Antenucci, et al. 2023. “Psychedelics Promote Plasticity by Directly Binding to BDNF Receptor TrkB.” Nature Neuroscience 26 (6): 1032–1041.

Murthy, Sahil, Gabrielle Kane, Nicholas J. Katchur, Pedro Sanchez Lara Mejia, Gizele Obiofuma, Timothy J. Buschman, Bruce S. McEwen, and Elizabeth Gould. 2019. “Perineuronal Nets, Inhibitory Interneurons, and Anxiety-Related Ventral Hippocampal Neuronal Oscillations Are Altered by Early Life Adversity.” Biological Psychiatry 85 (12): 1011–1020.

Nabel, Elisa M., and Hirofumi Morishita. 2013. “Regulating Critical Period Plasticity: Insight from the Visual System to Fear Circuitry for Therapeutic Interventions.” Frontiers in Psychiatry 4: 146.

Nardou, Romain, Eastman M. Lewis, Roman Rothhaas, Ranran Xu, Aiqing Yang, Ed Boyden, and Gül Dölen. 2019. “Oxytocin-Dependent Reopening of a Social Reward Learning Critical Period with MDMA.” Nature 569 (7754): 116–120.

Nardou, Romain, Edward Sawyer, Young J. Song, Marie Wilkinson, Yuri Padovan-Hernandez, Junia L. de Deus, Noelle Wright, et al. 2023. “Psychedelics Reopen the Social Reward Learning Critical Period.” Nature 618 (7966): 790–798.

Nichols, David E. 2016. “Psychedelics.” Pharmacological Reviews 68 (2): 264–355.

Nour, Matthew M., Lisa Evans, David Nutt, and Robin L. Carhart-Harris. 2016. “Ego-Dissolution and Psychedelics: Validation of the Ego-Dissolution Inventory (EDI).” Frontiers in Human Neuroscience 10: 269.

Olson, David E. 2018. “Psychoplastogens: A Promising Class of Plasticity-Promoting Neurotherapeutics.” Journal of Experimental Neuroscience 12: 1179069518800508.

Olson, David E. 2021. “The Subjective Effects of Psychedelics May Not Be Necessary for Their Enduring Therapeutic Effects.” ACS Pharmacology & Translational Science 4 (2): 563–567.

Pace, Thaddeus W. W., and Christine M. Heim. 2011. “A Short Review on the Psychoneuroimmunology of Posttraumatic Stress Disorder: From Risk Factors to Medical Comorbidities.” Brain, Behavior, and Immunity 25 (1): 6–13.

Palhano-Fontes, Fernanda, Dayanna Barreto, Heloisa Onias, Katia C. Andrade, Morgana M. Novaes, Jessica A. Pessoa, Sergio A. Mota-Rolim, et al. 2019. “Rapid Antidepressant Effects of the Psychedelic Ayahuasca in Treatment-Resistant Depression: A Randomized Placebo-Controlled Trial.” Psychological Medicine 49 (4): 655–663.

Picard, Martin, and Bruce S. McEwen. 2018. “Psychological Stress and Mitochondria: A Conceptual Framework.” Psychosomatic Medicine 80 (2): 126–140.

Pizzorusso, Tommaso, Paolo Medini, Nicoletta Berardi, Sabrina Chierzi, James W. Fawcett, and Lamberto Maffei. 2002. “Reactivation of Ocular Dominance Plasticity in the Adult Visual Cortex.” Science 298 (5596): 1248–1251.

Polepally, Prabhakar R., Vivek M. Goncharov, Mark V. Nelson, Thomas E. Prisinzano, and Eduardo R. Butelman. 2021. “Salvinorin A: From Plant Origin to Designer Pharmaceutical Probe.” Bioorganic & Medicinal Chemistry 32: 116033.

Reckweg, Johannes T., Cees J. van Leeuwen, Cécile Henquet, Therese van Amelsvoort, Eef L. Theunissen, Natasha L. Mason, Riccardo Paci, Theis H. Terwey, and Johannes G. Ramaekers. 2023. “A Phase 1/2 Trial to Assess Safety and Efficacy of a Vaporized 5-Methoxy-N,N-Dimethyltryptamine Formulation (GH001) in Patients with Treatment-Resistant Depression.” Frontiers in Psychiatry 14: 1133414.

Reckweg, Johannes T., Malin V. Uthaug, Attila Szabo, Alan K. Davis, Rafael Lancelotta, Natasha L. Mason, and Johannes G. Ramaekers. 2022. “The Clinical Pharmacology and Potential Therapeutic Applications of 5-Methoxy-N,N-Dimethyltryptamine (5-MeO-DMT).” Journal of Neurochemistry 162 (1): 128–146.

Reh, Roger K., Brett G. Dias, Charles A. Nelson, Daniel Kaufer, Janet Werker, Brian Kolb, Janice K. Levenson, and Takao K. Hensch. 2020. “Critical Period Regulation Across Multiple Timescales.” Proceedings of the National Academy of Sciences USA 117 (38): 23242–23251.

Reiff, Collin M., Elizabeth E. Richman, Charles B. Nemeroff, Linda L. Carpenter, Alik S. Widge, Carolyn I. Rodriguez, Ned H. Kalin, William M. McDonald, and the Work Group on Biomarkers and Novel Treatments. 2020. “Psychedelics and Psychedelic-Assisted Psychotherapy.” American Journal of Psychiatry 177 (5): 391–410.

Ren, Peng, Jingya Wang, Nanxi Li, Guangxiang Li, Hui Ma, Yongqi Zhao, and Yunfeng Li. 2022. “Sigma-1 Receptors in Depression: Mechanism and Therapeutic Development.” Frontiers in Pharmacology 13: 925879.

Reynolds, Joseph C., Roy G. Lai, Jonathan S. T. Woodhead, James H. Joly, Cameron J. Mitchell, David Cameron-Smith, Ryan Lu, et al. 2021. “MOTS-c Is an Exercise-Induced Mitochondrial-Encoded Regulator of Age-Dependent Physical Decline and Muscle Homeostasis.” Nature Communications 12: 470.

Riba, Jordi, Marta Valle, Gloria Urbano, Marta Yritia, Adelaida Morte, and Manel J. Barbanoj. 2003. “Human Pharmacology of Ayahuasca: Subjective and Cardiovascular Effects, Monoamine Metabolite Excretion, and Pharmacokinetics.” Journal of Pharmacology and Experimental Therapeutics 306 (1): 73–83.

Ross, Stephen, Anthony Bossis, Jeffrey Guss, Gabrielle Agin-Liebes, Tara Malone, Barry Cohen, Sarah E. Mennenga, et al. 2016. “Rapid and Sustained Symptom Reduction Following Psilocybin Treatment for Anxiety and Depression in Patients with Life-Threatening Cancer: A Randomized Controlled Trial.” Journal of Psychopharmacology 30 (12): 1165–1180.

Roth, Bryan L., Karol Baner, Richard Westkaemper, Daniel Siebert, Kenner C. Rice, Sergei Steinberg, Phil Ernsberger, and Richard B. Rothman. 2002. “Salvinorin A: A Potent Naturally Occurring Nonnitrogenous Kappa Opioid Selective Agonist.” Proceedings of the National Academy of Sciences USA 99 (18): 11934–11939.

Rucker, James J. H., Catherine Roberts, Marc Seynaeve, Allan H. Young, Bruce Suttle, Toshie Yamamoto, Roger Brand, et al. 2024. “Phase 1, Placebo-Controlled, Single Ascending Dose Trial to Evaluate the Safety, Pharmacokinetics and Effect on Altered States of Consciousness of Intranasal BPL-003 (5-Methoxy-N,N-Dimethyltryptamine Benzoate) in Healthy Participants.” Journal of Psychopharmacology 38 (5): 489–502.

Seeman, Philip, Karol P. Baner, Bryan L. Roth, John A. Bunzow, and David K. Grandy. 2009. “Dopamine D2 High Receptors Stimulated by Phencyclidines, Lysergic Acid Diethylamide, Salvinorin A, and Modafinil.” Synapse 63 (8): 698–704.

Shao, Lan-Xin, Clara Liao, Ian Gregg, Pasha A. Davoudian, Neil K. Savalia, Kathryn Delagarza, and Alex C. Kwan. 2021. “Psilocybin Induces Rapid and Persistent Growth of Dendritic Spines in Frontal Cortex In Vivo.” Neuron 109 (16): 2535–2544.

Sikirić, Predrag, Sven Seiwerth, Robert Rucman, Bozo Turković, Ivica Rotkvić, Mateja Jelovac, Marko Sebecić, et al. 2020. “Brain-Gut Axis and Pentadecapeptide BPC 157: Theoretical and Practical Implications.” Current Neuropharmacology 18 (12): 1144–1166.

Sreekumar, Parameswaran G., Arjuna J. Ranjit, Allen J. Galstoyan, Ali Karimi, Vahe Garibovic, Christine Spee, David R. Hinton, et al. 2018. “The Mitochondrial-Derived Peptide Humanin Protects RPE Cells from Oxidative Stress, Senescence, and Mitochondrial Dysfunction.” Investigative Ophthalmology and Visual Science 59 (3): 1006–1017.

Stauffer, Christopher S., Joshua D. Woolley, Jamie Yokel, Steven L. Batki, Adam C. Frohe, Sasha Davis, Anne Holmes, et al. 2022. “A Two-Session, Group-Based Ketamine-Assisted Psychotherapy for PTSD with Comorbid Substance Use Disorders: A Pilot Randomized Controlled Trial.” Journal of Psychoactive Drugs 54 (5): 446–455.

Szabo, Attila, Attila Kovacs, Ede Frecska, and Eva Rajnavolgyi. 2014. “Psychedelic N,N-Dimethyltryptamine and 5-Methoxy-N,N-Dimethyltryptamine Modulate Innate and Adaptive Inflammatory Responses Through the Sigma-1 Receptor of Human Monocyte-Derived Dendritic Cells.” PLoS ONE 9 (8): e106533.

Szeto, Hazel H. 2014. “First-in-Class Cardiolipin-Protective Compound as a Therapeutic Agent to Restore Mitochondrial Bioenergetics.” British Journal of Pharmacology 171 (8): 2029–2050.

Takesian, Anne E., and Takao K. Hensch. 2013. “Balancing Plasticity/Stability Across Brain Development.” Progress in Brain Research 207: 3–34.

Thomann, Jan, Dino Luethi, Edna Grünblatt, Henriette Meyer zu Schwabedissen, and Matthias E. Liechti. 2025. “Acute Effects and Pharmacokinetics of LSD After Paroxetine or Placebo Pre-Administration in a Randomized, Double-Blind, Cross-Over Phase I Trial.” Clinical Pharmacology and Therapeutics 117 (6): 1784–1792.

Timmermann, Christopher, Leor Roseman, Luke Williams, David Erritzoe, Charlotte Martial, Helena Cassol, Steven Laureys, David Nutt, and Robin L. Carhart-Harris. 2019. “DMT Models the Near-Death Experience.” Frontiers in Psychology 9: 1424.

Transcend Therapeutics. 2025. “Transcend Therapeutics Receives Breakthrough Therapy Designation for TSND-201 (Methylone) for the Treatment of PTSD.” Press release, July 10, 2025.

Transcend Therapeutics. 2026. “IMPACT-1 Study Published in JAMA Psychiatry: Results of TSND-201 (Methylone) in Patients with Severe PTSD.” Press release, February 18, 2026.

Vargas, Maxemiliano V., Lindsay P. Cameron, Calvin Ly, Eden V. Barragan, Zeinah Bashir, Hannah Saeger, Sammy Patel, et al. 2025. “Emerging Mechanisms of Psilocybin-Induced Neuroplasticity.” Trends in Pharmacological Sciences46 (10): 859–874.

Venturino, Alessandro, Rouven Schulz, Joana De Jesús-Cortés, Manuela H. Maes, Sébastien Nagy, Frances Reilly-Andújar, Joao Colaco-Mota, et al. 2021. “Microglia Enable Mature Perineuronal Nets Disassembly Upon Anesthetic Ketamine Exposure or 60-Hz Light Entrainment in the Healthy Brain.” Cell Reports 36 (1): 109313.

Volkova, A., Mikhail Yu Bobrov, Vladimir P. Lebedev, Vladimir Yu Pankratov, and Lyubov Yu Volkova. 2016. “Selank Administration Affects the Expression of Some Genes Involved in GABAergic Neurotransmission.” Frontiers in Pharmacology 7: 31.

Warner-Schmidt, Jennifer L., Christopher Pittenger, Matthew Stogniew, Blake Mandell, Sarah J. Olmstead, and Benjamin Kelmendi. 2022. “Methylone, a Rapid Acting Entactogen with Robust Anxiolytic and Antidepressant-Like Activity.” Frontiers in Psychiatry 13: 1041277.

Watts, Rosalind, and Jason B. Luoma. 2020. “The Use of the Psychological Flexibility Model to Support Psychedelic Assisted Therapy.” Journal of Contextual Behavioral Science 15: 92–102.

Weiss, Brennan, Roland R. Griffiths, Frederick S. Barrett, Katrin H. Preller, Manoj K. Doss, Albert Garcia-Romeu, and Matthew W. Johnson. 2025. “Psilocybin Enhances Hippocampal Neurogenesis and Reverses Stress-Induced Synaptic Deficits in Murine Models.” Translational Psychiatry 15: 142.

Wilkinson, Samuel T., Sanjay J. Mathew, John H. Krystal, Steven D. Hollon, Carlos A. Zarate, and James W. Murrough. 2018. “The Effect of a Single Dose of Intravenous Ketamine on Suicidal Ideation: A Systematic Review and Individual Participant Data Meta-Analysis.” American Journal of Psychiatry 175 (2): 150–158.

