On a Sunday afternoon in early December 2024, a musical with a score by Elton John played its final performance at the Palace Theatre, less than a month after it had opened. Its investors had put up at least $20 million (The New York Times, 2025). Nobody in the theater district would have described the closing as unusual.

Broadway, by every other measure, is booming. The June Tony Awards capped a season that broke every revenue record the industry keeps (The Broadway League, 2026), and the trade press has spent the summer using the word renaissance. Since the pandemic, 46 new musicals have opened on Broadway on roughly $800 million of investor money, according to securities filings reviewed by The New York Times (2025), and three have turned a profit. The industry’s best year and its worst run of losses are the same season, and neither fact is treated as remarkable by the people inside it. There really is no business like show business, and financially that turns out to be a uniquely bad thing.

Commercial musicals have failed to return their investors’ money roughly four times in five for as long as anyone has counted, through booms and recessions alike (The Hustle, n.d.), and the closing at the Palace was one more instance of the rule rather than an exception to it. The usual explanation, that New York has become too expensive to make them in, does not survive the data. What does survive it is less flattering. The people who finance Broadway musicals are not buying what a person buying $50,000 of stock is buying. They are buying opening night, their name in the Playbill, the party afterward, and the right to say they put a show on Broadway, with a financial return attached that is closer to a lottery ticket than to an investment (Forbes, n.d.; The Hustle, n.d.). That prestige lowers the return they need to justify the check, which is why Broadway can keep raising money at failure rates that would end a startup category, and it carries a larger implication than the industry likes to admit. Money that behaves this way is patronage, the same money that names concert halls and endows opera companies, sold with a one-in-five chance of a refund so that nobody involved has to call it a gift. The failure rate is a subsidy rate. The record season is what the subsidy bought. And the reason it has never been fixed is that everyone in a position to fix it is being paid by it.

By the Broadway League’s count, the 2025-26 season, which ran from May 26, 2025, to May 24, 2026, grossed $1,910,903,835 across 74 productions, 1,703 playing weeks, and 13,416 performances, on 14,577,322 admissions and 90.8 percent of available seats sold (The Broadway League, 2026). Adjusted for the extra week in the prior season, grosses rose 3.5 percent, attendance 1.8 percent, and the average paid admission 1.7 percent, to $131.09 (The Broadway League, 2026). The pre-pandemic record, set in 2018-19, was $1.83 billion (The Broadway League, 2019). The recovery ran from $845.4 million in the truncated 2021-22 season to $1.58 billion in 2022-23, $1.54 billion in 2023-24, $1.89 billion in 2024-25, and the record this May (The Broadway League, n.d.-c). Weekly attendance averaged roughly 280,000, against about 257,000 across the seven seasons before the pandemic (Giles, 2026).

Tammy Faye, the Elton John musical that closed at the Palace, Boop!, an origin story for the cartoon character Betty Boop, and Smash, adapted from a television series about the making of a Broadway musical, each cost at least $20 million and each closed within four months of opening at a total loss (The New York Times, 2025). Those three were the most conspicuous of the 43 musicals that lost money, but they had company at the top of the budget range. A revival of Cabaret whose budget reached $26 million after its producers converted the August Wilson Theatre into a working nightclub closed without recouping (The New York Times, 2025). A $19.5 million Gypsy starring Audra McDonald, with strong reviews, did not recoup (The New York Times, 2025). Sunset Boulevard, capitalized at between $13.5 million and $15 million depending on the report (The Economist, 2025; The New York Times, 2025), won the 2025 Tony Award for best revival of a musical and still failed to return its investment (The New York Times, 2025). The three that did make money were MJ, & Juliet, and Six (The New York Times, 2025).

A Bucknell University (n.d.) study of the 133 investor-backed musicals that opened between 2008 and 2017, well before the current cohort, found that 21.1 to 25.6 percent fully recouped. Only 8 of the 133 gave investors a realistic chance of doubling their money, because Broadway’s standard partnership terms raise royalty rates after recoupment and split the remaining profit with the producers, so a show typically needs profits of roughly three times its capitalization before an investor’s stake has doubled (Bucknell University, n.d.). Producer Ken Davenport’s (n.d.) tally of the five years before the pandemic put the recoupment rate at 20.45 percent. The industry’s own estimate that four out of five shows lose money has been in circulation since the 1960s, and no dataset since has contradicted it (The Hustle, n.d.). The rate has held through two recessions, the move from paper to digital ticketing, the pandemic closure, and a season in which the top ticket to see Denzel Washington in Othello reached $921 (The Washington Post, 2025). Whatever is wrong with the economics of the Broadway musical was wrong before anyone had heard of the current price of lumber.

Before anyone sings a note, a producer raises the capitalization, the money required to build the show and get it to opening night: sets, costumes, rehearsal salaries, studio rent, advertising, legal and management fees, and a reserve for the first weeks of performances. A major new musical is now capitalized at between $12 million and $25 million (The Hustle, n.d.), and the figure has inflated sharply: Hadestown, capitalized at $11.5 million in 2019, would cost an estimated $18 million to mount today (The Economist, 2025). Opening night does not end the spending so much as change its frequency. A large musical carries a weekly operating cost that industry estimates put between $650,000 and $800,000, before the theater takes a fixed weekly fee plus 6 to 7 percent of gross box office, collected for as long as the show runs and whether or not it has recouped. Every morning, before the first ticket of the week is sold, the show is several hundred thousand dollars underwater. A production capitalized at $15 million that grosses a strong $1 million a week, against $750,000 in operating costs before rent and royalties, might have $100,000 left for investors in a good week, and at that rate needs three years of uninterrupted business to get back to zero, with no cast turnover, no recession, no competing opening, and no additional marketing spend to hold the house. The industry calls that point recoupment, the moment investors have received their original money back, and most shows never reach it (Davenport, n.d.). A show cannot make the years go faster by growing, either. The most successful production in New York plays eight performances a week in a building with a fixed number of seats, and its only levers are price, premium inventory, and length of run. The Bucknell distribution, a fifth recouping and a handful doubling (Bucknell University, n.d.), is a venture portfolio without the one winner that pays for the other ten.

Producers, asked why, say cost, and the London comparison gives them some support. Jamie Lloyd’s Sunset Boulevard was mounted in the West End for roughly £2.5 million, about $3 million, and transferred to New York at a capitalization more than four times larger, for a staging that was substantially the same (The Economist, 2025). The West End’s average ticket, at around $81, is also roughly $50 cheaper than Broadway’s (The Economist, 2025). Once a show reaches New York it is dealing with a specific set of 41 theaters, labor agreements, rents, and operating costs that no individual producer can change, which is why the same staging costs four times as much on one side of the Atlantic as the other.

Yet a cost shock that broke a working model would show up as a break in the failure rate, and there is none. It would also, under ordinary conditions, be passed through to price, and it has not been. Broadway News (2025), which analyzes the League’s weekly grosses, put average paid admission at $128.43 in 2022-23, $125.27 in 2023-24, and $129.12 in 2024-25. Removing the three celebrity-led plays that skewed the 2024-25 average, Glengarry Glen Ross, Good Night, and Good Luck, and Othello, brings that season’s figure to $125.03, $3.40 below 2022-23 in nominal terms and, after inflation, well below the average paid in 2018-19 (Broadway News, 2025). This season’s $131.09 is an increase of about 2 percent over three years (The Broadway League, 2026), during which consumer prices in the New York area rose roughly 10 percent (U.S. Bureau of Labor Statistics, 2025). Labor tells the same story. The Production Contract ratified by Actors’ Equity and the League in October 2025 raised the minimum weekly salary for a Broadway performer from $2,638 to $2,717, with 3 percent increases in each of the following two years through September 2028 (Actors’ Equity Association & The Broadway League, 2025; Deadline, 2025), against New York area inflation of 4.1 to 4.2 percent at the time (U.S. Bureau of Labor Statistics, 2025). The people on stage are being paid less, in real terms, in each year of the agreement.

Musicals still accounted for 77 percent of tickets sold last season, but musical attendance fell from 11,758,735 to 11,232,843, a decline of more than half a million, while play attendance rose from 2.8 million to 3.32 million, up roughly 14 percent (The Broadway League, 2026). The 21 plays that opened during the season grossed a combined $463 million, more than twice what the category earned two seasons earlier (CNBC, 2026), and the average paid admission for a play, $139.55, was more than $10 above the average for a musical (The Broadway League, 2026). Plays with a celebrity in the cast sold, by The Washington Post’s (2025) analysis, at roughly twice the price of plays without one. George Clooney’s Good Night, and Good Luck, whose top ticket reached $799, had the highest single week of any of them at $4.3 million and was profitable eight weeks into a three-month run (The Washington Post, 2025). Every Brilliant Thing, a one-performer play capitalized at $5.75 million, recouped in under two months (Playbill, 2025). A new $20 million musical is an open-ended bet on sustaining demand for two or three years against an unknown audience; a 12-week run with a film actor attached can be forecast, and largely sold, before the first performance. Twenty-one plays opened this season against 12 musicals (The Broadway League, 2026). The industry has not made musicals cheaper, which the buildings and the contracts do not allow. It has made fewer bets it cannot price in advance.

Of Broadway’s 41 theaters, 33 belong to three organizations: the Shubert Organization with 17, the Nederlander Organization with 9, and ATG Entertainment with 7. A theater owner collects rent and a percentage of gross for the length of a run regardless of whether the production recoups, so a musical can lose its investors $20 million and still spend six months paying the building in full. Earlier this month, ATG agreed to sell itself to Mari, the entertainment company controlled by Ari Emanuel, in a transaction The Wall Street Journal (2026) reported at roughly $6 billion, covering ATG’s 7 Broadway houses, 10 West End theaters, and some 70 venues in total (The Hollywood Reporter, 2026). Providence Equity had bought ATG in 2013 for about $500 million (Deadline, 2026). Emanuel’s company already owns TodayTix, which handles at least 15 percent of ticket sales on Broadway and in the West End (Variety, 2026a). The landlord’s twelvefold return was earned across the same period in which four out of five of the productions inside its buildings lost money, at a valuation roughly three times the industry’s annual box office.

The League’s most recent demographics report puts the average household income of a Broadway theatergoer at $276,465, the average age at 41, and the average number of shows attended in the prior year at 5.4 (The Broadway League, 2025). The 8 percent of the audience that attended 15 or more shows bought 40 percent of all tickets sold (The Broadway League, 2025). A quarter of attendees lived in New York City, fewer than 13 percent in the suburbs, the lowest suburban share in 30 years, 42 percent elsewhere in the United States, and 20 percent abroad (The Broadway League, 2025), and international visitation to New York fell 3.2 percent in 2025, to 12.5 million (NYC Tourism + Conventions, 2026). The record was set by a smaller number of wealthier, more frequent buyers, and by a category of star-led play that is limited in supply.

In an ordinary capital market, a product that fails to return its investment four times in five would see the cost of its equity rise until only projects that could clear the higher bar got financed, and the weak ones would fail to raise money before a set was built. Nothing of the kind has happened on Broadway in 60 years (The Hustle, n.d.), and the reason is in how the shows are sold to the people who fund them.

A Broadway musical is financed through a limited liability company organized for that production alone, and its units, typically $25,000 or $50,000, are sold privately to accredited investors (Forbes, n.d.; The Hustle, n.d.). The offering documents describe the financial terms. They also describe, depending on the production and the size of the check, opening night tickets, access to the opening night party, house seats for the run, invitations to rehearsals and company events, billing, and, for the largest co-producers, eligibility for the Tony Award should the show win one. The investor base, by every industry account, is made up largely of people for whom Broadway itself is the attraction; The Hustle (n.d.) found that many are theater fans whose appreciation for the aura of Broadway outweighs the financial risk. One veteran producer’s estimate, at the pessimistic end, is that one in ten musicals recoups, and that recoupment is rare enough to warrant a press release when it happens (Forbes, n.d.). The older line, that on Broadway you cannot make a living but you can make a killing, gets quoted as folk wisdom (The Hustle, n.d.). It is a fairly exact description of the only payoff structure that attracts this kind of money.

Call it the prestige subsidy. An investor who receives value from opening night, the credit, and the association before the first distribution check is written needs less of a return to justify the check, and needs less regardless of whether the show ever recoups, because the party takes place and the Playbill is printed either way. The marginal investor is not asking for better odds; the failure rate is therefore never repriced; and the weaker projects keep getting funded on the same terms they were funded on when the four-in-five estimate was first made. The market is not failing to learn. It is clearing on a good the League does not count.

Unable to change what Broadway costs, producers have changed what Broadway is for, and they will say, correctly, that the failure rate measures the wrong window: tours, international productions, regional and school licensing, cast recordings, merchandise, and film rights generate income long after the New York production closes, and a portion flows back to the original investing entity (Davenport, n.d.). Seussical, which opened at the Richard Rodgers Theatre in 2000 and closed after 198 performances (Internet Broadway Database, n.d.) at a reported loss of around $11 million, is the standard example. By 2005 Music Theatre International had issued more than 650 licenses for the show in a single year, with 180 already booked for 2006 (Playbill, 2006), and the school edition ranked among the ten most-produced musicals in American high schools as recently as the 2021-22 school year (Educational Theatre Association, 2022). The successful version of the same pattern is Wicked, capitalized at $14 million in 2003 (Playbill, 2003) and recouped in 14 months (Fox Business, n.d.). By 2016 it had grossed $1.04 billion on Broadway and more than $2.8 billion from touring and sit-down productions across North America (Variety, 2016), and its two film adaptations have since grossed roughly $758 million and $526 million worldwide (Box Office Mojo, n.d.; Variety, 2026b). Its original backers, by one estimate, have earned annual returns of 85 to 110 percent (Fox Business, n.d.). The film grosses do not flow directly to those original stage investors, and the rights chain behind them is complicated, but the shape is clear enough: the Broadway run represents about a fifth of the franchise, and a musical can be financially mediocre as a musical and enormously valuable as intellectual property.