Williams, Nolan R., Boris D. Heifets, Carolyn Blasey, Karen Sudheimer, Justin Pannu, Heather Pankow, Jessica Hawkins, et al. 2018. “Attenuation of Antidepressant Effects of Ketamine by Opioid Receptor Antagonism.” American Journal of Psychiatry 175 (12): 1205–1215.

Yokoyama, Jennifer S., Iris J. Broce, Erika Sant’Ana, Christine M. Williams, Kurtresha Worden, Kaitlin B. Casaletto, Cleo Liang, et al. 2024. “KLOTHO KL-VS Heterozygosity Is Associated with Diminished Age-Related Neuroinflammation, Neurodegeneration, and Synaptic Dysfunction in Older Cognitively Unimpaired Adults.” Alzheimer’s & Dementia 20 (6): 4012–4024.

Zhao, Weixing, Zhipeng Xu, Jiangbei Cao, Qiang Fu, Yishuang Wu, Xiaoying Zhang, Yue Long, et al. 2019. “Elamipretide (SS-31) Improves Mitochondrial Dysfunction, Synaptic and Memory Impairment Induced by Lipopolysaccharide in Mice.” Journal of Neuroinflammation 16: 230.

Zhu, Yihao, Handong Wang, Jiang Fang, Wei Dai, Jiang Zhou, Xiaoliang Wang, and Mengliang Zhou. 2018. “SS-31 Provides Neuroprotection by Reversing Mitochondrial Dysfunction After Traumatic Brain Injury.” Oxidative Medicine and Cellular Longevity 2018: 4783602.

On the evening of February 25, 2026, in a converted warehouse in San Francisco’s Dogpatch neighborhood, several dozen artificial intelligence employees and civil society advocates gathered under a disco ball, La Croix cans in hand, to hear the Democratic pollster David Shor explain how their industry was about to destroy the American working class.

Shor had data. Seventy-nine percent of voters were worried about “the government not having a plan to protect workers.” Seventy-two percent were concerned that AI was going to drive down wages “for people like you.” A federal jobs guarantee polled well; universal basic income did not. The phrase “permanent underclass” tested better than “structural unemployment.” The single most effective political ad his firm had focus-grouped, against billion-dollar opposition spending, ended with the line: They work for the bots. We work for you.

The audience perched on benches and beanbags. The presentation closed with a fundraising pitch, for the price of one hour of what the AI industry was spending that year on data center construction, the room could equip Democratic candidates with the messaging architecture for the coming jobs apocalypse.

This is the scene Jasmine Sun reports, with anthropological precision, near the end of her New York Times essay published this week. It is, on its face, an unremarkable San Francisco fundraiser. It is also, read carefully, the most analytically useful image of the present moment in American political economy. What is being held in that warehouse, under that disco ball, is the bourgeoisie of a particular industry adjusting to its own coming redundancy, by purchasing political messaging consultants. It is Versailles in Allbirds. The aristocracy is throwing a fundraiser to soften its own descent.

That this scene is possible at all, that the people building the technology that is about to render their fellow citizens unemployable are simultaneously the people most actively organizing the political response to that unemployability, tells you everything you need to know about the political moment. The class that is producing the catastrophe is also producing the discourse that frames the catastrophe, the policy proposals that respond to the catastrophe, and the political consultants who will sell the response to the public. The catastrophe is not a market failure. It is the market succeeding at exactly what it is designed to do, in the presence of a technology that has finally made labor itself optional. And the people gathered in the warehouse understand this, in their bones, which is why they are there.

Sun’s piece is, in its way, a remarkable document. She has reported, on the record and from the inside, what the most powerful technology firms in the country actually believe about the future of work. The “San Francisco consensus,” in her formulation, is bleak: that advanced AI will soon surpass human capabilities; that this will displace millions of jobs; that the technology will depress economic mobility and exacerbate inequality; that wealth and power will accrue to AI companies and to the existing owners of capital; and that none of the people building this technology have any clear idea what to do about it. The Anthropic CEO Dario Amodei is quoted predicting “an unemployed or very-low-wage ‘underclass’” of people with “lower intellectual ability,” a phrase to which we will return. Sam Altman is quoted predicting that the technology he is building will “shift power from labor to capital,” which is approximately what the Communist Manifesto said capital was already doing in 1848. Jack Dorsey, after firing nearly half of Block’s employees in March, told Wired that coding agents had “presented an option to dramatically change how any company is structured.” The market rewarded him with a twenty-five percent surge in after-hours trading.

What Sun has captured, in other words, is the moment at which capital’s executives announce, on the record, that they have built the machine that will end the historical compromise between labor and capital that has been the foundation of every developed economy since 1945. They are announcing it because they cannot keep it secret any longer. The displacements have begun. The signals are visible in the labor market. The college graduates are sending out hundreds of resumes and receiving no replies. The software engineers are joking, with the gallows humor that always precedes mass unemployment in a particular sector, about replacing themselves with the tools they are building. The article quotes Steven Adler, a former safety researcher at OpenAI, on the dissonance: “Even if someone thinks it is personally distasteful to make money from building technology that companies say may literally kill everyone, many people are just cogs in the machine.”

Cogs in the machine. That is the phrase, and the cogs know it.

What Sun’s piece does not do, what it structurally cannot do, given its venue, is name what it has reported. It hovers around the phenomenon, gathers the evidence, presents the policy options as they are currently being developed inside the AI industry’s own legitimation apparatus, and stops short of the analysis that would render the evidence intelligible. This essay is an attempt to perform the analysis the Times declined to publish. It is also an attempt to argue that the political response now circulating among Silicon Valley’s house Democrats, the public wealth fund, the thirty-two-hour workweek, the corporate tax increase, the equity stake for citizens, is not the radical horizon it presents itself as. It is the liberal response, the bourgeois response, the response calibrated to preserve the property relations that produced the catastrophe while redistributing some fraction of the proceeds. The actually radical response, the one that has not yet been articulated in mainstream American political discourse, but whose shape is becoming visible through the work of figures like Zohran Mamdani, is structurally different. It does not redistribute the rents of AI. It democratizes the means of producing them.

Let’s take this in stages. First, what Marx anticipated and what contemporary Marxism extends. Second, what Sun and her sources are actually describing, the technical and economic mechanism by which AI is producing the underclass they fear. Third, why Amodei’s “lower intellectual ability” formulation is so significant, what its genealogy is, and why no one in the Times noticed it. Fourth, why the entire discourse Sun reports on: the “San Francisco consensus,” the OpenAI white paper, the Anthropic Institute, the Shor presentation, is structurally a legitimation apparatus, and why every policy it produces will fail. Fifth, what democratic socialism actually proposes, why Mamdani’s victory is its first political form adequate to the moment, and what an AI-specific democratic-socialist program would have to include. The argument will arrive at a single conclusion: the alternatives to democratic ownership are not just morally inferior. They are structurally incoherent. They cannot solve the problem they claim to address, because they preserve the relation that produced the problem.

I want to be honest about the stakes before we begin. The window in which the political response is determined is not long. Sun’s reporting confirms that the AI industry is preparing, now, to underwrite a particular set of liberal policy responses, the public wealth fund, the equity stake, the modest tax, and to deploy political consultants to sell those responses to the public before any more radical alternative can take shape. The argument of this essay is that those responses are, by design, inadequate. They cannot do what they claim to do. And the next eighteen months, the period leading into the 2028 Democratic primary, are the period in which it will be determined whether the political left in this country can articulate, with the clarity Sun’s anonymous AI researchers have already articulated to her, what is actually happening, and what would actually have to be done.

Let us begin where the analysis begins.

The Fragment on Machines

Karl Marx wrote a passage in 1858, in the notebooks that would later be published as the Grundrisse, that has become known as the Fragment on Machines. It is, by my reading, the single most prescient piece of political economy ever produced, and it has been waiting one hundred and sixty-eight years for the moment that David Shor was describing under the disco ball.

Marx was working out the logic of capital’s relationship to labor and was confronting a paradox. Capital, in his analysis, depends on labor for the production of surplus value. The worker produces value; the capitalist appropriates a portion of it; that appropriation is the mechanism by which capital accumulates and reproduces itself. But capital is also, Marx noticed, in a constant race to replace labor with machinery. Every capitalist faces a competitive incentive to reduce the amount of labor required per unit of output. Across the system, this produces what Marx called the rising organic composition of capital, more machines, less labor, per dollar of investment.

The contradiction is structural. Capital’s individual incentive is to eliminate labor. But labor is the source of the surplus value capital depends on. As the system progresses, capital approaches a limit at which it has reduced labor to a vanishing fraction of total production, at which point, Marx argued, the law of value itself begins to break down. The system reaches a configuration in which “labour time ceases and must cease to be its measure, and hence exchange value [must cease to be the measure] of use value.” This is the famous passage about the worker’s relationship to production: “Labour no longer appears so much to be included within the production process; rather, the human being comes to relate more as watchman and regulator to the production process itself.”

Marx wrote this in 1858 about industrial machinery in nineteenth-century Manchester. The worker, in his vision, would still be needed as watchman and regulator. The machines were powerful, but they were not yet capable of watching and regulating themselves.

We are now in the moment in which they are.

Read Sun’s article through the Fragment on Machines lens and what is happening becomes legible in a way that the article itself cannot make it. The “permanent underclass” is not a sociological category. It is the working class at the moment when capital no longer requires it as a source of surplus value. The “white-collar bloodbath” Amodei predicts is the watchman and regulator, the knowledge worker, the one Marx thought would be the residual category, being eliminated in turn, because the systems can now watch and regulate themselves. What we are watching is not a single industrial transition. It is the structural completion of the process Marx identified, the moment at which capital’s productive forces have developed to the point at which capital itself begins to lose its rationale, because the social form of capital is the wage relation, and the wage relation requires labor that can be exploited.

The system, of course, is not going to dissolve gracefully into post-scarcity communism just because Marx anticipated this. Capital is going to do what capital always does at moments of structural crisis. It is going to attempt to preserve the existing property relations through political means. It is going to reorganize the relationship between capital and the state to underwrite the value of the assets it holds, even when those assets no longer produce surplus value through the exploitation of labor in the classical sense. It is going to extract rents, from data, from infrastructure, from intellectual property, from regulatory capture, at scales that make the industrial fortunes of the nineteenth century look modest. And it is going to require, in order to do this, that the dispossessed population be kept quiescent through some combination of welfare provision, ideological capture, and outright repression.

This is the moment Sun is describing without describing it. The AI industry’s executives understand, at some level, that they are presiding over the structural completion of a process Marx identified in 1858, and they understand that the political consequences of that completion are unmanageable through normal liberal-democratic means. Amodei, in the longer essay Sun cites, is explicit about this: “The balance of power of democracy is premised on the average person having leverage through creating economic value. If that’s not present, I think things become kind of scary.” This is not a technologist’s worry. It is a class analysis, articulated by a member of the class that is producing the condition. He is correct. Democracy is premised on the average person having economic leverage. When that leverage disappears, democracy as we have known it cannot persist. What replaces it is one of two things: a regime in which capital openly governs without the legitimating fiction of popular consent, or a regime in which the dispossessed have organized themselves into a political force capable of seizing the productive apparatus and democratizing it. There is no third option, because the wage relation that produced the historical compromise between capital and labor is the thing that is being technologically dissolved.

This is the analytical core that Sun’s article does not reach. Her piece treats the situation as a policy problem, what should we do about AI displacement?, when it is, in fact, a structural crisis of the social form of capital itself. The policy question is downstream of the structural question. And until the structural question is articulated, no policy response can be adequate to the situation, because every policy response will be calibrated to preserve the property relations that are producing the crisis.

This is what contemporary Marxist analysis does that liberal political economy structurally cannot. It refuses to take the existing property relations as given. It asks: what is the social form? What does the system require in order to reproduce itself? What happens when the conditions of reproduction can no longer be met? The Fragment on Machines is not a prophecy. It is an analytical instrument. And applied to Sun’s evidence, it identifies the moment we are in: the social form of capital has reached the point at which the productive forces it has developed cannot be reconciled with the ownership relations under which they were developed. Either the ownership relations change, or the productive forces are deployed against the population that, under the old relation, constituted capital’s reason for existing.

We have a name, in Marxist analysis, for this configuration: real subsumption arriving at its terminus. Real subsumption is the stage at which capital does not merely take over the labor process but reorganizes it according to its own logic, when the assembly line replaces the workshop, when the corporation replaces the family firm, when the algorithm replaces the manager. The terminus of real subsumption is the moment at which capital can dispense with the labor process altogether, retaining the worker only as a residual cost item, awaiting deletion. We are there. The “white-collar bloodbath” is not a metaphor. It is the structural consequence of having reached this terminus.

What contemporary Marxist analysis adds, beyond the basic Marxian framework, is the analysis of the subjective and cultural mechanisms through which capital secures the consent of the populations it is in the process of dispossessing. Mark Fisher’s Capitalist Realism describes the atmospheric condition: the unconscious certainty that there is no alternative to the present arrangement, that even the disasters the system produces are preferable to the disasters that would attend trying to change it. Franco Berardi’s analysis of semiocapitalism describes how capital, in its post-Fordist phase, has begun to extract value not just from labor power but from the semiotic capacity of the human mind itself, from attention, from affect, from the production of meaning. Byung-Chul Han’s analysis of the achievement subject describes the internalization of capital’s demands by the workers themselves, who exhaust themselves in the pursuit of an optimization that only deepens their precarity. Gilles Deleuze, in his “Postscript on the Societies of Control,” describes the institutional form of present-day power: not the disciplinary enclosure of the factory but the modulation environment of continuous evaluation, the dividual replacing the individual.

Each of these thinkers describes a different facet of the same configuration. The achievement subject, the eighty-hour-a-week Anthropic engineer Sun describes, running multiple Claude Code agents simultaneously to keep “someone, human or machine, always on the clock,” is the subjective form of real subsumption. The semiocapitalist firm, the AI lab whose product is meaning generation at scale, is the institutional form. Capitalist realism is the cultural condition that makes the dispossession appear unavoidable. The dividual is the political subject who results: a data profile, continuously optimized, never finished, structurally incapable of the collective action that would resist the dispossession.

What is happening, in other words, is the conjunction of the structural-Marxian crisis (the terminus of real subsumption) with the cultural-Marxian conditions (the subjective and ideological forms that prevent organized response). This is what makes the present moment so dangerous. The objective conditions for a political-economic crisis of unprecedented magnitude have arrived. The subjective conditions for organized response have, simultaneously, been hollowed out by forty years of neoliberal restructuring and ten years of platform capitalism. The dispossessed cannot easily organize because the cultural infrastructure within which class-based political organization once occurred, the union, the political party, the ethnic-religious-civic association, the geographically proximate community, has been substantially dismantled by the very same technological forces that are now completing the dispossession.

This is the position. This is where we are. Sun’s article, read straight, is a description of capital’s executives recognizing the position and beginning, in their fragmentary and contradictory way, to plan for it. The remainder of this essay is about what that plan is, why it is inadequate, and what would actually have to be done.

The Mechanism

Before we get to the Amodei question, we need to be clear about what the technology actually does, because the technical details matter for the political analysis. Sun is a careful reporter, and her piece contains the relevant evidence. We can work from it.

The benchmarks tell the story. The AI Productivity Index measures frontier-model performance against four occupations: investment banking associate, management consultant, Big Law associate, primary care physician. OpenAI’s GDPVal benchmark covers forty-four occupations, from real estate broker to news analyst. These are economic-utility benchmarks, calibrated against the hourly cost of the human labor they are designed to replace. Tejal Patwardhan, who leads frontier evaluations at OpenAI, told Sun that her models had moved, in a few months, from being “not yet on par with human experts” to “over an 80 percent win rate” against them. Eighty percent. That is not the rate of incremental productivity gains characteristic of previous technological transitions. That is the rate at which a new technology renders an existing labor category economically obsolete.

The mechanism is direct. A firm employing one hundred associates at a major law firm, or a hundred junior bankers, or a hundred customer service representatives, can now perform the same volume of work with twenty associates and an enterprise license to a frontier model. The economics are crushing. The license is cheaper than the salaries by orders of magnitude. The legal liability is comparable, as long as a credentialed human signs off at the end. The output is, by an increasing range of measures, at parity. There is no individual firm-level decision available except to lay off the eighty unnecessary associates, because any firm that does not lay them off is structurally outcompeted by the firms that do. This is the Jack Dorsey calculation, performed in the context of every white-collar industry, simultaneously, over a period of perhaps three to five years.

What happens to the eighty laid-off associates is the political question. There are a few things they can attempt. They can retrain into a different field, but the same dynamic is operating in adjacent fields, so the retraining target keeps moving. They can move to fields where AI is structurally less competitive: the skilled trades, care work, certain forms of in-person service, but those fields are crowded and pay substantially less than what they were earning, and the flood of refugees from the white-collar economy will further depress wages in the receiving sectors. They can attempt entrepreneurship, which is what the warehouse-fundraiser audience is doing, but the path of starting a successful firm is now contested by AI-native competitors who can operate at a fraction of the headcount. Or they can fall out of the labor force, which the macroeconomic data already shows is happening at the entry level.

The labor share of national income, which has been declining since the late 1970s as a result of globalization, automation, and the dismantling of organized labor, is going to fall off a cliff. Sun cites the obvious comparison: the late-1990s and 2000s “China shock,” in which entire American manufacturing communities were hollowed out by the offshoring of production. That episode produced the political conditions for Donald Trump in 2016, the opioid epidemic, a measurable decline in life expectancy among non-college-educated white Americans, and the most sustained rise in deaths of despair the country has seen since the early twentieth century. The China shock unfolded over roughly fifteen years and affected approximately five percent of the American workforce. The AI shock is being predicted, by the people building it, to unfold over roughly three to five years and affect somewhere between thirty and fifty percent of the workforce. We are talking about a labor market disruption an order of magnitude larger than the China shock, compressed into a quarter of the time, hitting the educated population that has been told for thirty years that education was its protection against exactly this kind of disruption.

The macroeconomic consequences are also straightforward, even if the AI industry’s house economists are reluctant to draw them. Aggregate demand is a function of wages. If wages are collapsing and the labor share is plummeting, aggregate demand will contract, unless something replaces the lost wages as a source of consumer spending. In the short run, fiscal stimulus and welfare programs can backstop demand. In the medium run, they cannot, because the political economy of fiscal redistribution requires a productive working population willing to vote for the taxation that funds it, and what AI is producing is precisely a non-productive working population whose political economy points in the opposite direction. There is also the question of who the AI economy is for, if the population that traditionally constituted its market has been rendered unable to pay for its products. This is the contradiction that the most thoughtful AI executives are beginning, very quietly, to articulate. If you have built a productive machine that no longer needs human labor and no longer needs human consumers because the humans cannot afford anything, what, exactly, is the machine producing for?

The answer, increasingly, is: for capital itself, and for the state that has been captured to serve capital. The St. Louis Fed analysis Sun cites, that AI-related investments accounted for thirty-nine percent of US GDP growth in the first three quarters of 2025, is the leading edge of a configuration in which the productive economy is decoupling from the consumption economy. The data centers are being built. The chips are being designed. The models are being trained. The infrastructure is being constructed at scale. None of this is producing consumer goods in the direct sense. It is producing the productive apparatus that will, capital hopes, produce the consumer goods of the future. But the consumer goods of the future will be produced for whom? At what price? Paid for with what wages, by what workers?

The honest answer, which capital’s executives are not yet willing to give in public, is that the AI economy is producing the conditions for what some of them are now describing as “post-labor capitalism,” an economic configuration in which the productive apparatus is owned by a small fraction of the population, the productive output is consumed largely by that same fraction at scales that make the present luxury market look modest, and the rest of the population lives at the edge of subsistence on whatever welfare provisions the state can be persuaded to fund. This is feudal in form. It is not socialism. It is not communism. It is not even a conventional capitalist crisis. It is the structural reorganization of the political economy around an ownership stratum that no longer requires a working population for any of the functions that have historically defined capital’s relationship to labor.

The question Sun’s article hovers around but does not name is whether the population being thus dispossessed will accept the configuration. There are reasons to believe they will not. There are also reasons to believe, and the warehouse fundraiser is one of them, that capital is preparing, now, to manage the consent process through political messaging, modest welfare provision, and the controlled inflation of a redistributive discourse calibrated to absorb the discontent without altering the ownership relation that produced it. We are about to examine how that management apparatus operates.

But first, we need to deal with what Amodei said.

The Phrase

Dario Amodei, the chief executive of Anthropic, the AI company that is currently the most prestigious and the most ostentatiously safety-conscious in the industry, published an essay in January of this year. The essay is roughly twenty thousand words long. It addresses, among other things, the long-term economic consequences of the technology Anthropic is building. In the essay, Amodei warns that AI may create, and I am quoting exactly, from the New York Times article, “an unemployed or very-low-wage ‘underclass’” for people with “lower intellectual ability.”

I want to sit with that phrase for a moment. Lower intellectual ability. These are the four words a frontier AI CEO chose, in the year 2026, to describe the population that his technology is going to render economically redundant. Sun reports the phrase in passing. The Times did not editorialize about it. None of the people who have written about Sun’s piece in the days since publication have, as far as I can determine, drawn attention to it. The phrase has passed into the ordinary discourse of American technology journalism without producing any of the response it would have produced if it had been uttered in approximately any other context in the last fifty years.

It would be useful to examine the genealogy of this phrase, because the phrase has a history, and the history is not, as the Times would have it, ancient. The notion that human beings can be ranked along a single axis of “intellectual ability,” and that the resulting ranking is biologically real and economically determinative, is not a new idea. It is the idea that animated Francis Galton’s Hereditary Genius in 1869, the founding text of modern eugenics, in which Galton argued that intelligence was inherited and that society should be reorganized to maximize the breeding of the intelligent and minimize the breeding of the unintelligent. It is the idea that animated Charles Davenport and the American eugenic movement in the 1910s and 1920s, which produced the Immigration Restriction Act of 1924 and the involuntary sterilization programs that operated in thirty-two American states until the 1970s, and that sterilized roughly sixty thousand Americans for the offense of having scored badly on intelligence tests of varying scientific legitimacy. It is the idea that animated the German Nazi Party’s Rassenhygiene program, which took the American eugenic literature as its explicit model and which carried the logic to its endpoint at Auschwitz. It is the idea that Richard Herrnstein and Charles Murray defended in The Bell Curve in 1994, an argument that was understood at the time, by every serious reader, to be a reformulation of eugenic claims for a post-civil-rights political environment, and that produced an entire policy literature on the supposed cognitive limits of welfare recipients.

What is new, and this is the point that needs to be named directly, is that over the past decade, a particular subculture has emerged within Silicon Valley that has explicitly revived this genealogy and that has provided the intellectual environment in which the phrase “lower intellectual ability” can appear in a frontier-AI CEO’s essay without being immediately recognized for what it is. The figures are public. Curtis Yarvin, the neoreactionary blogger whose work explicitly defends a return to hereditary aristocracy, has become a fixture of Silicon Valley intellectual life and is read approvingly by, among others, J. D. Vance and Marc Andreessen. Richard Hanania, who began his career writing for white-nationalist publications under a pseudonym and who has built a substantial Substack readership defending the genetic basis of inequality, is a regular figure in the venture capital intellectual ecosystem. The pronatalist movement around figures like Malcolm and Simone Collins, who advocate explicitly for selective breeding to raise population intelligence, has received favorable coverage in major business publications and substantial investment from prominent technology figures. The longtermist philanthropic apparatus, funded heavily by Open Philanthropy and the now-collapsed FTX Future Fund, has provided institutional sponsorship to research on “dysgenic decline” and the cognitive characteristics of populations, including work that draws explicitly on the writings of Richard Lynn, the late British psychologist whose career was devoted to ranking the average IQ of the world’s racial and national groups, and whose data sets continue to be cited in mainstream AI policy work despite their well-documented methodological corruption. Nick Bostrom, whose 2014 book Superintelligence is among the most-cited works in the AI safety literature and whose Future of Humanity Institute at Oxford was foundational to the longtermist movement, was forced in 2023 to apologize for a 1996 email in which he used a racial slur and defended the proposition that Black Americans are, on average, less intelligent than white Americans. The apology did not result in his removal from any of the institutions through which longtermism has shaped contemporary AI policy.

This is the intellectual environment in which Amodei’s sentence was written. Anthropic was founded in 2021 by former OpenAI safety researchers explicitly aligned with longtermist philosophy. Its early funding included substantial contributions from longtermist-affiliated philanthropic vehicles. Its research culture has been shaped, in part, by the academic and intellectual networks that sustain the figures named above. None of this means that Dario Amodei holds, or has expressed, any of the explicit positions of the figures listed. It means that the phrase “lower intellectual ability,” as it appears in his essay, did not arrive there from nowhere. It arrived there from a discursive environment in which the proposition that human worth and economic value are determined by a heritable cognitive trait has become, over the past ten years, conversationally available in a way that it has not been in mainstream American intellectual life since the early 1960s.

I am not claiming that Dario Amodei privately holds the explicit views of Hanania or Yarvin. I am claiming something more precise, and more damaging. I am claiming that the phrase “lower intellectual ability” is doing specific work in his sentence, that the work it is doing is eugenic in structure whether or not Amodei intends it to be, and that the structural function of the phrase is identical regardless of the interior state of its author. The “underclass” is not, in Amodei’s formulation, a class produced by the technology Anthropic is building. It is a class that already exists, defined by a biological characteristic, that the technology will simply make visible by rendering its members economically redundant. The technology is the diagnostic instrument. The biology is the cause. This is, formally and structurally, the eugenic claim. It is the claim that economic position reflects biological capacity, and that the dispossessed are dispossessed because of who they are rather than because of what has been done to them. The fact that it can be written in a major technology executive’s public essay, in 2026, and reproduced in the New York Times without editorial comment, is not a fact about Amodei. It is a fact about how thoroughly the discursive environment of contemporary American technology capital has been shaped by the figures and institutions named above, to the point at which a position that would have been recognized as eugenic by any reader in 1965 is now recognized as a thoughtful contribution to the discourse on AI policy.

This is the same move that nineteenth-century industrialists made when they explained the existence of the urban poor by reference to the moral and intellectual deficiencies of the poor themselves rather than by reference to the property relations that had dispossessed them. It is the same move that twentieth-century segregationists made when they explained the position of Black Americans in the labor market by reference to “ability” rather than to four hundred years of slavery, peonage, and discrimination. It is the move that always gets made when a powerful class needs to explain the suffering produced by its own arrangements as a fact of nature rather than a consequence of policy. And it is being made, in the year 2026, by the chief executive of one of the three most powerful AI laboratories in the world, in the pages of his publicly available essay, and being reported by the New York Times without comment.

What Amodei should have written, if he were being honest, is something like this: “The technology my company is building is going to render economically redundant a portion of the population whose labor is currently being substituted by our machines. The size of this portion will grow as the technology improves. The portion will eventually include almost everyone, because the technology will eventually substitute for almost all forms of human labor. This is not a fact about the population’s intellectual ability. It is a fact about the productive forces my company controls and the property relations that determine who benefits from those productive forces.”

That is the honest version. It is also the version that names the problem in a way that requires a structural response. The “lower intellectual ability” version naturalizes the dispossession. It locates the cause in the population rather than in the technology and its ownership. And, this is the political point, it lays the groundwork for a particular policy response, the response of providing welfare provision to a population that is presumed to deserve it because of its biological limitations rather than because the productive apparatus has been organized to dispossess it.