But notice what the argument concedes. The return on a Broadway musical is not in the Broadway run. Broadway has become an extremely expensive research and development lab, the place where a producer spends $20 million proving that a story, a score, and a title can hold an audience, after which the property travels somewhere with better economics. The New York production is the stage at which a title acquires reviews, awards eligibility, a cast recording, and the right to call every later version a Broadway musical; the money is collected in Dallas and Dayton, in school gymnasiums, on tour, and on screen, mostly by rights holders rather than by the partnership that paid for New York. A musical can destroy its investors’ capital and still create a durable asset. That is an origination market, and it is the industry’s best defense of its own economics, made by the side that receives the payments.

So the record season and the failure rate were never in tension. The audience is buying entertainment, the investor is buying prestige, and the producer is building an asset whose most valuable years arrive after the marquee comes down. Three different products are being sold on one ticket, and only one of them is being counted.

Nor was the $800 million (The New York Times, 2025) lost, in any sense a landlord would recognize. It was spent, in New York, on sets built in union shops, on salaries starting at $2,717 a week (Actors’ Equity Association & The Broadway League, 2025), on orchestras, stagehands, advertising agencies, general managers, insurance, and on fixed rent plus 6 to 7 percent of gross to three landlords. A show that closes at a total loss has paid every one of those bills in full. To the theater owner, a flop and a hit are the same customer; one just stays longer. The investors’ losses and the industry’s record revenue are the same money, counted from opposite ends.

Opera companies, ballets, and orchestras solicit gifts from the same demographic that buys Broadway units, put the donors’ names on the wall, and call the money philanthropy. Broadway solicits the same money from the same people, attaches a one-in-five chance of a refund, and calls it an investment. The lottery ticket is the whole trick. It lets the donor call himself a producer, and it lets the industry take subsidy without the board seats, the disclosure, or the stigma that usually come with it. Lincoln Center’s patrons know what they are. Broadway’s are permitted not to.

Nobody involved would do better under the truth. That is why the misdescription has lasted since the 1960s (The Hustle, n.d.), and why the periodic calls to fix Broadway’s economics misread who wants them fixed. The only fix is to price the risk honestly, and honest pricing would cut the flow of capital by something like four-fifths, which would close theaters, eliminate jobs, and end the supply of new titles that touring, licensing, and film draw on. The economics of the commercial musical are how the roughly 97,000 jobs the League credits to the industry (The Broadway League, n.d.-a), three landlords, and a licensing catalogue get paid for by people who would never write a donation check, and who have been persuaded, at a cost of about $800 million over five years (The New York Times, 2025), that they are doing something else. Broadway has solved a problem every other art form in America is still working on. It gets paid by people who do not give, because it has never once asked them to.

The fix is not going to be popular, because the fix is to stop pretending. Sell the party as a party. Separate the opening night, the Playbill credit, and the Tony eligibility from the security and price them as what they are, a consumption package for people who love the theater, structured as the gift it already is. Then let whatever capital remains price the risk on the actual odds, with the theater taking its percentage from profit rather than gross so the building carries some of the downside it currently passes to the partnership, and with new titles proven in London, regionally, or Off-Broadway before anyone spends $20 million on 44th Street (The Economist, 2025). Broadway would produce fewer new musicals, capitalized on honest terms, by investors who expect to be paid back, and the ones that opened would have a materially better chance of running long enough to matter. The people who make their living on the current volume would hate it, the landlords who just sold for $6 billion (The Wall Street Journal, 2026) would hate it more, and the patrons would lose the pleasure of being called producers. That is the way back to being a business, if Broadway wants to be one. Then again, I am an economist on the internet, and the people who actually run Broadway have had 60 years to think about this and have decided, quite rationally, to keep the party going.

To anyone who has shopped for groceries in the past two years, the image is absurd before it is alarming. Olive oil is a kitchen staple, a cultural cliché, a fixture of Mediterranean tourism brochures. It is not a commodity one associates with armed robbery. And yet, across the southern Italian region of Puglia (which produces roughly 40 percent of the country’s olive oil from some 300,000 hectares of groves and generates more than €1 billion in annual sales), producers have quietly begun moving their shipments under conditions more typical of cash transit: helicopter surveillance overhead, GPS transponders welded to their tanks, and police escorts on the road between grove and mill. Pietro Piccioni, the director of Coldiretti’s Bari branch, has told Italian press that olives must now be “guarded like diamonds at night and escorted during transfers to the mills.”

What looks on the surface like a climate-driven food-price story is, in fact, the collision of two crises that had been running in parallel for most of a century. The first, better known, is ecological: the Mediterranean climate that olive trees evolved to exploit is now breaking under them. The second, largely invisible to shoppers outside Italy, is the deep infiltration of the country’s agricultural sector by organised crime, in what Italian prosecutors and journalists have come to call the agromafia. Climate broke the harvest. Criminal networks built over five decades of postwar smuggling were waiting to capture the shock. Olive oil is the commodity on which those two stories have become indistinguishable.

The scale of the climate damage is easiest to understand through Spain, the world’s largest producer and the origin of roughly 40 percent of global olive oil supply in a normal year. In a typical season Spain yields around 1.4 million tonnes of extra virgin. In 2022, after one of the worst droughts on record in the Andalusian olive belt, it produced 666,000 tonnes, less than half the typical total. Origin prices jumped 112 percent, European retail prices nearly tripled, and wholesale extra virgin hit an all-time high in January 2024. The subsequent recovery in the 2024/25 harvest, driven by better rains in Spain and a record 450,000-tonne crop in Turkey, offered brief relief. But the United States Department of Agriculture now forecasts another 10 percent global decline for the 2025/26 crop year. Turkey is expected to fall back to 275,000 tonnes, though producers on the ground have privately told Turkish agricultural press that they expect as little as 150,000 to 180,000, which would be the country’s lowest olive oil yield in more than a decade. Tunisia is projected to drop from 340,000 tonnes to 200,000. European Union production is forecast to slip from 2.08 million tonnes to 1.98 million. Global output is expected to fall from roughly 3.36 million tonnes to 3.02 million.

The mechanism behind these failures is narrow and unforgiving. When Mediterranean flowering temperatures exceed approximately 40°C (roughly 104°F), the flowers cook, and the trees do not set fruit. The 2025 season saw extended heat waves in Andalusia, southern Italy, and the Turkish Aegean during the critical flowering window in late spring, compounding the damage. Climate scientists now attribute the 2022 Iberian drought in substantial part to anthropogenic warming, and the pattern is expected to recur with increasing frequency through the 2030s. An olive tree planted today, meanwhile, does not yield commercially for roughly fifteen years and does not reach peak production for thirty-five. There is no market signal that can rush biology. For a perennial crop, there is no replanting one’s way out of a climate that has stopped working.

Against that backdrop of sudden scarcity, the criminal infrastructure was already in place, and had been for decades. Coldiretti, Italy’s largest farmers’ association, now estimates the annual turnover of the Italian agromafia at approximately €25.2 billion, up from around €16 billion in the mid-2010s and €21.8 billion in 2017, roughly double the figure the organisation reported only a decade ago. The Italian Confederation of Agriculture has estimated that there are now some 240 agro-mafia offenses committed in the country every day. Olive oil sits near the top of the product mix, alongside counterfeit cheese, wine, and cured meats.

The broader context is harder still to absorb. Italian organised crime receipts as a whole are estimated at between 7 and 9 percent of the country’s gross domestic product. A joint Eurispes / SOS Impresa study cited by ABC News placed combined mafia turnover across Italy’s four major syndicates at roughly €130 billion per year, with profits approaching €70 billion. If consolidated into a single enterprise, those figures would make the Italian mafia the largest single business in the country. The study estimated that the four groups extract approximately €250 million a day from legitimate Italian businesses through extortion and related means, the equivalent of €10 million an hour. The ‘Ndrangheta alone, now widely regarded as the wealthiest and most powerful of Italy’s crime organisations, was estimated by Europol and the Guardia di Finanza at around €55 billion in annual global income as long ago as 2008.

Italian organised crime is not a monolith. Four major syndicates hold significant positions in the agricultural sector. The Sicilian Cosa Nostra, weakened by the 1990s state crackdown but hardly extinct, retains roughly 5,000 active affiliates and a further 20,000 collaborators across Europe and North America. The Campanian Camorra is a decentralised network of approximately 7,000 members distributed across some 180 independent clans, an organisational model that makes it both unusually flexible and unusually violent. The Calabrian ‘Ndrangheta has approximately 6,000 Italian members and, by some estimates, as many as 10,000 worldwide; built on rigid blood ties and arranged marriages, it has proven remarkably resistant to law-enforcement infiltration, and has effectively replaced Cosa Nostra as the dominant force in transatlantic cocaine trafficking. The Sacra Corona Unita, Puglia’s “fourth mafia,” emerged in the 1970s as a southern offshoot of the Camorra, fought its way to independence in the 1980s under the leadership of Giuseppe Rogoli, and has since become the dominant player in the smuggling corridors off the Puglian coast.

Their entry into agriculture was not a recent improvisation. After the Second World War, all four organisations built a sophisticated maritime smuggling network across the Adriatic, running untaxed American cigarettes out of ports in Albania and Montenegro into Puglian beaches on fast boats known as scafi blu, literally “blue speedboats,” painted dark to disappear into the night sea. The narrowest crossing, from Vlorë to Otranto, is only about 70 kilometres, making the route effectively untraceable by the naval assets of the period. By the late 1980s, the Camorra alone had built what Italian prosecutors described as a billion-euro tobacco-tax-evasion empire; the Sacra Corona Unita, operating from the Strait of Otranto, controlled much of the traffic running from Vlorë into Brindisi. When the European Union and the Italian Guardia di Finanza finally broke the cigarette trade through sustained enforcement pressure in the 1990s, the networks did not dissolve. They pivoted, applying the same boats, the same trucking routes, the same corrupted port contacts and the same bribery channels to a new product line. That product line was food, and by the 2010s it already included counterfeit Parmigiano, fraudulent Chianti, mislabelled mozzarella, and, most profitably of all, adulterated extra virgin olive oil.

The scale of the fraud has been documented for years. In 2015, a single Italian police operation seized approximately 7,000 tonnes of counterfeit olive oil, the bulk of it North African oil chemically deodorised and rebranded as Italian extra virgin, destined for American retailers. In 2016, a Carabinieri investigation dismantled a Calabrian network that had been blending imported sunflower and soybean oils, labelling the result as Italian extra virgin, and exporting it to grocery chains in North America. A 2017 case uncovered 220 tonnes of adulterated product bound for the United States market. In 2023, a joint Spanish-Italian investigation codenamed Operation Omegabad seized 260,000 litres of counterfeit olive oil in a single action after a tanker truck was stopped on a routine inspection near Manzanares. Tom Mueller, in his 2011 book Extra Virginity, reported that more than half of all olive oil imported into the United States was adulterated in some form; independent testing since has not substantially overturned his finding. The CBS News program 60 Minutes documented the broader pattern in a 2016 feature titled simply “Agromafia,” in which Italian anti-fraud officials described the lowest grade of olive oil (lampante, officially unfit for human consumption) as the workhorse raw material of the counterfeit trade.

When the 2022 harvest collapsed and prices doubled, every forged Tuscan label on a bottle of anonymous oil was suddenly worth two or three times what it had been before. The networks responded as any rational enterprise would. In December 2025 alone, Italian authorities seized 14,000 litres of untraced extra virgin in a single Puglian raid and logged additional armed robberies of commercial shipments. Coldiretti recorded 200 separate extortion cases against Puglian olive farmers across 2024: burned fields, slaughtered livestock, damaged equipment, explicit threats to sell at forced prices. Renato Nitti, the public prosecutor at Trani, has warned publicly that the Puglian mafia has been systematically underestimated in national threat assessments. The helicopters, the GPS transponders and the police escorts are the legal response to a hostile environment that now, for practical purposes, resembles a conflict zone.

Puglia faces a second, entirely separate catastrophe compounding the first. The bacterium Xylella fastidiosa, which arrived in the Salento region around 2013, has since destroyed an estimated 21 million olive trees across southern Puglia, groves that in many cases had stood for centuries. The combined effect of drought in Spain, bacterial blight in Italy, and heat stress across the entire Mediterranean basin has created the tightest supply environment the European olive oil market has seen in modern commercial history.

None of the violence, strictly speaking, is new. It is the resumption of an older pattern. Throughout the 1990s, more than a thousand olive trees were burned in mafia retribution attacks against Italian farmers who had refused to fall in line. In 1991, the Calabrian landowner Antonio Cordopatri was murdered by ‘Ndrangheta gunmen for declining to sell his groves. More recently, a Sicilian producer named Nicola Clemenza organised roughly 200 fellow farmers into a cooperative designed to bypass mafia middlemen and sell their oil directly to distributors. The response was to set fire to his car and his home with Clemenza, his wife, and his young daughter inside. The family survived. Many who have tried the same thing have not. Italian investigators believe the order for the attack came from Matteo Messina Denaro, the Cosa Nostra boss who remained a fugitive for thirty years before his arrest in 2023.

On the detection side, the state is losing. Italian food inspectors tested 8,200 olive oil samples over the past year, and approximately 1,900 (nearly a quarter) came back with irregularities. When a Belgian newspaper independently tested 32 brands pulled from supermarket shelves in Brussels in October 2025, twenty failed quality standards and one was outright adulterated. Europol’s continent-wide Operation OPSON, a rolling food-fraud investigation across 29 European countries with olive oil as a flagship target, most recently dismantled 11 criminal networks, issued 104 arrest warrants, and referred 278 individuals to judicial authorities. And yet Coldiretti’s own estimate is that only about 10 percent of counterfeit olive oil in European circulation is ever intercepted. The organisation’s broader estimate of total annual losses from food fraud in Italy exceeds €5 billion.