This is the rhetorical first move in what is going to be a sustained ideological project. The AI industry needs to explain, to itself and to the public, why the people it is dispossessing deserve what is happening to them. The classical capitalist explanation, that the displaced workers are lazy, or undeserving, or insufficiently entrepreneurial, does not work in this case, because the displaced workers include lawyers and doctors and engineers and middle managers, the population that has spent thirty years being told that their education and their willingness to work hard were the guarantee against exactly this kind of displacement. A new explanation is required. And the explanation that is being assembled, in pieces, in the public essays of figures like Amodei and in the technical papers that anchor the AI productivity benchmarks, is that the displaced population was always going to be displaced because the displaced population was, in some essential biological-cognitive sense, less. The technology is not creating the underclass. It is revealing the underclass that nature has already created.

This is the function of the phrase. It is preparing the ideological ground for a configuration in which the dispossessed are receiving welfare provision not as a matter of political-economic right but as a matter of biological pity. The pity will be modest, because pity always is. The welfare will be conditional, because biological pity always is. The dispossessed will be told, in a hundred subtle ways, that the technology has shown them what they truly were, and that their position is the natural consequence of their nature rather than the consequence of an economic arrangement that could be changed by political action.

This is the moment to describe what the Anthropic Institute actually is. Sun reports that Anthropic established the Anthropic Institute in March of 2026, to “house its teams working on economics, societal impact and frontier safety.” The institute is led by Jack Clark, the company co-founder, and it has been positioned in Sun’s piece as the more thoughtful, the more critical, the more honest of the AI industry’s research apparatuses. There is a certain truth to this, in the sense that Anthropic researchers are more willing to publish findings that show the technology’s limits and harms than their counterparts at OpenAI. The “light and shade” framing Sun cites is real. The papers on overuse of AI tools by junior engineers, on the unhealthy delegation of personal decisions to chatbots, on the labor market consequences of model deployment, these are real research outputs, and they are more honest than what comes out of the firm’s larger competitor.

But the function of the institute, structurally, is the same function that the Chacruna Institute and its analogues serve in the psychedelic field. (I have written about this elsewhere, in the seven-part series on the political economy of psychedelic enclosure, and the parallel is exact, not metaphorical.) Chacruna hosts the indigenous critiques, the ethical concerns about commodification, the panels on benefit-sharing and reciprocal governance, the keynote addresses by source-community representatives. the entire discourse of acknowledgment that the underlying business model of psychedelic medicine requires in order to function without legitimation crisis. The conferences provide the discourse. The FDA priority vouchers and venture-funded clinical trials distribute the property. The two operations together constitute an apparatus that allows the founding move, the abstraction of the chemical from the culture, the patenting of indigenous knowledge by firms that exclude the source communities, to operate at scale, in front of audiences that include source-community representatives whose presence is being used to legitimate exactly the operation that excludes them. The model has been refined over a decade. It works. The deployment proceeds. The discourse flourishes. The communities remain unpaid.

The Anthropic Institute is the equivalent apparatus for the AI economy, and it has been assembled with the same structural function in mind. It hosts the discourse of concern, the worries about underclass formation, the careful papers on labor market effects, the philanthropic donations, the public pronouncements from the CEO about democratic erosion, that the underlying business model requires in order to function without legitimation crisis. The institute publishes the research. The company deploys the technology. The two operations together allow the founding move, the deployment of automation technology under conditions of private ownership, to operate at scale, with rhetorical legitimacy, in front of audiences that include the very populations being dispossessed. Jack Clark gives the careful interviews. Dario Amodei publishes the twenty-thousand-word essays. Anthropic researchers produce the papers documenting the concerning trends. Anthropic employees pre-commit billions in personal philanthropic donations to nonprofits dedicated to addressing the harms of AI. The ads run, the policies get drafted, the warehouse fundraisers happen. The technology is deployed on schedule. The employees feel their conscience has been honored because the institute exists. The public is informed that the industry is “thinking carefully” about the consequences. The deployment proceeds.

This is not an accusation of bad faith. Most of the people working at the Anthropic Institute are genuinely concerned about what is happening, and the work they are producing is the most honest research in the field. The point is structural. The institute exists, regardless of the intentions of its researchers, to perform the legitimation function that the company’s business model requires. And as long as the business model remains what it is, the production of automation technology under conditions of private ownership, deployed at scale into a labor market without democratic governance over the deployment, the legitimation function will be performed, the rhetoric of concern will continue to be produced, and the substantive change will continue not to occur.

Sun, to her credit, sees some of this. She writes that “Anthropic’s coffers probably won’t be emptied in the service of public work-force programs unless politicians compel the company to do so.” She notes that Anthropic has not yet released a set of policies it supports. She notes that Jack Clark “demurred” when asked whether the institute would lobby for redistributive measures. The structural picture is visible in her reporting if you look for it. The institute generates the discourse. The discourse circulates. The deployment continues. And the hour at which the company will be “compelled” by politicians to fund public workforce programs is the hour at which a political force has emerged outside the AI industry that is capable of compelling it. The institute itself is not, structurally, going to produce that force. It cannot. Its function is to manage the discontent, not to mobilize it.

This is what we are dealing with. A frontier-AI CEO has articulated a eugenic framing, eugenic in structure, regardless of his interior intent, of the population his technology is dispossessing. His company has established a research institute whose function is to host the discourse of concern that the deployment of his technology requires for legitimation. The institute is more honest than its competitors, which actually makes it more effective as a legitimation apparatus, because honest legitimation is the kind that wears longest. The same employees who, in private, express anxiety about the magnitude of the dispossession they are producing are deploying, in public, the institutional discourse that permits the dispossession to continue. The same researchers who produce the careful papers documenting the harms are taking the venture compensation that depends on the harms continuing. The same CEO who warns about democratic erosion is the CEO whose company has now, by his own implicit admission, produced the technology that erodes democracy. And this is being reported, with appropriate journalistic balance, in the New York Times, by a thoughtful writer who has correctly identified the symptoms but cannot, for reasons of venue, name the structure.

The structure, in plain language, is this. A small ownership class is producing a technology that will dispossess the bulk of the working population. The class understands what it is doing. The class is preparing, in public, the rhetorical and policy framework that will manage the political consequences of the dispossession. The framework includes a soft eugenic naturalization of the dispossession (the population was always going to be displaced, because of who they are), a research apparatus that performs the concern (the institutes, the white papers, the careful essays), a redistributive package calibrated to absorb the political pressure (the public wealth fund, the thirty-two-hour week, the jobs guarantee), and a political consultancy operation funded by the industry itself to sell the package to the dispossessed before any alternative can be articulated. This is what is being assembled, in real time, in the warehouses and the institutes and the policy offices and the New York Times. It is being assembled now because the dispossession is beginning now, and the framework has to be in place before the political response to the dispossession crystallizes into something the framework cannot manage.

We are in a particular moment in American political-economic history. Let us continue.

The Thinkable

The policy proposals Sun catalogues, in the central sections of her article, deserve careful attention, because they constitute what is currently being articulated, inside the AI industry’s own legitimation apparatus, as the radical horizon of the thinkable.

The OpenAI white paper, “Industrial Policy for the Intelligence Age,” published in April 2026, proposes a thirty-two-hour workweek, higher corporate and capital gains taxes, and a “public wealth fund” that provides all citizens an equity stake in AI companies. Sun describes these as “radically progressive” ideas. The Anthropic equivalent has not yet been published, though Jack Clark has gestured toward redistributive measures. The Shor presentation in the warehouse pushed toward a federal jobs guarantee, populist messaging against AI billionaires, and the strategic deployment of “they work for the bots, we work for you” as a unifying frame.

Each of these is, in its way, a serious proposal. The thirty-two-hour workweek would, if implemented, redistribute productivity gains from owners to workers. Higher capital gains taxation would slow the upward concentration of wealth. A public wealth fund would, if structured well, provide a stream of dividend income to the population. A federal jobs guarantee would establish a floor under labor market participation. These are not nothing. They are substantially more than what the Democratic Party has been willing to articulate in any of the last five presidential cycles, and they represent, to the credit of the people developing them, a recognition that the standard liberal toolkit of education, retraining, and human capital investment is going to be inadequate to what is coming.

But, and this is the analytical point, they are also, every single one of them, calibrated to preserve the property relations that are producing the catastrophe. They redistribute the rents. They do not change who owns the productive apparatus. The public wealth fund holds equity; it does not control governance. The thirty-two-hour workweek shortens the period of wage labor; it does not alter the wage relation. The corporate tax increase takes a larger share of the profits; it does not transfer ownership of the firms producing them. The jobs guarantee provides employment of last resort; it does not democratize the productive sector. Each of these proposals, in its structural form, accepts that the AI industry will continue to be owned and governed by the people who currently own and govern it, and proposes only that some larger fraction of the proceeds be transferred, on terms set by the existing owners, to the population the industry is in the process of dispossessing.

This is the liberal response. It is the bourgeois response. It is, structurally, the response that the AI industry would want, because it preserves the industry’s ownership of the productive apparatus while neutralizing the political pressure that would otherwise emerge from mass dispossession. It is the response that Chris Lehane, OpenAI’s lobbyist, is pushing for, the reason he is pushing for it is that it is the response that allows OpenAI to continue to exist as OpenAI. It is the response that David Shor is recommending to his Democratic clients, because it is the response that polls well and that the AI industry is willing to fund. And it is the response that, if implemented, will produce a configuration that can only be described, structurally, as feudal.

I want to be specific about this. The configuration in which a small ownership class controls the means of production, a state apparatus extracts a portion of the surplus and redistributes it as welfare provision, and a dispossessed population lives at subsistence on the redistributed surplus, is not capitalism. Capitalism, in its classical form, is a system in which the dispossessed sell their labor power to the owners and receive wages in exchange. The configuration being proposed in the OpenAI white paper, taken to its logical end, eliminates the wage relation entirely. The dispossessed population does not sell labor power, because there is no labor power to sell. It receives a share of the rents extracted by the ownership class through the mechanisms the ownership class has agreed to. This is, in the technical political-economic sense, a feudal relation. The dispossessed are tenants on the land of the owners, receiving subsistence in exchange for political loyalty, with no productive role in the economy that supports them.

The political consequence of a feudal relation is well documented in the historical record. The dispossessed do not, under feudal conditions, develop political agency, because political agency historically arises from the bargaining position that productive labor confers. They develop, instead, a politics of dependency, in which their relationship to the state apparatus that funds their subsistence becomes the central axis of their political life. The ownership class, having no need for the dispossessed as workers or as customers, comes to regard the welfare provision as a charitable concession that can be withdrawn at will. The political stability of the configuration depends entirely on the willingness of the ownership class to continue funding the welfare. When that willingness wavers, which it will, because rentier classes always come to resent the populations they support, the configuration becomes openly authoritarian. The pieces are already in place: the surveillance infrastructure, the militarized policing, the legal apparatus of preventive detention.

This is what universal basic income, on its own, produces. It is what a public wealth fund without democratic governance produces. It is what a jobs guarantee that does not transfer ownership of the productive apparatus produces. Each of these proposals, taken in isolation, becomes the welfare component of a feudal political economy whose ownership structure is preserved. The radical-sounding rhetoric, Lehane warning about “fascism and communism,” Shor running ads against the AI billionaires, is the political theater that accompanies the consolidation of the ownership class’s position. It distracts from the structural question by performing the redistributive question. The redistributive question, in this configuration, is downstream of the structural question, and cannot be resolved without reference to it.

What the AI industry is funding, then, is a political response calibrated to the industry’s own continued existence. The redistributive measures are real, in the sense that they would produce real material improvements for the dispossessed. They are also, structurally, the price the industry is willing to pay to retain ownership of the productive apparatus that is dispossessing them. The price is being calibrated, now, by the political consultants in the warehouse, the policy economists in the Anthropic Institute, and the lobbyists in the OpenAI Washington office. They are arriving at a number, call it twenty percent of corporate profits, call it a fifteen percent capital gains tax, call it a public wealth fund of three percent of company equity, that allows the industry to continue to exist while neutralizing the political pressure that would otherwise compel a more substantial restructuring.

This is the deal that is being assembled. It will be presented, in the 2028 Democratic primary, as the radical option, contrasted favorably with the Republican refusal to address AI displacement at all. It will be supported by figures on the moderate left who recognize that something must be done and who do not, themselves, have an alternative framework. It will be opposed, from the right, as confiscatory and excessive. The actual debate, in the mainstream press, will be about whether the redistributive measures are sufficient to address the crisis or whether they are too aggressive and will impede AI investment. The debate will not be about whether the underlying ownership structure should be altered, because the existence of that question requires a political vocabulary that has been substantially absent from American public discourse since the 1940s and that Sun’s reporting confirms is still not being articulated, in 2026, by the AI industry’s own internal critics.

The political vocabulary that names the structural question, however, exists. It has been assembled, over the past decade, in the work of the contemporary democratic-socialist movement. It was articulated by Bernie Sanders in 2016 and 2020. It was extended by the Democratic Socialists of America’s organizing in the late 2010s and 2020s. It was given its first major American electoral expression by the New York mayoral campaign of Zohran Mamdani in November 2025. And the test of whether it can rise, in the next eighteen months, to the level required by the AI moment will determine the political shape of the next thirty years.

We need to talk about Mamdani.

The Mamdani Inflection

In November 2025, the city of New York elected Zohran Mamdani as mayor, against the combined opposition of the city’s real estate industry, its hedge fund sector, its mainstream Democratic establishment, the leading newspapers, the centrist political action committees, and the considerable financial resources that the Democratic Party’s donor class poured into defeating him. He won anyway. He won because his program of frozen rents, free buses, public childcare, public groceries, a millionaires’ tax to fund all of it, was wildly popular among the population of the city that had been priced out, indebted, and exhausted by the configuration of capital that had taken hold during the previous twenty-five years. He won because that population, which the Democratic Party had assumed could be safely ignored as long as the donor class was kept happy, turned out in numbers that the donor class had not anticipated.