The technological tools to do better exist. Nuclear magnetic resonance spectroscopy can identify roughly 95 percent of adulterated oils. Isotope-ratio mass spectrometry can verify the geographic origin of a bottle by reading the atomic signature of the regional rainwater its olives drank. The method exploits the fact that each Mediterranean microclimate leaves a distinct molecular fingerprint in the water plants absorb, which persists through the pressing process. DNA-based methods can match residual olive DNA in the oil to specific varietals (Koroneiki, Arbequina, Picual) and identify claims of single-origin purity on what is in fact blended product. The Italian agricultural ministry operates a national traceability platform called SIAN that uses DNA testing at the mill, combined with an artificial-intelligence early-warning system designed to flag fraud patterns before product reaches shelves.

The limiting factor, however, is not the science but its cost. The mass spectrometers run to roughly $400,000 per unit, trained operators command six-figure salaries, and the entire European Union has only a few dozen laboratories performing serious authentication work at meaningful scale. The chlorophyll and carotenoid dyes required to turn a litre of sunflower oil into a plausible counterfeit cost approximately five dollars. When detection is nearly five orders of magnitude more expensive than fraud, the equilibrium is not in doubt.

A quieter crisis is closing in from the other side. In the most recent annual producer survey conducted by Olive Oil Times, 41 percent of Mediterranean olive farmers named labour shortages as their single greatest concern, ranking the issue ahead of climate, ahead of rising input costs, and ahead of pests. Traditional, non-irrigated, terraced groves (the kind that produce the world’s most distinctive extra virgin oils) cannot be mechanically harvested, and the younger generation has largely left them. The artisanal production base is ageing out at precisely the moment when the counterfeit supply is most aggressively scaling up. As genuine small producers exit the market, the gap they leave behind is filled by whatever is cheapest to move, label, and ship.

The wider significance of the olive oil story is not that a cherished Mediterranean product is now sometimes counterfeit. That has, in one form or another, been true for decades. The significance is that the current crisis offers a clear and troubling template for how climate shocks will interact with pre-existing criminal infrastructure in the years to come. When a staple commodity with a long production cycle is rendered suddenly scarce by a warming planet, the first beneficiaries are not the producers who still hold genuine inventory but the enterprises already equipped to move counterfeit product at scale. The shock is not absorbed by the market. It is captured by whoever is best positioned to capture it. In the Mediterranean, that happens to be four organisations that learned to move contraband across the Adriatic during the Marshall Plan, and have been refining the operation ever since.

The olive trees are still there. So are the families who have pressed oil from them for generations. Many of them are working terraced hillsides their grandfathers planted and their grandchildren will not inherit, on land now surveilled by helicopter, protected by GPS, and escorted to market under armed guard. They are simply no longer the ones selling the product to the rest of the world. Someone else is doing that now, and has been for some time, and will almost certainly be doing it for a long time to come.

Forty-five countries, at last count, offer some version of this experience. The governments of Denmark, Sweden, Spain, the United Kingdom, Australia, and South Korea, among many others, have concluded that it is absurd to ask citizens to spend hours or days manually transcribing financial data that the government has already received through other channels. The logic is not complicated. Employers report wages. Banks report interest. Brokerages report capital gains. The tax authority possesses, in most cases, every number that matters. So the authority fills in the form and sends it to the taxpayer, who checks it, corrects anything that looks wrong, and moves on.

The United States does not do this. What the United States does instead is require approximately 170 million individual filers to spend an average of thirteen hours and $290 each, per year, on the process of manually entering information that the Internal Revenue Service has possessed since January. Collectively, Americans devote 7.1 billion hours annually to tax compliance, a figure so large it resists intuition without analogy. It is the equivalent of 3.4 million people working full-time, year-round, doing nothing but filling out tax forms. That phantom workforce is thirty-eight times larger than the actual IRS. The economic cost of all this effort, accounting for both lost productivity and out-of-pocket expenses on software, accountants, and preparation services, is $464 billion per year. To put that figure in perspective, it exceeds the gross domestic product of Thailand and represents nearly two percent of the entire American economy. Not the taxes themselves. The paperwork. Americans spend more on complying with the tax code than the corporate income tax collects in revenue.

It is natural, when confronted with these numbers, to assume that something about American taxes must be fundamentally more difficult than taxes elsewhere. That the complexity is an inevitable byproduct of a large, sophisticated economy with a progressive rate structure and a federal system of government. There is some truth in this, and we will get to it. But the more immediate and more troubling explanation is simpler. There are people who make a great deal of money from the current arrangement, and they have spent decades and tens of millions of dollars ensuring that no one changes it.

The company at the center of this story is Intuit, the Silicon Valley conglomerate that owns TurboTax, QuickBooks, Credit Karma, and Mailchimp. Intuit reported $20 billion in revenue last year. The company’s public-facing identity is built around empowerment: financial confidence, economic opportunity, the democratization of money. The corporate website speaks of powering prosperity. The CEO gives interviews about removing barriers and serving the underserved. It is a compelling brand, and in many respects Intuit has built genuinely useful products. TurboTax, for all its controversies, does help millions of Americans navigate a tax system that is bewilderingly complex. The question that should trouble anyone paying attention is why the system remains so bewildering in the first place, and what role the company that profits most from the bewilderment has played in keeping it that way.

The answer is available in Intuit’s own regulatory filings, if you know where to look. In its 2012 Form 10-K, the annual report filed with the Securities and Exchange Commission, a document in which public companies are legally obligated to describe material risks to their business with something approaching candor, Intuit identified what it called “governmental encroachment” as a competitive threat. The term referred to any effort by a federal or state government to make tax filing simpler, cheaper, or free. The company warned shareholders that such encroachment “may present a continued competitive threat to our business for the foreseeable future.”

This is a remarkable disclosure, though it is rarely discussed. A company whose mission is to help people manage their finances is telling its investors that the government helping people manage their finances would be bad for business. The mission statement and the risk disclosure cannot both be true. One of them is marketing, and the other is the plan.

This particular species of corporate contradiction is not unique to Intuit. It has become one of the defining features of twenty-first-century American capitalism: the growing, often uncloseable distance between what a company says it exists to do and what the financial structure of the company actually requires it to do. The stated purpose faces outward, toward consumers, toward the public, toward the cultural conversation. The financial reality faces inward, toward the quarterly earnings call, toward the shareholder base, toward the analyst who will downgrade the stock if growth decelerates. When the two conflict, the financial reality wins. It wins every time. It wins not because the people running these companies are uniquely cynical, but because the incentive architecture of public markets makes any other outcome nearly impossible.

Consider the pattern. Facebook, now Meta, built its empire on the stated mission of connecting people and building community. Its financial model required maximizing the time users spent on the platform, which meant algorithmically promoting content that triggered emotional arousal, which meant that outrage, conspiracy, and division were not bugs in the system but the most reliable revenue generators the company had. Internal research, later leaked by a whistleblower, showed that the company’s own data scientists had concluded that Instagram was materially worsening body image and mental health outcomes among teenage girls. The company’s response was not to change the product. It was to suppress the research. Connecting people was the mission. Addicting people was the model. The model won.

Or consider the pharmaceutical industry, where companies like Purdue Pharma marketed OxyContin under the banner of compassionate pain management, of liberating patients from suffering, while simultaneously funding studies designed to minimize the drug’s addictive properties and dispatching sales representatives to encourage doctors to prescribe it at higher doses and for longer durations than the evidence supported. The stated purpose was healing. The financial incentive was volume. Volume won, and it won catastrophically.

Or consider the food industry, where major corporations spend billions on marketing that emphasizes nutrition, wellness, and feeding families well, while their product development teams engineer processed foods for what the industry internally calls “bliss point,” the precise combination of sugar, salt, and fat that maximizes craveability and repeat purchase. The mission is nourishment. The model is engineered overconsumption. The model wins.

Or consider the banking industry in the years before 2008, when the largest financial institutions in the world marketed themselves as stewards of their customers’ financial futures while simultaneously packaging subprime mortgages into securities they knew to be unstable, selling those securities to pension funds and municipalities, and in some cases actively betting against the products they were selling. The mission was financial stewardship. The model was fee extraction. The model won so thoroughly that it nearly destroyed the global economy.

The Intuit case is, in some respects, less dramatic than any of these. Nobody died because TurboTax hid its free product from Google. Nobody lost their home because filing taxes took thirteen hours instead of five minutes. But what makes the Intuit story so clarifying, so useful as a case study in corporate misalignment, is its purity. There is no ambiguity about what happened, no room for the company to claim that its product had unintended side effects or that the harm was an unforeseen consequence of scale. Intuit identified a specific public good, the simplification of tax filing, as a threat to its business. It then spent tens of millions of dollars to prevent that public good from being realized. It hired former government officials to do the lobbying. It funded the campaigns of the lawmakers who would vote on the relevant legislation. And when the government, despite all of this, managed to build a free filing tool that worked, Intuit helped orchestrate its destruction. The mission and the model did not merely diverge. They pointed in precisely opposite directions, and the company chose the model with the kind of institutional discipline that most organizations reserve for their core products.

The lobbying record traces the shape of that choice with unusual precision. Since 1998, Intuit has spent more than $45 million on federal lobbying. H&R Block has spent an additional $38 million. Together, the two dominant players in the industry have directed $83 million toward influencing the way Americans file their taxes. Intuit set a company record of $3.5 million in 2022, the year the IRS began developing a free filing tool, and then exceeded it in 2023 and 2024 with roughly $3.9 million each year. The company has employed dozens of revolving-door lobbyists, former IRS officials, former congressional staff, former administration appointees, people who helped build the regulatory architecture of the tax system and who now earn their living ensuring that no one improves it.

The strategy has a longer history than most people realize. In 2002, during the first term of George W. Bush, the Office of Management and Budget proposed a straightforward modernization: the IRS would develop a free online portal that would allow taxpayers to file electronically at no cost. The logic was unassailable, even by the standards of an administration philosophically committed to small government and private-sector solutions. Taxpayers were paying private companies to transmit information to the government that the government already had. It was, as one OMB official put it at the time, an obvious improvement. Nobody should have to pay extra just to file a return.

Intuit moved before the proposal was public. The company’s lobbyists contacted the White House directly. Sympathetic members of Congress, including the representative from Intuit’s own congressional district, began publicly accusing the IRS of harboring secret plans to undercut the private sector. Within months, the initiative was dead, replaced by a compromise that, in hindsight, was something closer to a capitulation. Under a memorandum of understanding between the IRS and a consortium of tax software companies calling itself the Free File Alliance, the federal government agreed not to build its own filing system. In exchange, the industry pledged to offer free filing to lower-income Americans through commercial software products. It was, on paper, a reasonable arrangement. It was, in practice, a disaster.

What followed was one of the more brazen episodes of consumer manipulation in recent corporate history. Intuit and H&R Block, both members of the Free File Alliance, embedded code on their websites instructing search engines not to index the free versions of their products. A taxpayer searching Google for “free tax filing” would find paid TurboTax products, TurboTax advertisements, and TurboTax landing pages that used the word “free” prominently, but not the actual Free File product the company was contractually obligated to offer. The free version was technically available, in the way that a fire exit is technically available in a building where someone has removed all the signs pointing to it. Meanwhile, the paid products employed what regulators would later describe as dark patterns: interfaces designed to funnel users through a process that appeared to be free until, after an hour of data entry, a fee materialized. By that point, most people paid rather than start over. The Federal Trade Commission filed a formal complaint accusing Intuit of deceptive advertising. All fifty state attorneys general sued. The resulting settlement was $141 million, and affected consumers received checks for as much as ninety dollars. Intuit’s annual revenue that year was north of fourteen billion. The settlement was not a penalty in any meaningful sense. It was the cost of doing business, and the business continued.

This is the juncture at which most analyses of corporate misalignment go wrong. They treat these episodes as failures of individual character, as evidence that specific executives made bad decisions or that specific corporate cultures became corrupted. But the Intuit story does not require a villain any more than the Facebook story requires a villain or the Purdue story requires one, though Purdue certainly had a few. What these stories require is a structure, and the structure is the same in every case.

A company identifies a mission that resonates with the public and with its own employees. It builds products that serve that mission, at least initially, and often quite well. It goes public, or it takes on growth-stage investors, and the financial expectations of external capital begin to exert gravitational force on every decision the company makes. Growth targets are set. Margins are monitored. Analysts model future earnings and punish any deviation from the trajectory. And gradually, sometimes over years and sometimes over decades, the company’s actual behavior migrates away from its stated purpose and toward whatever activity most reliably converts into quarterly revenue growth. The mission does not disappear. It remains on the website, in the investor presentations, in the CEO’s talking points at conferences. But it becomes decorative. The operating logic of the company is determined not by what it says but by what it measures, and what it measures is the thing that Wall Street rewards.

In Intuit’s case, what Wall Street rewards is the continued inability of the American taxpayer to file a return without paying for the privilege. Twenty billion dollars in annual revenue does not materialize from a tax code that works for people. It materializes from a tax code that confuses people. And the company has demonstrated, through a quarter-century of lobbying, litigation, and regulatory maneuvering, that it understands this with perfect clarity.

The saga might have ended with the Free File debacle, as one more cautionary tale about corporate influence in Washington, except that in 2024 the IRS finally built the thing that had been killed two decades earlier. The program was called Direct File, and it was, by every available measure, a success. It was a clean, government-operated website that allowed taxpayers with relatively straightforward returns to file directly with the IRS, for free, without commercial software, without upsells, without the dark patterns that had made the Free File experience so corrosive. The user satisfaction numbers were extraordinary: ninety-four percent of filers rated the experience as excellent or above average. The Government Accountability Office reviewed the pilot and recommended expanding it to all fifty states. The Taxpayer Advocate Service, an independent watchdog within the IRS itself, said Direct File allowed taxpayers to easily file accurate returns fast. Roughly 300,000 people used it in its first full year, a small fraction of the 170 million individual returns filed annually, but the tool was designed to scale, and it was scaling.