The Mamdani victory is significant, in the analysis I have been developing, for a specific reason that has not yet, to my knowledge, been clearly articulated. Mamdani’s program is not the AI agenda. Mamdani has barely talked about AI. His program is calibrated to the specific conditions of New York City: a housing crisis, a transit crisis, a childcare crisis, a wage crisis, that have been visible for fifteen years. But the form of his program is the form that an AI-era democratic-socialist program would have to take. The form is not redistribution of the surplus. The form is municipal, and ultimately collective, ownership of the conditions of social reproduction. The frozen rents are a partial expropriation of landlord rents. The free buses are a decommodification of urban mobility. The public childcare is a removal of childrearing from the wage market. The public groceries, in their pilot form, are a tentative gesture at the decommodification of food itself. Each of these is a micro-act of what, scaled to the AI moment, would constitute the only structurally adequate response to mass technological dispossession: the transfer of the conditions of life from market provision to democratic provision.

This is the form. The content is what we have to work out, and what Mamdani himself has not done because his political brief has been local rather than national and because the AI moment has not yet, in his city, become the central axis of political conflict. But the content can be specified. We know what the productive apparatus of the AI economy looks like. We know what its critical inputs are. We know who currently owns it. The content of an AI-era democratic-socialist program would consist of the public, democratic ownership of those inputs, that apparatus, and that productive capacity.

Let me sketch what this would actually look like, because no one in the current mainstream discourse has the courage to articulate it, and because the absence of the articulation is itself one of the reasons the discourse has been captured by the OpenAI white paper.

The first element is ownership of the data. The training data on which frontier models are built is, almost without exception, data that was produced by the population at large, the texts written, the images created, the conversations had, the records kept by hundreds of millions of people whose work and lives have been digitized over the past three decades. This data was, in essentially every case, taken without compensation. The legal frameworks that would have required compensation, copyright, trade secret, privacy law, were systematically circumvented through the doctrine of “fair use” in scraping, the bilateral agreements with platform owners who themselves had taken the data from their users, and the regulatory permissiveness that has characterized the period. The AI economy is built on a primitive accumulation of data that is, in scale and character, comparable to the primitive accumulation of land that Marx identified as the precondition for industrial capitalism.

A democratic-socialist response to this would begin with a recognition that the training data is, in its origins, a commons. The natural framework for governing a commons is collective ownership and democratic management. This points toward a public data trust, with ownership held in common by the population whose data it consists of, with governance through democratic mechanisms, and with licensing arrangements that direct a substantial portion of the value generated by AI models trained on the data back to the population that produced it. This is not utopian. The technical mechanisms exist; the political mechanisms exist; what is missing is only the political will to assemble them. A public data trust of this kind would, by itself, transform the economic structure of the AI industry, because the cost of training data, currently zero, because of the primitive accumulation, would become a real cost, paid into a fund whose proceeds are democratically distributed.

The second element is ownership of the computer infrastructure. The data centers being built across the United States, at a capital intensity that, as Sun notes, accounts for nearly forty percent of GDP growth, are the productive apparatus of the AI economy. They are also, increasingly, being built with substantial public subsidies, tax breaks, energy concessions, water allocations, infrastructure investments, that the localities granting them have rarely calculated against the long-term return. A democratic-socialist response would recognize that public subsidies of this scale should produce public ownership stakes of a comparable scale. Where the federal government, state governments, or local governments have funded data center construction, they should hold equity in proportion to the funding, with rights of governance attached to the equity, and with a substantial portion of the productive output reserved for public-interest applications determined democratically rather than commercially. This is what public investment in productive infrastructure has always meant in social-democratic political economies; it has been abandoned in the United States only because forty years of neoliberal restructuring have hollowed out the public capacity to demand it.

The third element is ownership of the model weights themselves. Frontier AI models are, at present, treated as the proprietary intellectual property of the firms that train them, even though the training process is funded by venture capital that is itself, in significant measure, drawn from public pension funds, sovereign wealth funds, and tax-advantaged investment vehicles whose privileges are public concessions. There is a strong case that frontier models trained at scales requiring billions of dollars of capital and gigawatts of public-grid electricity should be regarded as utilities, with the model weights themselves treated as public infrastructure on the open release model. The technical precedent exists: Meta has released the Llama series of models under terms approximating this arrangement, and the major Chinese laboratories have released several of their frontier models openly. The reason this has not become the American norm is not technical. It is that the American AI industry has structured itself around the proprietary model, and the proprietary model is what produces the rents that fund the venture capital cycle. A democratic-socialist policy would treat the proprietary model as an artifact of the present ownership arrangement, not as a technical necessity, and would condition continued public support, including public access to the electricity grid, the labor market, and the regulatory environment, on the public release of model weights at the frontier.

The fourth element is democratic governance of deployment. The decision of when, how, and at what scale to deploy automation technology into specific labor markets is, at present, made entirely by the firms that own the technology, in response to competitive pressures that none of the affected workers have any input into. A democratic-socialist response would establish, at every level, federal, state, sectoral, firm-level, democratic mechanisms through which workers have meaningful input into automation decisions. This is the German codetermination model, scaled to the AI moment and extended into the deployment decisions that are now the central determinant of labor market outcomes. It is also the historical demand of the American labor movement, abandoned in the 1970s and 1980s under the pressure of Reaganite restructuring, and now urgently needed in a configuration where the deployment decisions made over the next five years will permanently determine the labor share of income.

The fifth element is the social wage. The portion of the productive output that is not retained by the workers who produce it, and that is not directly distributed to consumers through market mechanisms, must be returned to the population through a social wage that decommodifies the conditions of life. This is where the public wealth fund of the OpenAI white paper meets its democratic-socialist transformation. The fund itself is acceptable; what is unacceptable is that it would distribute dividends to a population that remains otherwise dispossessed, leaving the rentier configuration intact. The democratic-socialist version would direct the proceeds of the fund toward decommodification: public housing at scale, free public transit nationally, public childcare and eldercare, public health care, public higher education, public broadband, public food provision in the sectors where market provision has failed. This is the Mamdani program, scaled to the national level and funded by the rents of the AI economy. It is also, in its structural form, the form of the welfare state that European social democracy built between 1945 and 1975 before it was systematically dismantled.

The sixth element, and this is the element that is most foreign to current American political discourse, but that is the element on which everything else turns, is the recognition that the productive forces being assembled by the AI industry are productive forces of a fundamentally social character, and that they cannot be coherently governed under the present property relations. The argument is not that AI should be nationalized in the form that nineteenth-century socialism imagined nationalization. The argument is that the present property relations, proprietary models, private data centers, billion-dollar valuations capitalized on rents extracted from socially produced data, are incoherent with the actual character of what has been built. The data is collective. The training is collective in its inputs and increasingly collective in its labor. The deployment affects the entire labor market. The infrastructure depends on the public energy grid. The valuation depends on legal arrangements granted by public authority. The argument is that these productive forces, having been built on collective foundations, must be governed through collective democratic mechanisms, and that the present private-ownership form is a transitional arrangement that the productive forces themselves are, in real time, rendering obsolete.

This is the argument that Marx made in 1858. It is the argument that contemporary Marxist analysis extends. It is the argument that, in 2026, has become not a theoretical position but an empirical observation. Sun’s article, read carefully, is the documentation of an industry that is in the process of confirming that the argument is correct. Amodei is correct that democracy is premised on labor having leverage. Altman is correct that AI will shift power from labor to capital. The Mechanize founders are correct that the technology is going to “fully automate the economy.” Each of them is articulating, in his own register, a recognition that the social form of capital has reached its terminus. They have no political vocabulary for what should follow, because their class position gives them no access to the vocabulary that has historically described what should follow. They are forced into the vocabulary of redistribution within the existing property relations, because they cannot, while remaining themselves, articulate the vocabulary of altering the property relations themselves.

That vocabulary exists. It has existed for one hundred and seventy years. It has been articulated, in fragments, in every major American political movement that has pushed against the prevailing capital-labor configuration. It has been kept alive, in the United States, by the labor movement, by the civil rights movement at its most ambitious, by the Sanders campaigns, by the DSA organizing, and now by Mamdani. The task of the next eighteen months is to bring that vocabulary into contact with the AI moment, to articulate the program with the specificity I have just attempted, and to insist on it in the public discourse with sufficient clarity that the OpenAI white paper cannot occupy the position of “radical horizon” that it currently occupies by default.

The Choice

There are, in the present moment, three positions on the AI question that are politically coherent. Sun’s article presents one of them, lets a second one show through the cracks of its sources’ nervousness, and does not articulate the third at all. Let me describe them.

The first position is the position of capital, articulated by the AI industry’s executives, refined by the lobbyists, polished by the political consultants, and presented in the OpenAI white paper. This position accepts the technological transformation, accepts the dispossession, accepts the labor-share collapse, and proposes to manage the political consequences through redistributive measures calibrated to preserve the ownership structure. The thirty-two-hour workweek, the public wealth fund holding minority equity, the corporate tax increase, the federal jobs guarantee, the data center subsidies that build out the infrastructure faster, each of these is a component of a package whose function is to absorb the political pressure produced by mass dispossession without altering the fundamental ownership relation that produces it. The position is, in its structural form, the construction of a welfare-feudal regime. The dispossessed receive subsistence; the owners retain the productive apparatus; the political conflict is managed through the redistributive parameters, which can be tightened or loosened as the relative bargaining power of the two classes shifts. This is the position that will dominate the AI industry’s political messaging through the 2028 cycle. It is the position that the Anthropic Institute, and the OpenAI policy team, and the Google DeepMind societal impact group are currently producing the discourse around. It is the position that David Shor was selling to the warehouse audience.

The second position is the position of right populism, gestured at by figures like Alex Karp in his Teamsters panel quote about the country “blowing up,” articulated more clearly by various MAGA-aligned commentators, and increasingly visible in the data center bans, the chatbot regulations, and the protectionist measures Sun mentions. This position rejects the technological transformation, or attempts to slow it down, through nationalist and protectionist mechanisms, restrictions on data center construction, tariffs on AI services, immigration restrictions to constrain the AI workforce, regulatory delay tactics. The position has the virtue of being honest about the magnitude of the disruption and the inadequacy of the liberal response. It has the fatal weakness of being structurally incoherent, because the technology is global, the firms are mobile, the capital is internationally deployable, and the only force that could enforce a national prohibition on AI deployment is the same state that has been captured by the AI industry’s lobbying apparatus. The right-populist response can produce dispossessed nationalism, can produce the fascism that Lehane warned about, can produce political violence, but it cannot produce a coherent alternative configuration of the political economy, because it does not address the property relations and it does not have a constituency among the owners of the productive apparatus.

The third position is the democratic-socialist position I have been sketching. It accepts the technological transformation, recognizes that it cannot be reversed and probably should not be, and proposes to alter the property relations that govern the deployment. It transfers ownership of the data, the compute, the models, and the deployment decisions, in democratic increments, from the present ownership class to the population at large. It uses the productive surplus generated by the technology to fund the decommodification of the conditions of life, on the Mamdani model scaled to the national economy. It establishes democratic governance over the labor market consequences of automation through codetermination and sectoral bargaining. It treats the AI economy as a public utility in the technical sense, a productive infrastructure of fundamentally social character that must be governed through democratic mechanisms because no other governance form is adequate to its scale and consequences. It is the only one of the three positions that addresses the structural question rather than the redistributive question. It is the only one that is structurally coherent. It is the only one that has, as its political subject, the population that is being dispossessed, rather than the class that is dispossessing them or the nationalist faction that is reacting against the dispossession.

The choice between these three positions is the choice that will define American politics over the next decade. Sun’s article documents the consolidation of the first position inside the AI industry’s own legitimation apparatus. The second position is consolidating in the right-populist movements that the article gestures at without examining. The third position has a beachhead in New York City, has a base in the Sanders-AOC formation in Congress, has an organizing infrastructure in DSA and the labor movement, and has, for the first time in a generation, a young population that has been trained, by the failure of the liberal redistributive consensus to deliver them economic security, to be receptive to a structural alternative.

The question is whether the third position can be articulated, with the specificity I have attempted, with sufficient clarity and force to compete with the first position for the political attention of the dispossessed. The first position has the funding, the political consultants, the access to the Democratic donor class, the access to the major media outlets. The third position has the mathematics. The first position is offering to transfer twenty percent of the surplus to the dispossessed in exchange for accepting permanent dispossession. The third position is offering to transfer ownership of the productive apparatus to the population that built it. The first position protects the ownership class from political consequences. The third position resolves the structural crisis by altering the property relations that produced it. There is no honest reading of the political economy that does not conclude that the third position is the only one that can do what it claims to do.

The reason it is not currently articulated, in the form I have just articulated it, is not that the analysis is wrong. The analysis is straightforwardly correct, and Sun’s article is, among other things, evidence for its correctness. The reason it is not currently articulated is that the institutions that would normally articulate it have been substantially captured by the first position. The Democratic Party, the academic political-economy establishment, the major newspapers, the policy think tanks, each of these is structurally constituted to articulate the first position and to filter out the third. The third position survives because the conditions on the ground are producing a generation of young Americans who are reaching it on their own, through their lived experience of the failure of the first position to deliver them anything resembling the security their parents had. The Mamdani campaign was the first major electoral expression of this generation reaching its analytical conclusions. The 2028 Democratic primary will be the next.

What Is to Be Done

I want to close where Sun’s piece closes, with the warehouse, the firebombings, the panel quote about the country “blowing up.” Karp is correct that none of the present configuration’s beneficiaries make any money in the world that he is gesturing toward, the world of mass political violence directed at the symbols of capital. He is also correct, by implication, that the redistributive measures the AI industry is willing to fund are calibrated to forestall that world, not to address its underlying causes. The redistribution is a tactic. The structural question remains.

The structural question can be addressed in two ways. It can be addressed by a democratic political movement that organizes the dispossessed population around the specific program I have sketched, that wins elections at the municipal, state, and federal levels, and that uses electoral power to alter the property relations governing the AI economy. This is the path that Mamdani’s victory in November 2025 indicates is possible. It is the path that the Sanders campaigns indicated, in 2016 and 2020, was possible at the national level. It is the path that, if pursued with sufficient discipline and clarity over the next eighteen months, could produce a 2028 Democratic primary in which the structural question is on the table for the first time in the lifetimes of the people pursuing it.

Or it can be addressed by the political violence that the warehouse audience is already, in their nervous gallows humor, beginning to anticipate. The firebombing of Sam Altman’s house was a warning. The data center bans being passed at the local level are a warning. The political polarization that AI displacement will produce, if no electoral path emerges to absorb it, is going to make the 2016-2024 period look like a prelude. The structural question will be addressed; it cannot remain unaddressed indefinitely; the only question is whether it is addressed through the democratic mechanisms or through the failure of the democratic mechanisms.

The argument I have been making is that the democratic mechanisms are still available, but only if the analysis underlying them is articulated with the specificity that the moment requires. The specificity is what is currently missing. Sun’s article documents what is happening, but the analytical vocabulary it deploys cannot name the structure. The OpenAI white paper proposes a redistributive package that preserves the ownership structure. The Anthropic Institute hosts the discourse of concern that legitimates the deployment. The David Shor presentation packages the political messaging that absorbs the discontent into the existing Democratic framework. None of these institutions is going to produce, on its own, the analytical clarity that the moment requires. That clarity has to come from outside the apparatus that has been assembled to manage the discontent.

It has to come from the political tradition of democratic socialism, broadly defined, in its American expressions from the labor movement through the civil rights movement through the Sanders campaigns through DSA through Mamdani, that has retained, against forty years of neoliberal restructuring, the analytical framework that names structural questions as structural and that proposes structural answers. It has to come from the writers, the organizers, the candidates, the academics, and the workers who have inherited that tradition and who have access to the venues that the AI industry has not yet captured. It has to come, in other words, from us. From the people reading this essay, from the people writing the essays I am responding to, from the people running for office in the 2026 midterms and the 2028 primaries, from the people teaching the courses that the next generation of American socialists will read.

The work is straightforward. We describe what is happening, in the language adequate to it. We articulate the program, with the specificity I have attempted to provide. We organize the political constituency: the dispossessed, the soon-to-be-dispossessed, the young, the indebted, the housing-cost-burdened, the wage-stagnated, the white-collar workers who are about to learn that their education was not the protection they were told it was. We contest the elections at every level, as Mamdani did, with programs that the donor class will not fund and that the electorate, increasingly, is prepared to vote for. We refuse the redistributive package being assembled in the Anthropic Institute and the OpenAI policy office, while accepting any individual element of it that can be wrested from the larger structure and deployed in service of the actual program. We treat the next eighteen months as the period in which the analytical clarity has to be established, before the first position consolidates its control of the discourse going into 2028.

This is what the moment requires. It is what the historical record indicates is possible, when the analytical clarity, the political organization, and the electoral opportunity converge. It is what Mamdani’s New York demonstrated, on a smaller scale, in November 2025. It is what the Sanders campaigns demonstrated, on a larger scale and against more substantial opposition, in 2016 and 2020. It is what the AI moment, in the magnitude of its dispossession and the rapidity of its onset, has now made not merely possible but necessary.

The alternative is the one that Sun’s article documents, in the warehouse fundraiser, in the disco-ball-illuminated political consultancy, in the careful papers of the Anthropic Institute, in the OpenAI white paper. The alternative is the welfare-feudal regime, the technological dispossession with redistributive consolation, the labor share collapse with public wealth fund palliative. The alternative is what the AI industry is funding, on the assumption that they can buy the political response before any other response can take shape.

They will succeed, unless they are stopped.

We work for you. They work for the bots. The line is not just a campaign slogan. It is an analytical claim. It identifies the structural division that the AI moment is producing. The class that owns the productive apparatus is, in real time, organizing the political response to its own production of mass dispossession. The class that is being dispossessed has, in real time, the opportunity to organize an alternative response. The next eighteen months are the period in which it will be determined whether that opportunity is taken.

The disco ball is still spinning. The La Croix is still cold. The presentation is still running. The aristocracy is still throwing the fundraiser to soften its own descent. They are not going to stop. Nothing in their position, their training, or their material interests would lead them to stop. They are going to continue, at the warehouse fundraisers and the white papers and the institutes and the policy offices, until the political response to their continued operation becomes irresistible.

The political response has to be assembled outside the warehouse. The vocabulary has to come from the political tradition that the warehouse cannot use. The program has to be articulated by the people who have been excluded from the warehouse, on terms the warehouse cannot accommodate.

That is the work.

It is, as Marx wrote about a different but structurally analogous moment, a long path. The hour at which the path ends is not yet visible. The hour at which the path must be taken, however, is visible. It is now. It is the eighteen months between this essay and the first nominating contests of the 2028 cycle. It is the period in which the analytical clarity will be established, or it will not be, and the political consequences of that establishment, or its absence, will play out across the rest of our lives.

They work for the bots.

We have to work for each other, or no one works for us at all.

Bibliography

Acemoglu, Daron, and Pascual Restrepo. “Robots and Jobs: Evidence from US Labor Markets.” Journal of Political Economy128, no. 6 (2020): 2188–2244.

Altman, Sam. “Moore’s Law for Everything.” Personal blog, March 16, 2021. https://moores.samaltman.com/.

Amodei, Dario. “The Adolescence of Technology.” Personal essay, January 2026. https://www.darioamodei.com/essay/the-adolescence-of-technology.

Amodei, Dario. “Machines of Loving Grace: How AI Could Transform the World for the Better.” Personal essay, October 2024. https://darioamodei.com/machines-of-loving-grace.

Autor, David H. “Why Are There Still So Many Jobs? The History and Future of Workplace Automation.” Journal of Economic Perspectives 29, no. 3 (2015): 3–30.

Berardi, Franco “Bifo.” The Soul at Work: From Alienation to Autonomy. Translated by Francesca Cadel and Giuseppina Mecchia. Los Angeles: Semiotext(e), 2009.

Berardi, Franco “Bifo.” Precarious Rhapsody: Semiocapitalism and the Pathologies of the Post-Alpha Generation. London: Minor Compositions, 2009.

Block, Sharon, and Benjamin Sachs. Clean Slate for Worker Power: Building a Just Economy and Democracy. Cambridge, MA: Labor and Worklife Program, Harvard Law School, 2020.

Bostrom, Nick. Superintelligence: Paths, Dangers, Strategies. Oxford: Oxford University Press, 2014.

Bostrom, Nick. “Apology for an Old Email.” Personal website, January 2023. https://nickbostrom.com/oldemail.pdf.

Brynjolfsson, Erik, and Andrew McAfee. The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies. New York: W. W. Norton, 2014.

Davenport, Charles B. Heredity in Relation to Eugenics. New York: Henry Holt, 1911.

Deleuze, Gilles. “Postscript on the Societies of Control.” October 59 (Winter 1992): 3–7.

Federal Reserve Bank of St. Louis. “Tracking AI’s Contribution to GDP Growth.” On the Economy blog, January 2026. https://www.stlouisfed.org/on-the-economy/2026/jan/tracking-ai-contribution-gdp-growth.

Fisher, Mark. Capitalist Realism: Is There No Alternative? Winchester, UK: Zero Books, 2009.

Fisher, Mark. K-Punk: The Collected and Unpublished Writings of Mark Fisher (2004–2016). Edited by Darren Ambrose. London: Repeater Books, 2018. Includes the unfinished manuscript fragments of “Acid Communism.”

Frey, Carl Benedikt. The Technology Trap: Capital, Labor, and Power in the Age of Automation. Princeton, NJ: Princeton University Press, 2019.

Galton, Francis. Hereditary Genius: An Inquiry into Its Laws and Consequences. London: Macmillan, 1869.

Han, Byung-Chul. The Burnout Society. Translated by Erik Butler. Stanford, CA: Stanford University Press, 2015.

Han, Byung-Chul. Psychopolitics: Neoliberalism and New Technologies of Power. Translated by Erik Butler. London: Verso, 2017.

Harvey, David. The New Imperialism. Oxford: Oxford University Press, 2003.

Herrnstein, Richard J., and Charles Murray. The Bell Curve: Intelligence and Class Structure in American Life. New York: Free Press, 1994.

Jäger, Simon, Benjamin Schoefer, and Jörg Heining. “Labor in the Boardroom.” Quarterly Journal of Economics 136, no. 2 (2021): 669–725.

Korinek, Anton. “Economic Policy Challenges for the Age of AI.” NBER Working Paper No. 32980. Cambridge, MA: National Bureau of Economic Research, September 2024.

Lewis-Kraus, Gideon. “The Reluctant Prophet of Effective Altruism.” The New Yorker, August 8, 2022.

Lombardo, Paul A. Three Generations, No Imbeciles: Eugenics, the Supreme Court, and Buck v. Bell. Baltimore: Johns Hopkins University Press, 2008.

Mamdani, Zohran. A New York for the Many: Campaign Platform. New York: Mamdani for New York Committee, 2025.

Marx, Karl. Capital: A Critique of Political Economy, Volume I. Translated by Ben Fowkes. London: Penguin Classics, 1990. Originally published 1867.

Marx, Karl. Grundrisse: Foundations of the Critique of Political Economy. Translated by Martin Nicolaus. London: Penguin Classics, 1993. Originally published 1858.

Marx, Karl, and Friedrich Engels. The Communist Manifesto. Translated by Samuel Moore. London: Penguin Classics, 2002. Originally published 1848.

OpenAI. “Industrial Policy for the Intelligence Age.” White paper, April 2026.

Polanyi, Karl. The Great Transformation: The Political and Economic Origins of Our Time. 2nd ed. Boston: Beacon Press, 2001. Originally published 1944.

Potter, Del. “The Generation That Forgot Itself: How the Sixties Generation Became the Base of American Fascist Reaction.” Del’s Substack, April 18, 2026.

Potter, Del. “The Conference and the Voucher: On Inclusion as Ritual and the Operation of Abstraction.” Part Seven of an Essay Series on the Enclosure of Psychedelics. Del’s Substack, April 29, 2026.

Sanders, Bernie. It’s OK to Be Angry About Capitalism. New York: Crown, 2023.

Stern, Alexandra Minna. Eugenic Nation: Faults and Frontiers of Better Breeding in Modern America. 2nd ed. Berkeley: University of California Press, 2016.

Sun, Jasmine. “Silicon Valley Is Bracing for a Permanent Underclass.” New York Times, April 30, 2026.

Sunkara, Bhaskar. The Socialist Manifesto: The Case for Radical Politics in an Era of Extreme Inequality. New York: Basic Books, 2019.

Torres, Émile P. Human Extinction: A History of the Science and Ethics of Annihilation. New York: Routledge, 2023.

On the evening of March 30, 2025, in the Mission District of San Francisco, the Brava Theater for Women hosted the closing panel of the Chacruna Institute’s Psychedelic Culture conference. Rick Doblin, founder of the Multidisciplinary Association for Psychedelic Studies, sat on the stage with his successors at the organization, Ismail Lourido Ali and Betty Aldworth. Behind them, Bia Labate, Chacruna’s co-founder, had opened the conference two days earlier with an instruction the field has been quoting back to itself ever since: stay in the tension, between celebration and critique, between what this movement has accomplished and what it still has not done. As the panel began its closing remarks, a small group of attendees stood up. They challenged Doblin and the MAPS leadership about the organization’s research and training work in Israel, specifically MAPS Israel’s program training therapists who had treated October 7 survivors and IDF veterans, conducted in the months following what the United Nations Commission of Inquiry would in September 2025 conclude was a genocide against Palestinians in Gaza. The protesters spoke. The panelists responded. The moderator suggested continuing the conversation after the closing ceremony. The conversation did continue, on the sidewalk outside the Brava Center, for nearly an hour. Then Doblin and members of the MAPS team went to a nearby restaurant. The protesters, by chance or design, walked into the same restaurant. More dialogue ensued. At the end of it, according to multiple eyewitnesses, one of the protesters expressed admiration for Doblin’s decades of work in the psychedelic field and asked him for his autograph. Doblin obliged.

Twenty-five days later, on the afternoon of April 24, 2026, the Food and Drug Administration announced the issuance of three Commissioner’s National Priority Vouchers for psychedelic drug programs. The recipients were Compass Pathways, the British-incorporated firm in which Christian Angermayer’s Atai Life Sciences holds a substantial stake, for its proprietary psilocybin formulation COMP360 in treatment-resistant depression. The Usona Institute, a Wisconsin nonprofit research organization with no apparent commercial team and no current launch infrastructure, for its psilocybin program in major depressive disorder. And Transcend Therapeutics, then in the final stages of a $1.225 billion acquisition by Otsuka Pharmaceutical, for its methylone candidate TSND-201 in post-traumatic stress disorder. The voucher compresses the FDA review timeline from ten or twelve months to one or two. Together, the three vouchers establish which fraction of the field will operate the regulatory pathway the April 18 executive order opened.

Conspicuously absent from the list was Resilient Pharmaceuticals, the entity formerly called Lykos Therapeutics, the corporate descendant of MAPS itself, the institution Rick Doblin had spent thirty-eight years building toward exactly this moment. Lykos had received its Complete Response Letter in August 2024. The letter became public on September 4, 2025, in a transparency release by the agency. Among its disclosures was an instruction that had circulated inside the trial protocols: that participating sites were not to record certain euphoria-like effects as adverse events, since these were considered therapeutically expected rather than reportable safety signals. Allegations of therapist-patient sexual misconduct at one trial site were flagged for independent audit. Lykos’s response was to dismiss seventy-five percent of its staff, rebrand as Resilient, and prepare a resubmission that the field has been waiting for ever since. It did not arrive in time for the priority voucher cycle. The methylone voucher went to Otsuka instead.