The tax preparation industry responded as if its survival depended on killing Direct File, which, in a sense, it did. Intuit spent nearly four million dollars in both 2023 and 2024 fighting the program specifically. The company donated one million dollars to Trump’s inaugural committee. Twenty-nine House Republicans wrote a letter to the incoming administration asking to shut Direct File down on the first day of the new presidency. Public Citizen subsequently reported that those twenty-nine legislators had collectively received $1.8 million in campaign contributions from opponents of the program over the course of their political careers. That works out to roughly $62,000 per congressman, a figure that seems, if anything, insultingly modest given what was at stake.

The effort succeeded. By the spring of 2025, the Trump administration had reduced the Direct File team by eighty percent. By summer, the IRS commissioner confirmed in a Senate hearing that the program was gone. By November, the IRS sent a quiet email to all twenty-five participating states informing them that Direct File would not be available for the 2026 filing season and that no future launch date had been set. The team that built it had been scattered through a combination of resignations and separations. The website went dark. A program that taxpayers loved, that independent reviewers praised, that cost thirty-two million dollars to build, roughly the price of a single military helicopter, was dismantled at the urging of an industry that earns twenty billion dollars a year from the problem the program was designed to solve.

It would be satisfying, at this point, to conclude that the story is primarily about Intuit, that if one company could be restrained, the filing experience would improve dramatically. This is partially correct but fundamentally incomplete. The tax preparation industry has exploited a system that was already broken. The deeper question is why the system was broken in the first place, and the answer to that lies not in Silicon Valley but in Congress.

The federal tax code is approximately 75,000 pages long. It has tripled in size since the late 1970s, when Jimmy Carter called it “a disgrace to the human race.” It contains more than 165 individual tax expenditures, the credits, deductions, exclusions, and preferential rates that reduce the amount of revenue the government collects. The number has more than doubled since 1974, when the Treasury Department first began tracking them. The aggregate cost of these provisions is staggering: in 2025, tax expenditures cost the federal government more than two trillion dollars. That is larger than the entire Social Security budget. Larger than national defense. Larger than Medicare. The American tax code gives away more money than the government spends on any single program.

The reason for this is both simple and, once you understand it, difficult to unsee. Congress has discovered that the tax code is the most politically convenient vehicle in Washington for distributing public money. When a lawmaker wants to subsidize childcare, or encourage homeownership, or reward investment, or promote clean energy, or help families afford health insurance, there are two paths available. One is to create a new government program, to propose legislation, stand before cameras, vote on a spending bill, and accept the political consequences of being seen to expand the federal government. The other is to insert a tax credit or deduction into the internal revenue code, which achieves substantially the same economic result but is framed as tax relief rather than new spending. The first path invites attack ads. The second invites applause. The money is identical. The optics are not.

The result is that the American 1040 has become, without most people realizing it, the largest social welfare delivery system in the country. It is simultaneously a $226 billion health insurance subsidy, through the exclusion of employer contributions to medical premiums. It is a $304 billion preferential rate for capital gains and dividends. It is a $120 billion child welfare program, through the child tax credit. It is a $67 billion wage supplement for low-income workers, through the earned income tax credit. It is a housing program, through the mortgage interest deduction. It is an education policy, through tuition credits. It is a retirement savings incentive, through the exclusion of 401(k) contributions. It is a climate policy, through credits for electric vehicles and solar panels. It is, in the unkind but accurate formulation, six federal agencies in a trench coat, arriving at your kitchen table every April with 179 pages of instructions.

And every one of these provisions, once created, becomes functionally immortal. The mortgage interest deduction has been defended for decades by a coalition of realtors, homebuilders, and mortgage lenders who treat any proposed modification as an existential threat. The employer health insurance exclusion persists in part because the health insurance industry has built its entire distribution model around it. The child tax credit has a constituency of millions of families who depend on the annual refund. The pass-through business deduction, created in 2017, instantly generated a lobbying apparatus of small business groups and accounting firms dedicated to its preservation. This is the ratchet at the heart of the tax code: provisions are easy to create, because every credit has a beneficiary, and almost impossible to remove, because every beneficiary has a lobbyist.

This dynamic, it is worth noting, is itself another manifestation of the misalignment between stated purpose and structural incentive, only this time the institution is not a corporation but a legislature. Members of Congress will tell you that they want to simplify the tax code. Many of them genuinely believe it. Simplification polls extraordinarily well. It is one of the few policy positions that commands broad support across partisan lines. But the incentive structure of congressional politics pushes relentlessly in the opposite direction. A lawmaker who adds a new tax credit gets to announce a benefit to a specific constituency. A lawmaker who eliminates an existing credit makes an enemy of that constituency and every lobbyist attached to it. The political reward for adding complexity is immediate and concentrated. The political cost of reducing complexity is immediate and concentrated. The benefit of simplification is diffuse and abstract. Nobody has ever won a congressional race on the platform of eliminating the mortgage interest deduction, even though most economists agree the policy is regressive and inefficient. The realtor lobby, on the other hand, has ended more than a few political careers.

The result is a system in which every participant, from the software company to the congressman to the lobbying firm to the trade association, is individually behaving in a way that is entirely rational given the incentives they face, and collectively producing an outcome that is irrational by any measure. This is the deep structure of the problem. It is not a conspiracy. It is not coordination. It is something more durable than either, because conspiracies can be uncovered and coordination can be disrupted, but a system in which misaligned incentives produce harmful outcomes through ordinary institutional behavior is almost impossible to reform. Everyone is following the rules. The rules are the problem.

What emerges from all of this is not a single failure but a feedback loop in which each component reinforces the others. Congress adds complexity to the tax code because hiding policy in the revenue code is easier than governing in the open. The complexity creates an enormous market for tax preparation services. Americans spend $148 billion per year in out-of-pocket costs alone. The tax preparation industry captures a share of that market and redirects millions of dollars into lobbying aimed at ensuring that nobody simplifies the filing process. And the painful, expensive, time-consuming experience of filing persuades ordinary Americans that their taxes are just inherently difficult, that the confusion is a feature of the math rather than a product of choices made by identifiable people with identifiable financial interests in maintaining it.

Every actor in this loop is behaving rationally. Lawmakers are optimizing for re-election. The industry is optimizing for revenue. The IRS, chronically underfunded and operating on technology built in the 1960s, is optimizing for institutional survival. None of them are optimizing for the person sitting at a kitchen table on a Sunday in March, staring at a W-2, trying to figure out the difference between box 12a and 12b.

There is a term of art in organizational theory for the moment when a company’s financial imperatives fully eclipse its founding purpose. It is called “mission drift,” and it is treated, in business school case studies, as something that happens to organizations, a kind of institutional weather event that descends on companies that fail to maintain cultural discipline. But this framing is misleading, because it implies that the drift is an aberration, that the natural state of a company is alignment between its mission and its incentives, and that misalignment represents a departure from equilibrium. The Intuit story, like the Facebook story, like the Purdue story, like the pre-2008 banking story, suggests the opposite. Misalignment is the equilibrium. It is the inevitable destination of any organization that operates under financial incentives sufficiently powerful to override its stated values. The mission does not drift. It is pushed, steadily and predictably, by the same forces that push every public company: the quarterly earnings cycle, the analyst consensus, the share price, and the fiduciary obligation to maximize returns for shareholders. What we call mission drift is really just the market working as designed.

Intuit’s executives almost certainly believe in financial empowerment. Many of the company’s employees are undoubtedly motivated by the desire to help people navigate complex financial systems. But belief is not the same as behavior, and in a publicly traded company, behavior is determined by incentives, not convictions. The incentive is to preserve the market. The market exists because the system is confusing. And so the company, regardless of what anyone inside it believes or intends, operates as a mechanism for perpetuating confusion. This is not a moral judgment about the people involved. It is a structural observation about the institution they inhabit. The machine does what the machine is built to do.

Researchers at the Treasury Department, the Federal Reserve Bank of Minneapolis, and Dartmouth College have estimated that the IRS could correctly auto-complete returns for between 62 and 73 million individual filers, somewhere between forty-one and forty-eight percent of all taxpayers, using data it already possesses. The accuracy was highest for low- and moderate-income households, the very people who can least afford to pay for professional preparation. The information is there. The technology is proven. The concept has been validated in nearly half the developed world. What is not there, and what has never been there, is the political will to overcome the constellation of interests that benefits from the status quo.

This is perhaps the most useful thing the Intuit story can teach us, not as a story about taxes specifically, but as a parable about how modern institutions fail the people they claim to serve. We live in an economy saturated with mission statements. Every company has one. They are printed on walls, embedded in onboarding decks, recited at all-hands meetings. They speak of empowerment, of community, of making the world better, of putting the customer first. They are not, in most cases, lies. They are aspirations that the financial structure of the company renders impossible to fulfill. The mission points north. The incentives point south. And the company, which must answer to its shareholders every ninety days, walks south while talking about north. The gap between the two is where the actual story of American commerce is written, in the space between the press release and the balance sheet, between the Super Bowl ad and the SEC filing, between what a company tells you it is and what its money says it needs to be.

Intuit’s mission statement says empowerment. Its balance sheet requires confusion. Its lobbying budget ensures that confusion prevails. This is not, as a sympathetic observer might frame it, a tension the company is navigating with imperfect but good-faith effort. It is the business. A company worth over $150 billion on the public markets did not reach that valuation by hoping the government would make its core product unnecessary. It reached that valuation by making very sure, year after year, dollar after dollar, that the government never would.

The tax code is complicated because Congress wants a junk drawer of policy tools and nobody will throw anything out. The filing process is complicated because an industry that earns billions from the complexity has spent decades and tens of millions of dollars making sure nobody cleans it up. One of these is a governance problem that would be difficult to solve under any circumstances. The other is a business model masquerading as a public service, with a mission statement about empowerment and a lobbying disclosure that says the opposite.

The next time you lose a Saturday to your taxes, it may be worth remembering that the IRS had your numbers before you sat down. That a postcard would have sufficed. That the software you are paying for exists to solve a problem that the company that made it paid to preserve. That somewhere in the Netherlands, a man who earns more than you and has a more complicated financial life cracked a beer, clicked a button, and never thought about it again. And that the company standing between you and that experience has a mission statement on its website and a 10-K on file with the SEC, and if you read them both carefully, you will find that they are describing two entirely different companies. One of them wants to help you. The other needs you to stay confused. You are already doing business with the second one.

Subscribe now

For more than a century, De Beers controlled between 80 and 90 percent of the global diamond supply. They didn’t sell everything they mined. They stockpiled. They bought up competing production to keep it off the market. They created a culture, “A Diamond is Forever,” that turned a transparent rock into an emotional obligation. The strategy worked so well that by the mid-20th century, a commodity that exists in enormous geological abundance had become one of the most expensive consumer goods on earth per unit of weight.

The mechanism was simple: control supply. Let demand build against a wall you own. The spread between your production cost and the consumer’s willingness to pay is your profit. And that profit persists exactly as long as the wall holds.

De Beers held the wall for over a hundred years. Then, in the early 2000s, lab-grown diamonds entered the market. Not as novelties or imitations, as chemically identical products at a fraction of the cost. Lab-grown stones now account for roughly 20 percent of the engagement ring market in the United States. De Beers responded by launching its own lab-grown line, Lightbox, at aggressively low price points, an attempt to segment the market and protect natural diamond premiums by positioning synthetics as casual, disposable, lesser. Whether this works long-term is an open question. What is not an open question is that the wall cracked, and the crack came from outside.

If you wanted to find a different version of the same mechanism operating at civilizational scale, you would look at petroleum.

OPEC, the Organization of the Petroleum Exporting Countries, was founded in 1960 by five nations that collectively controlled enough crude oil production to move global prices by turning a valve. The mechanism is identical to De Beers’: coordinate supply, let demand build, harvest the spread. OPEC calls them production quotas. The economic logic is the same as a stockpile.

The critical difference is governance. De Beers was a single company. OPEC is a cartel of sovereign nations, each with its own budget pressures, political crises, and incentive to cheat. And they do cheat. When a member exceeds its quota, pumping more than agreed to capture short-term revenue, the additional supply depresses prices for everyone. The 2014 oil price collapse, which saw crude fall from $107 to $26 per barrel in eighteen months, was triggered in significant part by members producing above quota while North American shale simultaneously flooded the market with new supply.

OPEC’s wall is the same as De Beers’. The threat is different. For De Beers, the crack came from outside, a technological alternative the monopoly didn’t control. For OPEC, the crack comes from inside, a member who breaks discipline for short-term gain. External disruption versus internal indiscipline. Same wall. Two different ways it falls.

Hermès International does not discuss its production numbers with particular enthusiasm. What is known is that the company manufactures roughly 70,000 Birkin bags per year across its ateliers in France. It maintains a waiting list that can extend past a year. It does not advertise. It does not discount. It does not sell through third parties. A standard Birkin retails for between $10,000 and $15,000. On the secondary market, certain configurations trade for $30,000 to $200,000. In 2023, a Himalaya Niloticus Crocodile Birkin sold at auction for $450,000.

The mechanism is, again, supply control, but with a refinement the commodity operators don’t employ. Hermès has made the act of purchasing the product into a performance of worthiness. You don’t simply have the money. You must cultivate a relationship with a sales associate. You must demonstrate, through a pattern of prior purchases, that you are a client who deserves access. The wait is not a bug. The wait is the product. The scarcity is not an externality of production. It is the experience being sold.

This is the most sophisticated version of the mechanism because it converts the supply constraint from a frustration into a feature. De Beers makes you want the rock. OPEC makes you need the fuel. Hermès makes you want to be the kind of person who is allowed to buy the bag. Each is a scarcity economy. Each is built on the same rule. But the Hermès version is self-reinforcing in a way the others are not, the more restricted the access, the more the consumer desires it, which makes further restriction more profitable, which makes the consumer desire it more. It is an engine that runs on its own exhaust.

For this reason, Hermès is the hardest version of the wall to crack. There is no lab-grown Birkin. There is no shale oil equivalent for leather goods. The only thing that could break the Hermès model is Hermès itself, a decision, for whatever reason, to produce more. To date, they have not made that decision. The wall holds.