These two events, the closing panel at Brava Theater, the voucher distribution at the FDA, are not unrelated incidents in a field whose discourse and whose economy operate independently. They are the same operation conducted twice. The conference and the voucher are paired apparatuses. The first produces the rhetoric of acknowledgment that the second requires in order to function without legitimation crisis. The second distributes the property the first has rendered abstractable by separating it from its source. The neo-colonial project at the heart of the contemporary psychedelic enterprise is to abstract the chemical from the culture; the conference is where the abstraction is performed as ritual, the voucher is where it is performed as policy. To understand how the field arrived at the configuration this essay’s predecessor described: three priority vouchers, a Texas ibogaine consortium, a Compass-Otsuka-Usona axis, a founding institution in collapse, requires holding both apparatuses in view at once.

The Operation Named

Let me state the proposition the rest of this essay will defend.

The neo-colonial project is to abstract the chemical from the culture. This is not a decorative gloss on a more fundamental economic operation. It is the economic operation. The chemical separated from the culture is the only form in which the medicine can be patented, scaled, capitalized, assigned to a particular fraction of capital, and moved through a regulatory pathway that costs upwards of fifty million dollars to traverse. The culture, by contrast, cannot be enclosed. It belongs to people who would have to be acknowledged, paid, and consulted; whose protocols would have to be honored; whose presence would generate obligations the regulatory pathway cannot accommodate. The project therefore requires the abstraction. Abstraction is the founding move on which everything downstream depends.

This formulation differs in emphasis from the argument my Part Six essay developed. That essay treated abstraction as an effect, the pharmacopoeia ended up without a people because of how capital structures pharmaceutical medicine. The formulation here treats abstraction as the project itself, the operation the apparatus exists to perform. The chemical-without-culture is the colonial product. Distributing it more equitably, to more demographics, at more accessible prices, is distributing colonial extraction more equitably. It does not address the founding move. It accepts the abstraction and bargains over its terms.

What follows in the rest of this essay is an examination of the two main institutional sites at which the abstraction is currently being performed: the conference economy and the voucher economy, and an argument that they are mutually constitutive. Without the conference, the voucher would face a legitimation crisis. Without the voucher, the conference would face a relevance crisis. They need each other.

The Voucher Economy and Its Logic

The April 24 voucher distribution is the cleanest available reading of which faction of capital the apparatus has chosen to capitalize. The signals are unambiguous. Compass Pathways is publicly listed in London, was substantially financed by Peter Thiel and Christian Angermayer’s Atai Life Sciences, and is structured as a commercial pharmaceutical operation in the conventional pharma mold. Its COMP360 program for treatment-resistant depression is a synthetic psilocybin formulation administered in a credentialed-facilitator clinical setting; the company holds patents on aspects of the formulation, the synthesis, and the delivery protocol. Usona Institute is the outlier among the three in being a nonprofit, but an open-pathway nonprofit aligned with the Compass approach to psilocybin standardization rather than with the MAPS approach to extended psychotherapeutic protocol. Transcend Therapeutics, the third voucher recipient, is the most diagnostic of the trio. Its methylone candidate is an MDMA analog being developed for the same indication MAPS spent four decades pursuing with MDMA itself. Otsuka Pharmaceutical is acquiring it for $1.225 billion. The compound is entering Phase 3, making it the least mature of the three and the most surprising of the voucher selections.

That last detail is the one worth holding. The methylone voucher could have gone to Resilient. Resilient holds two completed Phase 3 studies. Its CRL had been issued nineteen months earlier and made public seven months earlier; its resubmission was understood to be in preparation. The field expected the voucher to go there. Industry analysts expected it. The administration was reported to have signaled it. Instead, the voucher went to a methylone program that did not exist five years ago, will be operated by a Japanese pharmaceutical company that has built its U.S. position primarily through the antipsychotic Abilify and the ADHD treatment Rexulti, and will commercialize a compound chemically related to MDMA without inheriting any of MAPS’ protocol architecture, training infrastructure, or twenty-year ethnographic relationship to the molecule’s therapeutic application.

The signal is unambiguous in two directions. First, the apparatus has chosen the pharma model over the founding-institution model. The fraction of capital operating the regulatory pathway is not the fraction the founding generation of psychedelic researchers expected. Compass, Otsuka, and an aligned-nonprofit Usona will run the pathway; Resilient will not, at least not in the inaugural cohort. Second, the apparatus has chosen the non-classical, MDMA-analog, pharma-friendly molecule over the classical compound carried by the field’s founder. Methylone is an MDMA analog with a substantially different abuse-liability profile, fewer associations with the rave-and-recreational history that has dogged MDMA, and a chemistry that fits more comfortably inside conventional pharmaceutical workflows. It is, in every respect, the more abstractable molecule. The apparatus chose it.

The Resilient exclusion also rewards careful reading of the CRL itself. The August 2024 letter, made public in September 2025, raised three categories of concern: durability, blinding, and safety reporting. The safety-reporting concern is the one that matters here. The FDA flagged that participating sites had been instructed not to record certain euphoria-like effects as adverse events, on the theory that euphoria was therapeutically expected and therefore not a safety signal. This is a small detail with a large implication. The field’s founding institution, in pursuing approval for a molecule whose entire therapeutic theory rests on the integration of phenomenology with chemistry, had been instructing its trial sites to abstract the phenomenology out of the safety data. The same operation the broader apparatus performs at the level of the regulatory pathway, the founding institution was performing inside its own protocol. The abstraction is everywhere.

The Conference Economy and Its Function

If the voucher distributes the abstracted molecule as property, the conference distributes the abstraction itself as discourse, and produces, in the process, the rhetoric of acknowledgment that allows the broader apparatus to function without crisis.

To see this clearly, return to the Brava Theater. Labate’s opening remarks framed the two days that followed. Stay in the tension, she said, between celebration and critique. Between what this movement has accomplished and what it still has not done. The line is well-crafted, lyrically pitched, and indispensable to the discursive form the conference operates inside. Its function is to permit indefinite occupancy of an unresolved position. To stay in the tension is not to resolve the tension; it is to continue inhabiting it as a matter of practice. Critique is welcome inside the tension. Action against the structure that produces the tension is not a tension-thing. It would resolve the tension by ending the structure that requires the tension to exist. Staying in the tension, therefore, is what the apparatus prefers its critics to do. The opening keynote of the field’s leading critical-knowledge institution issued, as its central instruction, the methodology the apparatus needed.

This is not a charge against Bia Labate personally, who has done substantial substantive work on indigenous reciprocity, plant medicine policy, and the conditions of the field’s source communities. It is a structural observation about what keynotes do. The keynote is a particular discursive form. Its function is to produce the affective and rhetorical conditions inside which the conference can proceed. The conference, in turn, produces the discursive conditions inside which the apparatus can proceed. The keynote is to the conference what the conference is to the apparatus. Each layer absorbs the critique of the layer above it and converts it into content. Critique becomes panel discussion. Panel discussion becomes published proceedings. Published proceedings become citations in further panels. At no point in the cycle is anything materially required of the apparatus that hosts the conference, that funds the keynote, that publishes the proceedings, or that distributes the priority vouchers six weeks later.

The Brava Theater protest exposed the limit of this form. When the protesters challenged Doblin about MAPS Israel’s training relationships with Israeli healthcare facilities operating with Israeli government support, the discourse-as-ritual broke down. The question was no longer “what is the field’s relationship to indigenous knowledge in the abstract?” It was “what is your training partner doing right now with the protocols you developed?” The first question is a tension to be stayed in. The second question is a material relationship to be answered or refused. The panel could not absorb the second question. Doblin and the MAPS leadership engaged the protesters, the moderator suggested moving the conversation outside, the conversation moved outside, the panel returned to its closing remarks. By the end of the evening Doblin was signing an autograph at a restaurant. The protest had not changed any institutional relationship. It had not interrupted the closing ceremony. It had not produced a single material commitment from MAPS, from Chacruna, or from any organization in the room. What it had produced was content. The Lucid News reporting; the social media circulation of the video; the academic commentary; the opportunity for the field to debate, in panels and podcasts and Substack essays, what the protest meant. All of which is to say: the apparatus absorbed the protest exactly as it was designed to absorb such protests, and emerged with its rhetorical equipment refurbished. Stay in the tensionwas reaffirmed precisely because the tension had been, briefly, threatened with resolution.

The historical record of the conference economy makes this pattern visible across multiple cycles. At Psychedelic Science 2023 in Denver, indigenous protesters interrupted Doblin’s closing keynote with chants and drums. The Native Coalition of Colorado posted, two days later, what may be the most accurate description of the conference economy ever produced inside the field: that organizations including MAPS, Chacruna, New Approach, and Decriminalize Nature had “manufactured consent, which is colonization.” The Indigenous Medicine Conservation Fund delegation, composed of representatives from Oglala Lakota, Sioux-Lakota, Diné, Comanche, Havasupai, Bwiti, Hopi, Tewa, Cheyenne, Arapaho, Navajo, Menominee, Oneida, Mazatec, and Yawanawa traditions, issued a formal declaration calling on the psychedelic boom not to become “another wave of colonialism.” The declaration is dated June 2023. Two years later, at Psychedelic Science 2025, the Plant Medicine Track of that same conference featured an opening panel titled “Setting the Principles for Cultural Appreciation and Respectful Relationship,” moderated by an IMC Fund co-executive director, with Georges Gassita of Blessings of the Forest Gabon, the very organization stewarding the Nagoya-compliant iboga supply chain my Part Six essay tracked, among the panelists. The Bwiti, in other words, were already on the main stage. They had been on the main stage in 2023 (signed onto the IMC declaration), in 2024, in 2025. And during precisely the window in which they were appearing on these stages, the Texas ibogaine consortium was being assembled, the Trump executive order was being drafted, and the FDA voucher process was being designed, none of which acknowledged Gabon, Nagoya, the Bwiti, or any of the institutions whose representatives the conferences had been hosting.

The conference and the voucher are not antagonists. They are complementary. The conference produces the appearance of the field listening to source communities. The voucher distributes the abstracted molecule to the firms that will scale it without those communities. The two operations together produce a configuration in which the apparatus can claim acknowledgment as discourse while denying acknowledgment as policy. This is the operation. This is what the conference is for.

Ngũgĩ’s Cultural Bomb and the Verbs of Abstraction

Ngũgĩ wa Thiong’o, who died in May 2025 at the age of eighty-seven, named this operation in 1986 with a precision the psychedelic field has not yet equaled. In Decolonising the Mind, he described what he called the cultural bomb of imperialism: the weapon that “annihilates a people’s belief in their names, in their languages, in their environment, in their heritage of struggle, in their unity, in their capacities and ultimately in themselves.” The bomb’s effect, Ngũgĩ wrote, is to make the colonized “see their past as one wasteland of non-achievement,” to make them “want to identify with that which is furthest removed from themselves,” to plant “serious doubts about the moral righteousness of struggle.” Its intended results are despair, despondency, and a collective death-wish. And the most diagnostic line in Decolonising the Mind, the one that bears most directly on the conference economy: “It is the final triumph of a system of domination when the dominated start singing its virtues.”

The psychedelic field has reached the configuration Ngũgĩ described. Its main-stage architecture now incorporates source-community representatives who deliver, on the field’s own platforms, the rhetoric the apparatus needs in order to proceed without crisis. This is not because those representatives are insincere, most are doing serious substantive work, but because the form of their inclusion has been calibrated to absorb their critique as content rather than to translate it into structural change. The Bwiti have been at MAPS conferences. The Mazatec have been at MAPS conferences. The Yawanawá have been at MAPS conferences. None of their presence has produced a Nagoya-compliant supply chain inside the U.S. regulatory pathway, a benefit-sharing provision in the Texas ibogaine legislation, or a revenue-share agreement between any voucher recipient and any source community. The apparatus has incorporated the keynote and refused the policy.

What Ngũgĩ’s framework makes visible is that the operation has a specific mechanism. The bomb works through a particular set of verbs. In language, those verbs are standardize, modernize, develop, civilize. In pharmacology, they are translate, characterize, isolate, standardize, identify the active principle. Each verb performs the same operation. Each requires that the source be reduced to raw material in order that proper knowledge can be produced through Western processing.

Translate implies that what the source community has is pre-clinical material awaiting translation into clinical knowledge through Western methodology. The Bwiti have not produced clinical evidence; they have produced data that the West will translate into evidence. The verb assumes the direction of translation runs only one way.

Characterize implies that the molecule’s properties were not known until they were characterized using Western analytical methods. Ibogaine had no properties before the hERG channel was characterized in modern electrophysiology. The two centuries of Bwiti cardiac-screening intuition did not constitute knowledge of those properties; they constituted folk practice that happened to track properties Western methods would later reveal.

Isolate implies that the active principle is the molecule, and that everything else, the nganga, the temple, the ngombi harp, the multi-day ceremony, the cosmological framing of death and rebirth, is non-active material to be isolated from. The verb encodes a metaphysics in which only chemistry is causal. The ethnopharmacological literature, the set-and-setting literature, and the Stanford-Ambio data all contradict this metaphysics. The verb survives anyway.

Standardize implies that variability in the source-community protocol is a defect rather than a feature, and that progress requires replacing the variable protocol with a standardized one that can be deployed at scale. The verb assumes that what cannot be standardized cannot be real medicine. The Bwiti’s calibrated dosing, the nganga adjusting in real time across the eight-to-twelve-hour course of the experience, is therefore not medicine but folk variability. The Knuijver 2022 Radboud study, which standardized the protocol and produced QTc prolongation in every participant and one cardiac event requiring resuscitation, is therefore evidence about ibogaine’s safety profile rather than evidence about what standardization itself does.

These are the verbs of the cultural bomb in pharmacological form. They perform the abstraction. They make the abstraction sound like science rather than like the operation it is. They are, in Ngũgĩ’s exact sense, the language of a system of domination, and they have been so thoroughly absorbed into the field’s discourse that to refuse them is to sound, inside the field, as if one is refusing science itself.