In March of 2020, the Federal Reserve cut interest rates to near zero and began purchasing $120 billion per month in Treasury securities and mortgage-backed bonds. The stated objective was to support the economy through the shock of a global pandemic. The operational effect was to flood the financial system with liquidity on a scale that had no precedent outside of wartime.

What followed was a two-year period in which virtually every asset class on earth appreciated simultaneously. Equities. Real estate. Cryptocurrency. NFTs. SPACs. Vintage cars. Trading cards. And, notably, sneakers. A pair of limited-edition Air Jordans purchased on release day for $180 could be resold within a week for $400, $500, sometimes more. The secondary sneaker market, facilitated by platforms like StockX and GOAT that functioned effectively as exchanges, complete with bid-ask spreads and historical price charts, crossed $6 billion in annual transaction volume by 2025.

The mechanism at work was not, strictly speaking, supply control. It was its mirror image: demand inflation. The Federal Reserve did not limit the supply of dollars. It expanded it. The result was that dollars became cheaper relative to goods, and goods, including speculative goods with no intrinsic utility beyond status and scarcity, became more expensive relative to dollars. The sneaker flippers of 2020 and 2021 were not primarily responding to the product. They were responding to the monetary environment. Easy money needed somewhere to go. Sneakers, with their transparent resale data and built-in authentication infrastructure, became one of the most accessible speculative vehicles available to a twenty-three-year-old with a smartphone.

This is the part that almost nobody discussed at the time, and that most retrospective analyses still underweight: the sneaker boom was not a sneaker phenomenon. It was a monetary phenomenon wearing a sneaker costume. And like every monetary phenomenon, it was temporary, not because sneakers stopped being desirable, but because the conditions that made flipping them profitable were a function of interest rates, stimulus payments, and excess liquidity, all of which reversed beginning in 2022.

The economist Richard Cantillon observed in the 18th century that when new money enters an economy, it does not arrive everywhere at once. It arrives first at the point of injection, typically banks and financial institutions, and then disperses outward, with each subsequent recipient benefiting less than the one before. The people closest to the source of new money get rich. The people furthest away experience the inflation without the windfall. This is known as the Cantillon Effect, and it describes, with uncomfortable precision, the distribution of gains in the sneaker resale market between 2020 and 2023. The resellers with SNKRS access, bot networks, and retail backdoors were closest to the supply. The consumer paying $500 for a shoe that cost $180 to produce was furthest away. The distance between those two positions was the speculation. When liquidity tightened, that distance collapsed.

James Jebbia opened Supreme’s first store on Lafayette Street in Manhattan in 1994. The store was small. The production runs were small. The drops were weekly, unannounced, and limited. If you got there late, the product was gone. There was no restock. There was no apology. There was next Thursday.

Supreme built a $2.1 billion valuation on this principle. The scarcity was so disciplined that a red box logo on a white T-shirt could resell for twenty times its retail price. A Supreme brick, a literal clay brick stamped with the brand’s logo, sold for $30 retail and traded for $150 on secondary markets. The absurdity was the point. The product was irrelevant. The access was the product.

In 2020, VF Corporation acquired Supreme for $2.1 billion. What followed was predictable to anyone who had watched De Beers or OPEC or the Federal Reserve interact with a scarcity mechanism: the new owners expanded supply. More stores. More drops. Wider distribution. The calculus was reasonable from a corporate finance perspective, you acquire an asset, you scale it, you harvest the margin at volume. The calculus was catastrophic from a scarcity economics perspective, because the margin at volume was precisely the margin that scarcity had created. Remove the scarcity, and the margin doesn’t compress. It evaporates. The secondary market for Supreme products has not recovered. It probably will not.

VF Corporation sold Supreme in 2024 for approximately $1.5 billion, a loss of roughly $600 million in four years on a brand whose entire value proposition was the one thing the new owners chose to dismantle.

These are five different industries, five different products, five different time horizons, and five different governance structures. But they are all one story, and the story has one rule: the value of a scarcity-dependent asset is a function of supply discipline, and the discipline is only as durable as the authority enforcing it.

De Beers held discipline for a century and cracked from outside.

OPEC holds discipline intermittently and cracks from within.

Hermès holds discipline absolutely and hasn’t cracked at all.

The Federal Reserve expanded supply deliberately, inflated every asset class on earth, and then reversed course, revealing who was swimming naked, as the saying goes.

Supreme held discipline for twenty-six years and was then acquired by someone who didn’t understand what they had purchased.

Each of these is a useful case study. None of them is a perfect one, because in each case the data is either proprietary, delayed, or incomplete. De Beers does not publish real-time transaction prices for rough diamonds. OPEC’s actual production figures are notoriously unreliable. Hermès does not disclose secondary-market data. Federal Reserve policy operates through so many intermediaries that isolating its effect on any single asset class requires modeling assumptions that reasonable economists will disagree about. And Supreme’s resale data, while once robust on platforms like StockX, was always a small and idiosyncratic market.

But there is one scarcity economy where the collapse happened in public, at scale, with transaction-level data available to anyone with an internet connection. Where you can watch the monopoly flood its own supply, observe the price premium disintegrate, track capital rotate to alternatives, and measure the market find its new floor, all on a publicly accessible price chart, updated daily, denominated in dollars, with authentication receipts.

That economy is sneakers. And the company that broke the wall is Nike.

In 2020, Nike and Jordan Brand controlled approximately 71 percent of the sneaker secondary market by transaction volume on StockX. Seventy-one percent. One corporate entity, seven out of every ten dollars spent. This was, in structural terms, closer to De Beers’ historical diamond monopoly than to any other scarcity economy operating in consumer goods. And like De Beers, the monopoly was maintained not through superior product but through superior supply control, limited production runs, algorithmic release mechanisms, controlled retail distribution, and a culture of collection that Nike had spent four decades cultivating.

The scarcity was not accidental. It was engineered. A pair of Air Jordan 1 Retro High OGs cost Nike roughly $25 to manufacture. They retailed for $170 to $180. On the secondary market, sought-after colorways traded for $400 to $600 within days of release. The distance between the manufacturing cost and the resale price, the full spread, was the scarcity premium. And Nike captured the retail slice while allowing resellers to capture the secondary slice, which in turn generated the cultural mythology of exclusivity that justified the retail price in the first place. The resellers were not gaming the system. They were the system. They were Nike’s unpaid marketing department, converting limited supply into visible demand that made every subsequent release feel more urgent.

Then Nike broke the rule. Between 2021 and 2024, the company systematically increased production on its most popular silhouettes. The Nike Dunk Low, which had been a resale staple commanding $250 to $350, was produced in such volume that pairs began appearing on outlet shelves. The Air Jordan 1 was restocked repeatedly. The Reimagined series, positioned as premium, flooded the market. Nike’s stated rationale was to meet demand and drive direct-to-consumer revenue. The actual effect was to dismantle the scarcity architecture that had generated the demand in the first place.

The company eventually acknowledged the error. In late 2024, Nike cut production on limited Jordan releases by approximately 35 percent, an attempt to rebuild the wall. But the correction arrived after the collapse. In commodity economics, this is the equivalent of OPEC slashing production after a price war has already destroyed margins. You can restrict supply going forward. You cannot un-mine the diamonds. You cannot un-pump the oil. You cannot un-flood the Dunks.

Simultaneously, and remarkably, Nike made a second decision that compounded the first. Under its previous CEO, the company had aggressively shifted toward a direct-to-consumer sales model, pulling product from wholesale partners, the Foot Lockers, the Finish Lines, the regional sneaker boutiques that had been Nike’s distribution backbone for decades. The strategic logic was margin capture: selling directly to consumers through Nike.com and Nike’s own stores eliminated the wholesale markup. The structural effect was to voluntarily surrender shelf space at every major footwear retailer in the United States.

De Beers, for all its errors, never left the distribution channel. Nike did.

What happened next is the part that transforms this from a brand management case study into an economics case study. In commodity markets, when a monopoly loses supply discipline, capital does not evaporate. It rotates. If OPEC floods and oil prices collapse, investment flows to natural gas, renewables, or shale. If De Beers loses pricing power, consumers shift to lab-grown or alternative gemstones. The capital finds the next scarcity.

In sneakers, the rotation is visible on StockX with a clarity that no commodity market can match. New Balance grew to $9.2 billion in global revenue in 2025, marking its fourth consecutive year of growth above 20 percent. The Asics Gel-1130, a retro running shoe with no celebrity collaboration and no hype infrastructure, became the single best-selling sneaker on StockX in 2025, outselling every Jordan, every Dunk, every Nike silhouette on the platform. Mizuno, a brand previously associated almost exclusively with volleyball and baseball, posted 124 percent year-over-year growth on StockX to become the fastest-growing sneaker brand on the platform. Brooks grew 1,508 percent. Anta, propelled by Kyrie Irving’s signature shoe after his departure from Nike, posted 1,901 percent growth in 2024 and continued surging with an additional 51 percent increase in the first half of 2025.

Nearly two hundred brands, not twenty, not fifty, two hundred, set all-time annual sales records on StockX in 2025. The global sneaker market, valued at approximately $104 billion, grew. What contracted was not the market. What contracted was Nike’s share of it. This is a capital rotation, not a market collapse. They are opposite economic events. They are being reported as the same story.

The demand that replaced the speculation is, in its own way, as revealing as the rotation itself.

Nike’s fastest-growing product category on StockX in 2025 was not a basketball shoe. It was not a retro release. It was the ReactX Rejuven8, a recovery slide, a foam slipper designed for post-workout comfort, which posted 5,811 percent year-over-year growth on the platform. Nike’s newest silhouette, the Mind 001, a $95 mule featuring 22 foam nodes that the company claims stimulate sensory pressure points in the foot, sold out in under five minutes on its initial release and resells for approximately double retail. A collaboration between Nike Mind and Hiroshi Fujiwara’s Fragment Design traded at $523 against a $140 retail price.

The financial press has covered this as a “comfort trend.” It is not a trend. It is a demographic transition. The consumer cohort that built the sneaker speculation, millennials who entered the hype economy between 2015 and 2021, are now between 30 and 42 years old. They have aging bodies and competing financial obligations. They are not lining up for SNKRS raffles at 7 AM. They want shoes that address plantar fasciitis, not shoes that address social status. This is not a shift in taste. It is a shift in the musculoskeletal reality of the customer base.

In any speculative market, the transition from speculator to end user represents the discovery of fundamental value. When the people buying a house intend to live in it, the price reflects utility. When the people buying a shoe intend to wear it, the price reflects comfort, durability, and design. The speculative premium, the portion of the price attributable to artificial scarcity and the expectation of resale profit, disappears. What remains is the floor. And the floor of the sneaker market, it turns out, is higher than most obituaries suggest. A $104 billion global market growing at 6 to 7 percent annually is not a dying industry. It is an industry that has discovered, perhaps for the first time, what it is actually worth when the speculation is subtracted.

There is one more variable, and it has nothing to do with sneakers.

In April 2025, the United States imposed tariffs of 46 percent on imports from Vietnam, 32 percent on Indonesia, and 145 percent on China. These are the three countries where the vast majority of global sneaker production is concentrated. Nike manufactures approximately 50 percent of its footwear in Vietnam, across 131 contract factories employing nearly 460,000 workers. Adidas sources 39 percent of its footwear from the same country. Ninety-nine percent of all shoes sold in the United States are imported. There is, in practical terms, no alternative manufacturing base of sufficient scale.

UBS analysts estimated that offsetting the Vietnam tariffs alone would require retail price increases of 10 to 12 percent. The Footwear Distributors and Retailers of America calculated that a sneaker with a pre-tariff retail price of $155, manufactured in Vietnam, would need to retail at approximately $220 under the full tariff schedule. Nike has already raised prices on selected higher-end models. Seventy-five percent of American consumers reported in a 2025 FDRA survey that they had noticed higher shoe prices over the prior twelve months.

This is the external shock meeting the internal collapse. Nike’s scarcity premium was already gone. Now the cost floor is rising. This is a double repricing event, and it creates an unusual condition in the secondary market. Resellers who purchased inventory at pre-tariff retail prices are holding goods whose replacement cost is rising. If retail reprices upward by 10 to 15 percent, the spread between the reseller’s acquisition cost and the new retail floor constitutes an unintentional arbitrage opportunity. The tariff, designed as a punitive measure on imports, may function as a margin restoration mechanism for the very resale market that had just been declared dead.

Whether this window is large enough and durable enough to matter is an open question. What is not an open question is that the sneaker secondary market now has a pricing dynamic that no other consumer goods category has, real-time public data, transparent transaction history, and a tariff-induced cost shock whose effects can be measured daily.

One brand is positioned for this environment in a way that no competitor can replicate on any relevant timeline.

New Balance, founded in 1906 in Boston, operates manufacturing facilities in Maine, Massachusetts, and a newly opened $70 million plant in New Hampshire. It is the only major sneaker brand with meaningful domestic production capacity in the United States. Its competitors, Nike, Adidas, Asics, Puma, have 90 percent or more of their production concentrated in Vietnam, Indonesia, and China, the three countries subject to the most severe tariff increases. New Balance is not immune to tariffs; it imports product as well. But it has a domestic manufacturing base that provides structural insulation its competitors lack entirely.

The company grew 19 percent in 2025 to $9.2 billion in global revenue and has indicated it is targeting $10 billion in 2026. Its CEO, Joe Preston, told CNBC that the company “could be bigger if we wanted to, but we use a selective distribution approach to manage the brand and not oversaturate.” This is, whether Preston intends it as such, a direct articulation of the scarcity rule. It is also a notable contrast to Nike’s behavior over the same period.