This is the deepest layer of what the apparatus has accomplished. It has captured the verbs. It has made the abstraction the unmarked default. To object to the abstraction now requires marking it, naming it, bringing it into visibility, and the conference economy has been calibrated to absorb that marking as further content. Stay in the tension, indeed.

What the Apparatus Cannot Absorb

There is, however, a class of statement the apparatus cannot absorb. The Brava Theater protest produced a brief instance of it. The PS2023 protest produced another. The IMC Fund Declaration produced a third. What these moments share is that they each posed concrete material questions in place of abstract discursive ones.

A discursive question can be absorbed: what is our relationship to indigenous knowledge? That question can occupy panels, generate proceedings, populate keynotes, and produce no obligation of any kind. A material question cannot be absorbed: who is your training partner and what are they doing right now? That question requires either an answer or a refusal to answer. Both responses are damaging to the apparatus. To answer is to expose the relationship. To refuse to answer is to expose the refusal. The protest at Brava Theater forced exactly this choice for ninety seconds. What the apparatus did with those ninety seconds tells us what its limit looks like. It did not answer the question. It contained the protest, moved it outside, dispersed it across an autograph and a sidewalk conversation, and reabsorbed the energy as content. But for ninety seconds the abstraction failed.

The methodological lesson is that the field’s most important political question is not whether to engage the apparatus on its own terms but whether to pose questions the apparatus cannot metabolize. Three such questions are visible right now:

The first is whether the priority voucher recipients will share revenue with the source communities of the molecules they are commercializing. Compass with the Mazatec. Usona with the Mazatec. Transcend, and Otsuka after acquisition, with whatever cultural archive methylone draws on. None of these arrangements currently exists. None has been proposed. None has been demanded by the field’s main-stage institutions. The question is materially answerable in dollars and percentages. The apparatus cannot accept the answer without restructuring its economics. Posing the question with specificity: what percentage of COMP360 revenue will flow to Mazatec community institutions? is what the apparatus cannot absorb.

The second is whether the Texas ibogaine consortium will operate under Nagoya. The United States is not a Nagoya signatory; Texas can claim no legal obligation. But the source community can be paid voluntarily, and the consortium can structure its supply chain through Terragnosis or a comparable Nagoya-compliant pipeline. The question is materially answerable as a procurement decision. The apparatus has not made it. Demanding that it make it is what the apparatus cannot absorb.

The third is whether the field’s main-stage institutions will refuse to host source-community representatives whose presence is being used to produce legitimation while no policy follows. This is the hardest of the three because it appears, on its surface, to be a refusal of indigenous voices. It is in fact the opposite. To refuse to host the keynote without the policy is to refuse to allow the keynote to be the substitute for the policy. Stay in the tension permits the substitution. Refusing the keynote until the policy arrives breaks the substitution. The Bwiti can speak at PS2025; the apparatus can claim acknowledgment; the voucher can issue six weeks later without the acknowledgment producing any policy. The mechanism by which this works is the mechanism that source-community representatives can refuse to participate in. Some have. Many have not, for understandable reasons including the importance of the platform itself. But the choice exists.

What these three questions share is that they shift the field’s discursive register from cultural abstraction to material claim. Whose practice should inform the protocol is a discursive question. Who is paid is a material one. What is the appropriate relationship to indigenous wisdom is discursive. What is the revenue share is material. The apparatus survives the first kind of question by absorbing it. It does not survive the second.

What This Costs

Return to the patient. The pharmacopoeia my Part Six essay described, the molecules the Texas ibogaine consortium and the April 24 voucher recipients will deliver into clinical settings over the next decade, will be administered in environments stripped of the architectures that make those molecules work. The Lykos CRL revealed the mechanism with unintended clarity. Trial sites instructed not to record euphoria-like effects as adverse events. Trial sites in which therapist-patient sexual misconduct allegations required independent audit. Trial protocols, in other words, that had already been performing the abstraction internally, separating the molecule’s phenomenology from its safety profile, treating the relational architecture of the session as either expected-and-unreportable or as a site of unmonitored harm. The field’s founding institution, in pursuing approval for the first psychedelic medicine, had been gaming the abstraction. The apparatus that has now displaced that institution will perform the same operation at greater scale and with less internal critique.

The Knuijver 2022 study at Radboud, the most rigorous prospective ibogaine safety dataset in the European literature, provides the natural experiment. A medical-monitoring-only protocol, stripped of the integration architecture and ceremonial container the Bwiti developed, produced QTc prolongation in every participant and one cardiac event requiring resuscitation. The Stanford magnesium-ibogaine cohort at Ambio, using a Bwiti-derivative protocol, produced zero serious adverse events across thirty Special Operations veterans. The difference between these two outcome profiles is the difference between the molecule held inside the container and the molecule abstracted from it. The apparatus is preparing to deploy the abstracted version.

The patients who will pass through the Texas pathway, through the Compass clinic networks, through the Otsuka methylone protocol if it reaches market, will receive medicines whose efficacy has been demonstrated in containers, in conditions of clinical operation that will not include the containers. They will be told the medicines worked because their depression scores came down at thirty days, their PTSD symptoms attenuated, their craving patterns interrupted. They will not be told that the scores would have come down further, and stayed down longer, and produced fewer adverse events, in containers the apparatus refused to fund. The cost of the abstraction is paid downstream, in the form of inferior medicine at greater scale, in patients who never knew that the medicine they received was a less effective version of the medicine that was available to be built.

The cost is also paid in the source communities themselves. Gabon’s iboga supply chain is currently under siege from international trafficking that the Gabonese state, weakened by the August 2023 coup and the political instability that followed, cannot effectively police. The Mazatec face an ongoing depletion of their traditional psilocybin habitats due to Western harvesting pressure that the Compass and Usona programs will only intensify once the synthetic-supply-chain economics are validated by approved pharmaceutical product. The peyote crisis in southern Texas, the central concern of the IMC Fund delegation in 2023, has accelerated through the 2024 and 2025 cycles, with peyote prices approaching $650 for two pounds where Native American Church members had once paid $350 for forty pounds. None of this appears in the priority voucher applications. None of it appears in the executive order. None of it appears in the keynotes that produce the rhetoric of acknowledgment. It appears only in the protests the apparatus absorbs as content.

The Choice Ahead of the Field

The conference and the voucher are paired apparatuses. They produce, between them, the contemporary configuration of the psychedelic enterprise: a field in which the rhetoric of indigenous reciprocity is delivered on main stages while the economic flows run in the opposite direction; in which source communities provide the protocols that make the medicines work and receive, in exchange, panel invitations and acknowledgments at conferences whose closing ceremonies they are sometimes permitted to interrupt; in which the founding institution of the modern psychedelic movement has been outflanked by a methylone program owned by a Japanese pharmaceutical company that did not exist in the field five years ago.

The neo-colonial project is to abstract the chemical from the culture. The conference economy performs the abstraction as discourse. The voucher economy performs it as policy. The two operations together constitute the apparatus of capture my essays in this series have been tracking. Naming the project is the precondition for refusing it. Refusing it is the precondition for any honest psychedelic medicine at all.

The field has a choice ahead of it that it has not yet acknowledged. It can continue performing the conference economy alongside the voucher economy and watch the abstraction proceed until the first deaths in supervised clinical settings make the abstraction visible to the FDA, at which point the field’s progressive critics will be invited to participate in the post-mortem, on panels, at conferences, with appropriately respectful keynotes. Or it can name what is happening, refuse the verbs of abstraction, and demand the material relationships the abstraction was designed to obscure. The protests at Brava Theater and PS2023, the Native Coalition statement, the IMC Fund Declaration of June 2023, the Bwiti at PS2025, these are not interruptions of the field’s discourse. They are the field’s remaining honest archive. They name what the keynotes do not. The question is whether the field will let that archive shape its next decade or whether it will continue letting the keynotes shape it.

Ngũgĩ wa Thiong’o died on May 28, 2025, two months after the Brava Theater protest. It is the final triumph of a system of domination when the dominated start singing its virtues. The line was written in 1986 about language and African literature. It applies, with painful exactness, to the configuration the psychedelic field has built for itself in the four decades since. Whether the field will recognize the description in time to refuse it is the political question the next decade will answer. This essay is one attempt to clarify what the question actually is.

The neo-colonial project is to abstract the chemical from the culture. The chemical can be patented. The culture cannot. That is the founding move. Everything downstream follows from it. To refuse it is to begin the work the field has not yet begun.

This is the seventh installment in an ongoing series on the political economy of psychedelic enclosure. Earlier essays examined the colonial inquisition, the Wasson rupture, the era of pharmaceutical patenting, the rise of the clinical-retreat model, the April 2025 executive order, and the iboga case as the pharmacopoeia without a people. The present essay turns to the institutional apparatus through which the abstraction is performed: the conference economy and the voucher economy as paired mechanisms of capture. Subsequent essays will examine the supply-chain dimension in greater detail and the question of what an honest psychedelic medicine, built on material reciprocity rather than rhetorical acknowledgment, would actually require.

References and Bibliography

Primary Documentation

FDA Commissioner’s National Priority Vouchers, April 24, 2026 — issued to Compass Pathways (psilocybin/COMP360 for treatment-resistant depression), Usona Institute (psilocybin for major depressive disorder), and Transcend Therapeutics (methylone/TSND-201 for PTSD).

Lykos Therapeutics Complete Response Letter, August 2024; publicly released by FDA, September 4, 2025. Available via FDA transparency initiative; analyzed by Psychedelic Alpha, BioSpace, Labiotech.eu, and Psychiatric Times in September 2025 reporting.

Otsuka Pharmaceutical / Transcend Therapeutics acquisition, announced March 2026, valued at $1.225 billion.

Indigenous Medicine Conservation Fund Delegation Declaration, June 2023, issued following the MAPS Psychedelic Science 2023 conference. Available at imc.fund/journal/declaration. Signatories represented Oglala Lakota, Sioux-Lakota, Diné, Comanche, Havasupai, Bwiti, Hopi, Tewa, Cheyenne, Arapaho, Navajo, Menominee, Oneida, Mazatec, and Yawanawa traditions.

UN Independent International Commission of Inquiry on the Occupied Palestinian Territory, including East Jerusalem, and Israel. Report on genocide finding, September 2025.

Conference Documentation and Reporting

Lucid News. (2025). Protests Highlighting Conflict in Gaza Disrupt Psychedelic Culture Conference. Coverage of the Chacruna Psychedelic Culture conference at Brava Theater, San Francisco, March 29–30, 2025.

Filter Magazine. (2023, August 31). Why Indigenous Protesters Stopped a Global Psychedelic Conference. Coverage of the PS2023 indigenous protest of MAPS’ Doblin keynote.

Lucid News. (2025, July 13). Psychedelic Science 2025 Closes with Talks From Indigenous Activists, Psychedelic Parents, and Hamilton Morris.

MAPS. (2025, June 12). Plant Medicine Track at Psychedelic Science 2025 Spotlights Ancestral Wisdom, Indigenous Sovereignty, and Being in Right Relationship.

Chacruna Institute press release and conference materials, Psychedelic Culture 2025, Brava Theater for Women in the Arts, San Francisco, March 29–30, 2025.

Theoretical Frameworks

Ngũgĩ wa Thiong’o. (1986). Decolonising the Mind: The Politics of Language in African Literature. Heinemann.

Fanon, F. (1961). Les Damnés de la Terre. (English: The Wretched of the Earth, 1963.)

Rodney, W. (1972). How Europe Underdeveloped Africa. Bogle-L’Ouverture Publications.

Mies, M. (1986). Patriarchy and Accumulation on a World Scale: Women in the International Division of Labour. Zed Books.

Slobodian, Q. (2023). Crack-Up Capitalism: Market Radicals and the Dream of a World Without Democracy. Metropolitan Books.

Federici, S. (2004). Caliban and the Witch: Women, the Body, and Primitive Accumulation. Autonomedia.

Galeano, E. (1971/1997). Open Veins of Latin America: Five Centuries of the Pillage of a Continent (C. Belfrage, Trans.). Monthly Review Press.

Clinical and Pharmacological Sources

Cherian, K. N., et al. (2024). Magnesium–ibogaine therapy in veterans with traumatic brain injuries. Nature Medicine, 30(2), 373–381.

Knuijver, T., et al. (2022). Safety of ibogaine administration in detoxification of opioid-dependent individuals: A descriptive open-label observational study. Addiction, 117(1), 118–128.

Maas, U., & Strubelt, S. (2006). Fatalities after taking ibogaine in addiction treatment could be related to sudden cardiac death caused by autonomic dysfunction. Medical Hypotheses, 67(4), 960–964.

Golden, T. L., et al. (2022). Effects of setting on psychedelic experiences, therapies, and outcomes: A rapid scoping review of the literature. Current Topics in Behavioral Neurosciences, 56, 39–70.

Kaelen, M., et al. (2018). The hidden therapist: Evidence for a central role of music in psychedelic therapy. Psychopharmacology, 235(2), 505–519.

Kettner, H., et al. (2021). Psychedelic communitas: Intersubjective experience during psychedelic group sessions predicts enduring changes in psychological wellbeing and social connectedness. Frontiers in Pharmacology, 12.

Industry Reporting

Psychedelic Alpha. (2026, April 24). Breaking: FDA Awards Priority Review Vouchers to Otsuka, Compass, and Usona.

STAT. (2026, April 24). FDA fast-tracks review of new psychedelic mental health treatments.

HealingMaps. (2026, April 24). FDA Grants Priority Vouchers to Three Psychedelic Drug Programs.

Fierce Biotech. (2026, April 24). Compass, Usona and Transcend score FDA national priority vouchers amid Trump administration’s psychedelic push.

National Law Review. (2026, April 28). Psychedelics and the Executive Order — From Schedule I to Treatment Priority.

Labiotech.eu. (2025, September 8). With Lykos CRL now public, FDA opens new era of accountability.