New Balance’s tariff position is not widely discussed in sneaker media, where the brand’s growth is typically attributed to collaborations with designers like Aimé Leon Dore and JJJJound, and to athlete signings including Shohei Ohtani and Coco Gauff. These are real factors. But they are not the structural factor. The structural factor is that New Balance can hold pricing in a tariff environment where every other major competitor is forced to raise prices or absorb margin compression. That is not a brand story. It is a cost-of-goods story. And cost-of-goods stories, over long enough time horizons, are the ones that determine market share.

The rule is always the same. Control supply. Let demand build against the wall. Harvest the spread. The rule works for diamonds and oil and handbags and monetary policy and streetwear. It works until it doesn’t, and the reasons it stops working fall into a taxonomy that is surprisingly small.

The wall cracks from outside, as it did for De Beers when laboratory synthesis offered a technological alternative to geological extraction.

The wall cracks from inside, as it does for OPEC when a member state breaks production discipline for short-term revenue.

The wall is deliberately demolished by an acquirer who doesn’t understand what they purchased, as it was for Supreme when VF Corporation scaled a brand whose entire value was its refusal to scale.

The wall is inflated and then deflated by monetary policy, as it was for the entire speculative asset landscape between 2020 and 2023.

Or the wall is broken by the monopolist itself, voluntarily, systematically, and with full access to the data that should have prevented it, as it was for Nike when the company chose to flood the very supply it had spent forty years restricting.

Nike’s version is the rarest and, for economists, the most instructive. External threats and internal cheating and acquisitive incompetence and monetary cycles are well-documented failure modes. A monopolist breaking its own scarcity with full information, and then abandoning its distribution channel simultaneously, is unusual enough to be genuinely novel as a case study. And because StockX exists, because GOAT exists, because the secondary sneaker market generates publicly visible, timestamped, authenticated transaction data at a volume and granularity that no other consumer goods category can match, the entire sequence, the flood, the premium collapse, the capital rotation, the discovery of the utility floor, and the tariff-induced repricing, is available for analysis by anyone with an internet connection and a basic understanding of supply and demand.

De Beers’ century of diamond data is proprietary. OPEC’s production figures are estimates. Hermès’ secondary market operates through private auction houses and consignment shops with no centralized price reporting. The Federal Reserve’s effects on asset prices are modeled through layers of abstraction.

Nike’s data is on StockX. Right now. Updated hourly.

Every scarcity economy runs on one rule. This is the first time in history you can watch what happens when the rule breaks, in real time, at scale, with receipts.

Please consider subscribing if you enjoyed and learned something friom this article :)

Subscribe now

Works Cited

Disco Didn't Die. Your Savings Did.

On July 12, 1979, tens of thousands of fans packed Comiskey Park in Chicago for Disco Demolition Night. Between games of a White Sox doubleheader, a local radio DJ detonated a crate of disco records on the field. The explosion tore a hole in the outfield grass, fans stormed the diamond, and the second game was forfeited. Disco, the culture declared, was dead.

Forty-seven years later, a disco song from that same summer is the hottest financial strategy on TikTok. And the story behind why is far stranger, and more revealing, than the trend itself.

The Ring My Bell Theory

In January 2026, a TikTok creator named Jennifer Marie posted a series of videos claiming that Anita Ward’s 1979 hit “Ring My Bell” could manifest wealth. The theory is rooted in numerology: certain characteristics of the song allegedly align with “coded phrases” that attract abundance. Examples include “I am wealthy,” “manifesting,” and, inexplicably, “Johnny Cash,” a man not historically known for his disco career or his financial planning.

The explanation is borderline incoherent. But the results, at least on the surface, were immediate. Marie’s audience multiplied. Followers sent her cash. She received a refund on an old electric bill. Other creators followed. A 31-year-old salon owner named Sasha Carcasses posted a clip dancing to the song that racked up more than two million views. Soon after, brands reached out about paying her for promotional partnerships for the first time. Another woman commented that after starting her day with the track, she received her biggest tip ever: $175. A 29-year-old video editor sold handmade necklaces and got commissioned to create a stop-motion animation, her “passion project,” for an upcoming event.

The song went from roughly 200,000 weekly U.S. streams to over two million in three months, a 1,000% increase according to the data company Luminate. Paris Hilton played it while ringing the closing bell at the New York Stock Exchange. Malaco Records, the small Mississippi label that acquired the rights in the 1980s, called the spike “Christmas in March.”

It is, by any measure, absurd. But dismiss it too quickly, and you miss the more interesting story underneath.

The Frequency That Actually Matters Isn’t Numerology. It’s 122 BPM.

Here’s what nobody in the manifestation discourse is talking about: “Ring My Bell” sits at approximately 122 beats per minute. That number matters more than any coded phrase.

Research published in Scientific Reports in 2025 found that music in the 85 to 120 BPM range falls within what’s called the “Medium Tempo Attraction” window, the tempo range most naturally suited to human auditory processing. The theory, first proposed by researchers Allan and Schab in the late 1980s, shows that beats spaced between 500 and 700 milliseconds apart create no perceptual bias in listeners. They simply feel right. At 122 BPM, “Ring My Bell” sits just above this window, in the zone that tips the balance from comfortable into gently energizing. It closely matches resting physiological rhythms, including breathing and heart rate under light excitement, but pushes them slightly upward.

A 2023 study in the Journal of Sports Sciences went further, measuring plasma dopamine, epinephrine, and norepinephrine concentrations before and after music exposure. Fast-tempo music (above 120 BPM) significantly increased adrenergic activation, the neurochemical cascade associated with alertness, confidence, and readiness to act, while slow-tempo music depressed it. Separate neuroscience research using PET scanning has shown that music triggers dopamine release in the striatum, the same brain region activated by food, sex, and psychoactive substances. A causal study published in Proceedings of the National Academy of Sciences went even further: when researchers gave participants a dopamine precursor (levodopa) before listening to music, musical pleasure increased. When they gave participants a dopamine blocker (risperidone), the pleasure disappeared. The brain doesn’t just enjoy music. It chemically rewards you for listening to it.

So when you play “Ring My Bell” at 7 AM on repeat, you aren’t aligning with the universe. You’re giving yourself a neurochemical boost that makes you slightly more confident, slightly more alert, and measurably more likely to notice opportunity when it shows up. The song is working. Just not for the reason TikTok thinks. The frequency isn’t some mystical vibration. It’s 122 beats per minute, and decades of peer-reviewed research can explain exactly why it makes you feel good.

This is also, incidentally, why disco as a genre keeps getting recycled for these trends. House music, disco’s direct descendant, lives between 115 and 130 BPM. It is, neurologically speaking, optimized for mood elevation. Dancing to it triggers a second layer: endorphin release compounds the dopamine hit, creating what neuroscientists describe as a dual-action pleasure response. The genre isn’t magic. It’s just well-engineered dopamine delivery. And it has been for nearly five decades.

The Manifestation-Industrial Complex

“Ring My Bell” isn’t an isolated event. It’s the latest entry in a pattern that TikTok has been running on roughly eight-to-ten-month cycles, each one following the same playbook.

In 2022, it was 417Hz frequency tones, solfeggio sounds supposedly tuned to “the vibration of financial abundance.” The hashtags accumulated millions of views. Creators posted videos of themselves listening with headphones, eyes closed, waiting for money to arrive. There was just one problem: a spiritual coach who has studied solfeggio frequencies told the Daily Dot that 417Hz is actually designated for the removal of trauma, not creating money. The frequency TikTok was using for wealth manifestation was, according to the very tradition it claimed to draw from, the wrong frequency entirely. Nobody noticed. Nobody cared. The videos kept going viral.

In early 2023, it was Lucky Girl Syndrome. The premise was simple: if you say “I am so lucky, everything always works out for me” with enough conviction, you will become lucky and everything will work out. The hashtag racked up over 700 million views. Manifestation coaches endorsed it. A TikTok user reported that after she and her husband started saying “everything works out for us,” he won a $900 sports bet that same night. Observers pointed out that Lucky Girl Syndrome was essentially a repackaging of “The Secret,” the 2006 film-turned-book about the law of attraction, which was itself a repackaging of Napoleon Hill’s “Think and Grow Rich” from 1937. Same idea, new packaging, different platform.

Now, in 2026, it’s a disco song that allegedly contains the name “Johnny Cash” as a wealth code.

Each time, the mechanic is identical. Someone posts a claim. People try it. Those who experience a coincidental win post about it. Those who don’t, and there are inevitably far more of them, just scroll past. The algorithm, optimized for engagement rather than accuracy, surfaces the positive testimonials and buries the silence. Within days, the trend looks like proof. It is, in fact, a textbook demonstration of survivorship bias operating at platform scale across two billion users. The wins are visible. The losses are invisible. And the recommendation engine ensures that the ratio between the two always appears to favor the believers.

But mocking the trend misses the point entirely. The more important question is why millions of people keep reaching for it.

Financial Nihilism and the Lottery Ticket Economy

Bloomberg has a name for what’s happening: financial nihilism. It describes a generation’s increasingly self-destructive relationship with money, one characterized by crypto bets, prediction markets, sports gambling, and retirement accounts raided to pay credit cards. The term gained traction after a Northwestern Mutual study released in March 2026 found that 80% of Gen Z respondents say they feel financially left behind. Seventy-three percent of Americans who feel that way said high-risk, speculative investments can help them reach their financial goals more effectively than traditional methods. Among Gen Z, that number climbed to 80%.

The numbers behind the feeling are stark. In 1990, the median American home cost 3.2 times the median household income. Today, according to the World Economic Forum, it’s five times. For someone aged 20 to 34, it’s closer to eight times their annual salary. The median wage for a bachelor’s degree holder, adjusted for inflation, has barely moved from $58,138 in 1990 to $60,000 today. Gen Z’s unemployment rate sits at 8.3%, double the national average. Entry-level jobs have declined by 35%. The average Gen Z adult carries $94,101 in personal debt. Only 27% own a home. The average first-time buyer is now 40 years old. And here’s the number that makes the math visceral: Gen Z’s total monthly debt payments have increased 27% while wages grew just 8%. The gap isn’t closing. It’s accelerating.

In 1989, adults under 40 held 12% of household wealth while those over 70 held 19%. Today, those under 40 command just 7% while those over 70 control 30%. Boomers hold 54% of all U.S. stocks, worth over $25 trillion. Millennials hold about 8%.

When the traditional path to wealth feels like a locked door, people start looking for side doors. And the data shows exactly which ones they’re trying.

Sports betting participation jumped from 25% to 30% of all U.S. consumers in a single year, driven almost entirely by Gen Z and millennials. Gen Z is now spending more per capita on gambling than any other generation, and 37% self-identify as addicted. Fifty-one percent own or have owned cryptocurrency. Meme coin trading volume surged 5x compared to other asset classes. A third of Gen Z investors have used prediction markets, platforms where you bet on the outcome of real-world events, blurring the line between investing and gambling entirely. Forty-six percent have already withdrawn from their retirement savings. And now, playing a disco song on repeat to attract money.

NYU professor Scott Galloway framed it bluntly at SXSW earlier this month. For the last 40 years, he argued, every genuine economic shock that threatened to destroy capital, the dot-com crash, 2008, COVID, was met with government intervention designed to protect asset owners, not earners. “Rather than let assets collapse and take money from the owners and give advantage back to the earners,” Galloway said, speaking directly to Gen Z, “we’re going to pull out your credit card and ensure, in the form of debt and stimulus, that I stay rich.” Bloomberg, surveying the Northwestern Mutual data, called Gen Z’s resulting investment behavior “financial nihilism.” The study attributed the trend to a belief that previous rules of growth and finance had broken down, creating a generation that suspected market manipulation and sought better returns in new, unregulated markets.

Economist Kyla Scanlon, herself a Gen Zer who coined the popular economic term “vibecession,” wrote in the Wall Street Journal that young adults aren’t confused by the risks they’re taking. They understand them. They’re responding to an economy where the usual advice no longer lines up with outcomes.

They’re not stupid. They’re out of options. And every ten months, TikTok gives them a new one.

This Isn’t New. It’s a Pattern as Old as Economic Downturns.

The connection between manifestation culture and economic hardship isn’t a TikTok invention. It’s one of the most reliable patterns in American cultural history.

Napoleon Hill published “Think and Grow Rich” in 1937, in the depths of the Great Depression. The book, which argued that desire, faith, and persistence could attract wealth through a vaguely defined “infinite intelligence,” sold over 15 million copies. It became one of the best-selling self-help books of all time. Rhonda Byrne released “The Secret” in 2006, right on the eve of the worst financial crisis since then. The book and film, built on the “law of attraction,” the idea that positive thoughts attract positive outcomes, became a global phenomenon. Oprah endorsed it. It sold 35 million copies.

As the Daily Dot noted as far back as 2022, when the 417Hz trend was still new, ideas about changing your life by vibrating at a higher frequency are popular during economic precarity for the same reason they were popular during the Depression: they feel personally empowering at a time when the masses are grappling with forces beyond their control. When you can’t control the economy, the housing market, or your student loan balance, you can control whether you press play on a song. And that tiny sense of agency, however illusory, is worth something to people who feel they have none.

TikTok didn’t invent manifestation culture. It just gave it an algorithm, a recommendation engine, and a two-billion-person audience primed by economic anxiety to believe.

The Real Money: Who Actually Profits From Viral Nostalgia

While millions of people play “Ring My Bell” hoping for a windfall, the actual financial winners of these viral moments are the companies that bet billions on owning old songs.

Over the last decade, the music catalog market has exploded into an institutional-grade asset class. Blackstone acquired Hipgnosis Songs Fund’s portfolio of 45,000 songs in a $1.6 billion deal in 2024, beating out a rival bid from Concord. Warner Music and Bain Capital launched a joint venture with up to $1.2 billion earmarked specifically for catalog acquisitions. Bob Dylan reportedly sold his catalog for $300 to $400 million. Bruce Springsteen’s went for a similar figure. Justin Bieber’s sold for over $200 million. Catalogs now routinely trade at 20 to 30 times their annual revenue, up from a historical range of 8 to 12 times.

The thesis behind these purchases is simple: old songs generate stable, recurring royalty income that functions like a bond with upside. Streaming guarantees a baseline. Sync licensing (placement in TV, film, and ads) adds spikes. And TikTok virality, the exact phenomenon driving the “Ring My Bell” theory, creates unpredictable but increasingly frequent jackpots. When Fleetwood Mac’s “Dreams” went viral on TikTok in 2020 after a man posted a video skateboarding and drinking cranberry juice, streams of the 1977 track surged so dramatically that Lindsay Buckingham and Mick Fleetwood both sold portions of their catalogs shortly after. The virality wasn’t just a cultural moment. It was a liquidity event.

Recognition Music, the company that now manages the old Hipgnosis catalog under Blackstone, has built an entire infrastructure around monitoring TikTok for exactly these moments. When Rihanna’s 2007 track “Breakin’ Dishes” went viral on TikTok in late 2025, Recognition’s monitoring systems caught the spike. They partnered with Roc Nation and Def Jam to re-pitch the song to streaming services, coordinated international influencer campaigns, and commissioned a dance music remix that gained traction on European radio. The song charted on the Hot 100 for the first time, eighteen years after its release. Recognition’s CEO told Billboard that in the last twelve months alone, they had released over 60 interpolations of catalog songs and generated about 500 million streams from those efforts.

Malaco Records, the small Mississippi label that owns “Ring My Bell,” doesn’t operate at that scale. But the principle is identical. They own a piece of a 47-year-old disco song that just experienced a 1,000% streaming increase because of a numerology theory on TikTok. Their VP called it “Christmas in March.” He’s not wrong. For catalog owners, every viral manifestation trend is Christmas.

This is the quietly enormous business story hidden inside the “Ring My Bell” theory: while Gen Z plays old songs hoping to manifest wealth, the companies that own those songs are the ones actually getting rich. The manifestation trend is, in a precise and literal sense, working. Just not for the people who believe in it.

The Only Frequency That Ever Paid Anyone

Meanwhile, there is exactly one person for whom “Ring My Bell” has reliably generated wealth across five decades, and she’s never posted a TikTok in her life.

Anita Ward was an unknown artist in 1979. She wanted to record a ballad. Her label handed her a steamy disco track instead, a propulsive song with a chorus that masquerades as an invitation: “You can ring my bell.” She hated it. She recorded it anyway. It went straight to No. 1 on the Hot 100 that summer, the last golden window for disco, competing for the top spot with the Bee Gees, Donna Summer, and Chic. Radio was a disco lovers’ paradise. Then it was over.

Ward never got close to another hit. She worked as a teacher. She worked in retail. Malaco Records, which specializes in soul, blues, and gospel, acquired the song in the 1980s. Ward faded from the public eye almost entirely. But she kept 40% of the publishing rights. That single decision, made 47 years ago by a young artist who didn’t even like the song she was signing away partial ownership of, is now paying off at a rate she never anticipated.

“I don’t really pay much attention to TikTok,” Ward told the Wall Street Journal. “But I’m getting royalties, and I’m not upset about that.”

She also offered a characteristically understated reflection on the song itself: “’Ring My Bell’ was a song that I thought I didn’t like. I’ve had to eat crow, because it became big all over the world.”

So while millions of people play disco on repeat, hoping it’ll pay their rent, the one person it’s actually working for is a retired woman in Memphis who secured a publishing deal in 1979. She didn’t need a frequency. She didn’t need a coded phrase. She didn’t need a numerological connection to Johnny Cash. She needed 40% of the publishing.

Disco didn’t die at Comiskey Park. It just started collecting residuals. And that’s the only frequency that’s ever paid, anyone.

SOURCES:
Disco Was Declared Dead in 1979. It Just Became Gen Z’s Financial Advisor.

On July 12, 1979, tens of thousands of fans packed Comiskey Park in Chicago for Disco Demolition Night. Between games of a White Sox doubleheader, a local radio DJ detonated a crate of disco records on the field. The explosion tore a hole in the outfield grass, fans stormed the diamond, and the second game was forfeited. Disco, the culture declared, was dead.

Forty-seven years later, a disco song from that same summer is the hottest financial strategy on TikTok. And the story behind why is far stranger, and more revealing, than the trend itself.

The Ring My Bell Theory

In January 2026, a TikTok creator named Jennifer Marie posted a series of videos claiming that Anita Ward’s 1979 hit “Ring My Bell” could manifest wealth. The theory is rooted in numerology: certain characteristics of the song allegedly align with “coded phrases” that attract abundance. Examples include “I am wealthy,” “manifesting,” and, inexplicably, “Johnny Cash,” a man not historically known for his disco career or his financial planning.

The explanation is borderline incoherent. But the results, at least on the surface, were immediate. Marie’s audience multiplied. Followers sent her cash. She received a refund on an old electric bill. Other creators followed. A 31-year-old salon owner named Sasha Carcasses posted a clip dancing to the song that racked up more than two million views. Soon after, brands reached out about paying her for promotional partnerships for the first time. Another woman commented that after starting her day with the track, she received her biggest tip ever: $175. A 29-year-old video editor sold handmade necklaces and got commissioned to create a stop-motion animation, her “passion project,” for an upcoming event.

The song went from roughly 200,000 weekly U.S. streams to over two million in three months, a 1,000% increase according to the data company Luminate. Paris Hilton played it while ringing the closing bell at the New York Stock Exchange. Malaco Records, the small Mississippi label that acquired the rights in the 1980s, called the spike “Christmas in March.”

It is, by any measure, absurd. But dismiss it too quickly and you miss the more interesting story underneath.

The Frequency That Actually Matters Isn’t Numerology. It’s 122 BPM.

Here’s what nobody in the manifestation discourse is talking about: “Ring My Bell” sits at approximately 122 beats per minute. That number matters more than any coded phrase.

Research published in Scientific Reports in 2025 found that music in the 85 to 120 BPM range falls within what’s called the “Medium Tempo Attraction” window, the tempo range most naturally suited to human auditory processing. The theory, first proposed by researchers Allan and Schab in the late 1980s, shows that beats spaced between 500 and 700 milliseconds apart create no perceptual bias in listeners. They simply feel right. At 122 BPM, “Ring My Bell” sits just above this window, in the zone that tips the balance from comfortable into gently energizing. It closely matches resting physiological rhythms, including breathing and heart rate under light excitement, but pushes them slightly upward.

A 2023 study in the Journal of Sports Sciences went further, measuring plasma dopamine, epinephrine, and norepinephrine concentrations before and after music exposure. Fast-tempo music (above 120 BPM) significantly increased adrenergic activation, the neurochemical cascade associated with alertness, confidence, and readiness to act, while slow-tempo music depressed it. Separate neuroscience research using PET scanning has shown that music triggers dopamine release in the striatum, the same brain region activated by food, sex, and psychoactive substances. A causal study published in Proceedings of the National Academy of Sciences went even further: when researchers gave participants a dopamine precursor (levodopa) before listening to music, musical pleasure increased. When they gave participants a dopamine blocker (risperidone), the pleasure disappeared. The brain doesn’t just enjoy music. It chemically rewards you for listening to it.

So when you play “Ring My Bell” at 7 AM on repeat, you aren’t aligning with the universe. You’re giving yourself a neurochemical boost that makes you slightly more confident, slightly more alert, and measurably more likely to notice opportunity when it shows up. The song is working. Just not for the reason TikTok thinks. The frequency isn’t some mystical vibration. It’s 122 beats per minute, and decades of peer-reviewed research can explain exactly why it makes you feel good.

This is also, incidentally, why disco as a genre keeps getting recycled for these trends. House music, disco’s direct descendant, lives between 115 and 130 BPM. It is, neurologically speaking, optimized for mood elevation. Dancing to it triggers a second layer: endorphin release compounds the dopamine hit, creating what neuroscientists describe as a dual-action pleasure response. The genre isn’t magic. It’s just well-engineered dopamine delivery. And it has been for nearly five decades.

The Manifestation-Industrial Complex

“Ring My Bell” isn’t an isolated event. It’s the latest entry in a pattern that TikTok has been running on roughly eight-to-ten-month cycles, each one following the same playbook.

In 2022, it was 417Hz frequency tones, solfeggio sounds supposedly tuned to “the vibration of financial abundance.” The hashtags accumulated millions of views. Creators posted videos of themselves listening with headphones, eyes closed, waiting for money to arrive. There was just one problem: a spiritual coach who has studied solfeggio frequencies told the Daily Dot that 417Hz is actually designated for the removal of trauma, not creating money. The frequency TikTok was using for wealth manifestation was, according to the very tradition it claimed to draw from, the wrong frequency entirely. Nobody noticed. Nobody cared. The videos kept going viral.

In early 2023, it was Lucky Girl Syndrome. The premise was simple: if you say “I am so lucky, everything always works out for me” with enough conviction, you will become lucky and everything will work out. The hashtag racked up over 700 million views. Manifestation coaches endorsed it. A TikTok user reported that after she and her husband started saying “everything works out for us,” he won a $900 sports bet that same night. Observers pointed out that Lucky Girl Syndrome was essentially a repackaging of “The Secret,” the 2006 film-turned-book about the law of attraction, which was itself a repackaging of Napoleon Hill’s “Think and Grow Rich” from 1937. Same idea, new packaging, different platform.

Now, in 2026, it’s a disco song that allegedly contains the name “Johnny Cash” as a wealth code.

Each time, the mechanic is identical. Someone posts a claim. People try it. Those who experience a coincidental win post about it. Those who don’t, and there are inevitably far more of them, just scroll past. The algorithm, optimized for engagement rather than accuracy, surfaces the positive testimonials and buries the silence. Within days, the trend looks like proof. It is, in fact, a textbook demonstration of survivorship bias operating at platform scale across two billion users. The wins are visible. The losses are invisible. And the recommendation engine ensures that the ratio between the two always appears to favor the believers.

But mocking the trend misses the point entirely. The more important question is why millions of people keep reaching for it.

Financial Nihilism and the Lottery Ticket Economy

Bloomberg has a name for what’s happening: financial nihilism. It describes a generation’s increasingly self-destructive relationship with money, one characterized by crypto bets, prediction markets, sports gambling, and retirement accounts raided to pay credit cards. The term gained traction after a Northwestern Mutual study released in March 2026 found that 80% of Gen Z respondents say they feel financially left behind. Seventy-three percent of Americans who feel that way said high-risk, speculative investments can help them reach their financial goals more effectively than traditional methods. Among Gen Z, that number climbed to 80%.

The numbers behind the feeling are stark. In 1990, the median American home cost 3.2 times the median household income. Today, according to the World Economic Forum, it’s five times. For someone aged 20 to 34, it’s closer to eight times their annual salary. The median wage for a bachelor’s degree holder, adjusted for inflation, has barely moved from $58,138 in 1990 to $60,000 today. Gen Z’s unemployment rate sits at 8.3%, double the national average. Entry-level jobs have declined by 35%. The average Gen Z adult carries $94,101 in personal debt. Only 27% own a home. The average first-time buyer is now 40 years old. And here’s the number that makes the math visceral: Gen Z’s total monthly debt payments have increased 27% while wages grew just 8%. The gap isn’t closing. It’s accelerating.

In 1989, adults under 40 held 12% of household wealth while those over 70 held 19%. Today, those under 40 command just 7% while those over 70 control 30%. Boomers hold 54% of all U.S. stocks, worth over $25 trillion. Millennials hold about 8%.

When the traditional path to wealth feels like a locked door, people start looking for side doors. And the data shows exactly which ones they’re trying.

Sports betting participation jumped from 25% to 30% of all U.S. consumers in a single year, driven almost entirely by Gen Z and millennials. Gen Z is now spending more per capita on gambling than any other generation, and 37% self-identify as addicted. Fifty-one percent own or have owned cryptocurrency. Meme coin trading volume surged 5x compared to other asset classes. A third of Gen Z investors have used prediction markets, platforms where you bet on the outcome of real-world events, blurring the line between investing and gambling entirely. Forty-six percent have already withdrawn from their retirement savings. And now, playing a disco song on repeat to attract money.

NYU professor Scott Galloway framed it bluntly at SXSW earlier this month. For the last 40 years, he argued, every genuine economic shock that threatened to destroy capital, the dot-com crash, 2008, COVID, was met with government intervention designed to protect asset owners, not earners. “Rather than let assets collapse and take money from the owners and give advantage back to the earners,” Galloway said, speaking directly to Gen Z, “we’re going to pull out your credit card and ensure, in the form of debt and stimulus, that I stay rich.” Bloomberg, surveying the Northwestern Mutual data, called Gen Z’s resulting investment behavior “financial nihilism.” The study attributed the trend to a belief that previous rules of growth and finance had broken down, creating a generation that suspected market manipulation and sought better returns in new, unregulated markets.

Economist Kyla Scanlon, herself a Gen Zer who coined the popular economic term “vibecession,” wrote in the Wall Street Journal that young adults aren’t confused by the risks they’re taking. They understand them. They’re responding to an economy where the usual advice no longer lines up with outcomes.

They’re not stupid. They’re out of options. And every ten months, TikTok gives them a new one.

This Isn’t New. It’s a Pattern as Old as Economic Downturns.

The connection between manifestation culture and economic hardship isn’t a TikTok invention. It’s one of the most reliable patterns in American cultural history.

Napoleon Hill published “Think and Grow Rich” in 1937, in the depths of the Great Depression. The book, which argued that desire, faith, and persistence could attract wealth through a vaguely defined “infinite intelligence,” sold over 15 million copies. It became one of the best-selling self-help books of all time. Rhonda Byrne released “The Secret” in 2006, right on the eve of the worst financial crisis since then. The book and film, built on the “law of attraction,” the idea that positive thoughts attract positive outcomes, became a global phenomenon. Oprah endorsed it. It sold 35 million copies.

As the Daily Dot noted as far back as 2022, when the 417Hz trend was still new, ideas about changing your life by vibrating at a higher frequency are popular during economic precarity for the same reason they were popular during the Depression: they feel personally empowering at a time when the masses are grappling with forces beyond their control. When you can’t control the economy, the housing market, or your student loan balance, you can control whether you press play on a song. And that tiny sense of agency, however illusory, is worth something to people who feel they have none.

TikTok didn’t invent manifestation culture. It just gave it an algorithm, a recommendation engine, and a two-billion-person audience primed by economic anxiety to believe.

The Real Money: Who Actually Profits From Viral Nostalgia

While millions of people play “Ring My Bell” hoping for a windfall, the actual financial winners of these viral moments are the companies that bet billions on owning old songs.

Over the last decade, the music catalog market has exploded into an institutional-grade asset class. Blackstone acquired Hipgnosis Songs Fund’s portfolio of 45,000 songs in a $1.6 billion deal in 2024, beating out a rival bid from Concord. Warner Music and Bain Capital launched a joint venture with up to $1.2 billion earmarked specifically for catalog acquisitions. Bob Dylan reportedly sold his catalog for $300 to $400 million. Bruce Springsteen’s went for a similar figure. Justin Bieber’s sold for over $200 million. Catalogs now routinely trade at 20 to 30 times their annual revenue, up from a historical range of 8 to 12 times.

The thesis behind these purchases is simple: old songs generate stable, recurring royalty income that functions like a bond with upside. Streaming guarantees a baseline. Sync licensing (placement in TV, film, and ads) adds spikes. And TikTok virality, the exact phenomenon driving the “Ring My Bell” theory, creates unpredictable but increasingly frequent jackpots. When Fleetwood Mac’s “Dreams” went viral on TikTok in 2020 after a man posted a video skateboarding and drinking cranberry juice, streams of the 1977 track surged so dramatically that Lindsay Buckingham and Mick Fleetwood both sold portions of their catalogs shortly after. The virality wasn’t just a cultural moment. It was a liquidity event.

Recognition Music, the company that now manages the old Hipgnosis catalog under Blackstone, has built an entire infrastructure around monitoring TikTok for exactly these moments. When Rihanna’s 2007 track “Breakin’ Dishes” went viral on TikTok in late 2025, Recognition’s monitoring systems caught the spike. They partnered with Roc Nation and Def Jam to re-pitch the song to streaming services, coordinated international influencer campaigns, and commissioned a dance music remix that gained traction on European radio. The song charted on the Hot 100 for the first time, eighteen years after its release. Recognition’s CEO told Billboard that in the last twelve months alone, they had released over 60 interpolations of catalog songs and generated about 500 million streams from those efforts.

Malaco Records, the small Mississippi label that owns “Ring My Bell,” doesn’t operate at that scale. But the principle is identical. They own a piece of a 47-year-old disco song that just experienced a 1,000% streaming increase because of a numerology theory on TikTok. Their VP called it “Christmas in March.” He’s not wrong. For catalog owners, every viral manifestation trend is Christmas.

This is the quietly enormous business story hidden inside the “Ring My Bell” theory: while Gen Z plays old songs hoping to manifest wealth, the companies that own those songs are the ones actually getting rich. The manifestation trend is, in a precise and literal sense, working. Just not for the people who believe in it.

The Only Frequency That Ever Paid Anyone

Meanwhile, there is exactly one person for whom “Ring My Bell” has reliably generated wealth across five decades, and she’s never posted a TikTok in her life.

Anita Ward was an unknown artist in 1979. She wanted to record a ballad. Her label handed her a steamy disco track instead, a propulsive song with a chorus that masquerades as an invitation: “You can ring my bell.” She hated it. She recorded it anyway. It went straight to No. 1 on the Hot 100 that summer, the last golden window for disco, competing for the top spot with the Bee Gees, Donna Summer, and Chic. Radio was a disco lovers’ paradise. Then it was over.

Ward never got close to another hit. She worked as a teacher. She worked in retail. Malaco Records, which specializes in soul, blues, and gospel, acquired the song in the 1980s. Ward faded from the public eye almost entirely. But she kept 40% of the publishing rights. That single decision, made 47 years ago by a young artist who didn’t even like the song she was signing away partial ownership of, is now paying off at a rate she never anticipated.

“I don’t really pay much attention to TikTok,” Ward told the Wall Street Journal. “But I’m getting royalties, and I’m not upset about that.”

She also offered a characteristically understated reflection on the song itself: “’Ring My Bell’ was a song that I thought I didn’t like. I’ve had to eat crow, because it became big all over the world.”

So while millions of people play disco on repeat hoping it’ll pay their rent, the one person it’s actually working for is a retired woman in Memphis who secured a publishing deal in 1979. She didn’t need a frequency. She didn’t need a coded phrase. She didn’t need a numerological connection to Johnny Cash. She needed 40% of the publishing.

Disco didn’t die at Comiskey Park. It just started collecting residuals. And that’s the only frequency that’s ever paid anyone.

Works Cited

That is, more or less, the origin story of the $2 billion insect farming bubble. Between 2014 and 2025, venture capital firms and governments (an alliance that should alarm anyone who has watched either group spend money) poured roughly $2 billion into startups promising to farm bugs at an industrial scale. The thesis had that particular Silicon Valley quality of being technically true and commercially suicidal. Insects convert feed into protein far more efficiently than cattle. They require a fraction of the land and water. They can theoretically be raised on organic waste, which allowed pitch decks to include the phrase “circular economy” on at least three slides, and if there is one thing that loosens a European government’s purse strings faster than the words “circular economy,” science has yet to discover it.

But to understand why anyone thought farming mealworms in the Netherlands would work, you have to understand the commodity they were trying to replace. And that commodity is soybeans. Specifically, soybean meal is the protein-rich byproduct of crushing soybeans for oil. It is the single most important protein ingredient in global animal feed. The market was valued at over $100 billion in 2025. It is what chickens eat, what pigs eat, and what farmed salmon eat. Without soybean meal, the modern factory farming system as we know it simply does not function.

And soybean meal has a problem. Several problems, actually, though the industry would prefer you think of them as externalities.

Start with the land. Brazil is the world’s largest soybean exporter, and the expansion of soy farming into the Cerrado, the vast tropical savanna that sits south of the Amazon, has been one of the great quiet ecological disasters of the past two decades. At least 46 percent of the Cerrado’s native vegetation has already been cleared for pasture and crops. In 2023, deforestation in the Cerrado reached 1.1 million hectares, more than twice that of the Amazon in the same year. The Matopiba region, which is the agricultural frontier where soy is pushing deepest into uncultivated land, saw soybean area increase by 253 percent between 2000 and 2014, at the cost of half the region’s native vegetation. The Soy Moratorium, a voluntary private-sector agreement from 2006 that prevented traders from purchasing soybeans grown on recently deforested Amazon land, did manage to reduce deforestation-linked soy in the Amazon from 30 percent of expansion to less than 1 percent. But in August 2025, a Brazilian court suspended the moratorium after a soy industry lobby group argued it was costing the state of Mato Grosso $3.7 billion a year. Environmental groups warned that the suspension could reopen the Amazon to large-scale soy-driven deforestation. Whether it will is still playing out. But the direction of travel is not encouraging.

Then there is the geopolitics. China accounts for roughly 60 percent of global soybean imports, and the U.S.-China trade wars have repeatedly reshuffled the entire market. After the 2018 trade frictions, China redirected purchases toward Brazil, which went from 53 percent of Chinese soy imports in 2017 to 78 percent by 2019. In 2025, renewed tariff escalation drove U.S. soybean exports to China effectively to zero for several months. The USDA projects that U.S. exports to China could fall to just 15 percent of pre-war levels. Meanwhile, the EU slapped a 25 percent retaliatory tariff on U.S. soybeans in late 2025, further scrambling trade flows. Soybean meal prices have been volatile, hitting $450 per ton in 2024, up 40 percent from $320 in 2019. Floods in Brazil’s Rio Grande do Sul cut the 2024 crop by nearly 3 million metric tons. Port congestion in Santos, the world’s largest grain hub, has pushed vessel wait times to 15 days during harvest season, each idle day costing traders thousands in fees that get passed back to farmers. The whole system, in other words, is enormous, politically fraught, ecologically destructive, and supply-chain fragile.

And it was this system that the insect farming startups proposed to disrupt.

The United Nations had published a landmark 2013 report arguing that edible insects could help feed a growing global population. A Dutch entomologist went on TED (the platform where ideas go to become LinkedIn posts) and showed slides of mealworm stir-fry with a chocolate cricket dessert, as if he were unveiling the iPhone. The money moved fast. Annual investment climbed from $19 million in 2014 to over $500 million by 2022. France alone absorbed $1.3 billion, roughly 65 percent of the global total. This makes a kind of poetic sense. The French have spent centuries insisting that what you eat is a matter of civilisational importance. That they would attempt to elevate larvae to the level of haute cuisine was less a surprise than an inevitability.

The thesis was not wrong, exactly. It was merely applied to the wrong market at the wrong scale with the wrong capital structure in the wrong climate. Other than that, it was flawless.

I keep thinking of this as a case study in something specific: how a genuine biological insight gets transmuted, through the dark alchemy of venture capital, into one of the most expensive category errors in agricultural history. The reason it matters beyond the entomology is that the same pattern is playing out right now across lab-grown meat, vertical farming, and several other sectors, where investors have confused disruption (an outcome) with a business model (which it is not). The insect economy is not just a story about bugs. It is a story about what happens when the financial system that funds new food technologies cannot understand how food technologies actually scale, and is too well-capitalised to learn from its mistakes before inflicting them on everyone involved.

The collapse felt like a punchline the audience saw coming, but the comedian did not. Investment cratered 88 percent from the 2022 peak. The entire sector raised just $60 million through the first nine months of 2025. For context, that is less than what a mid-tier SaaS company might pull in over a Tuesday afternoon Series B. Of the twenty largest insect farming companies, almost a quarter ceased operations. The shuttered firms account for nearly half of all capital ever deployed into the industry. That statistic should, in a just world, be tattooed on the forearm of every partner at every climate-tech fund that participated. It will not be. Venture capital has a convenient institutional amnesia about the sectors it immolates. The learning curve resets with every fund cycle. The LPs never find out. Everyone moves on to fusion energy.

The most spectacular wreckage belongs to Ÿnsect, the French mealworm producer that became the poster child of insect protein. The company (whose name includes a Y-umlaut, because nothing communicates “approachable food product” quite like a diacritical mark) raised over €600 million. A full third of the sector’s total funding. A quarter of that came directly from the French government. In 2023, Ÿnsect recorded revenues of €656,000 against losses of €80 million. Sit with that for a second. For every euro the company earned, it burned more than a hundred. There are cryptocurrency projects with more disciplined unit economics. The company entered judicial liquidation in December 2025. Its flagship facility near the town of Dole, which was supposed to be the largest insect farm in the world, a cathedral of larvae, will close. A smaller pilot site will reportedly pivot to producing fertiliser from insect by-products. That is roughly the corporate equivalent of a Michelin-starred restaurant that failed at fine dining reopening as a compost bin.

Ÿnsect is the tallest headstone, but the graveyard is crowded. Agronutris, another French producer, entered safeguard proceedings with $100 million invested. Scandinavia’s Enorm Biofactory burned through $55 million and is reporting severe losses. Beta Hatch, once North America’s largest mealworm facility, opened in Washington state in summer 2022 and auctioned its equipment by autumn 2023. It operated for roughly fourteen months. Innovafeed, the sector’s largest surviving company, opened a USDA-funded pilot plant in Decatur, Illinois, in partnership with ADM, then suspended operations eighteen months later, citing “funding challenges.” That is corporate for “we ran out of money but would prefer you didn’t phrase it that way.” The USDA awarded the project $11.7 million in public funds, and when a nonprofit filed records requests for details, roughly half of the 600 pages were redacted. The nonprofit is now suing. Democracy works in mysterious ways.

And then there is the Tyson detail, which I think captures the whole trajectory better than any chart could. Tyson Foods, the largest meat company in America, a corporation that knows more about turning animals into products than perhaps any entity in human history, quietly withdrew its air permit application in late 2024 for a Nebraska insect farm that was supposed to process 300 billion larvae annually in partnership with Dutch producer Protix. The project is, according to internal emails obtained through public records requests, “on hold indefinitely.” When Tyson Foods looks at the economics of farming an animal and then walks away, you are not dealing with a market-timing problem. You are dealing with a market that does not exist.

And here is where the soybean story becomes essential to understanding what went wrong. Because the numbers are not subtle.

Soybean meal costs roughly $450 to $500 per ton. Even with the price volatility of recent years, even with Brazilian floods, Chinese tariffs, and port congestion pushing costs around, it remains one of the cheapest high-quality protein sources on the planet. That is what a century of optimisation, massive scale, and a global supply chain involving three continents and hundreds of millions of acres of farmland gets you. Fishmeal, the other feed ingredient that insect startups targeted, runs about $1,800 per ton. Insect meal? Between $3,800 and $6,000. The startups were walking into a market where buyers switch suppliers over a $ 10-per-ton difference and asking them to pay three to ten times more. That is approximately as viable as pitching artisanal spring water to a municipal utility and acting confused when they say no.

The soybean meal market is valued at $100 billion annually. It is projected to reach $140 billion by 2033. The U.S. alone crushed a record 2.58 billion bushels of soybeans in the 2025/26 marketing year, with new plants opening in Nebraska and South Dakota to feed the biofuel boom. Brazil exported a record 23.3 million metric tons of soybean meal in 2025, with half of it going to the EU. That volume has surged 70 percent over the past decade. Argentina’s crop rebounded 37 percent in 2024 after recovering from drought. You are not going to outcompete this system with a warehouse full of mealworms, no matter how compelling the TED talk was. The soybean supply chain has problems. It has enormous, well-documented environmental and geopolitical problems. But being expensive is not one of them.

And the insect startups made the cost problem worse through their own supply chain choices. What do you feed the insects? Agricultural co-products. Wheat bran, corn gluten, oat husks. The same ingredients are already used in conventional livestock feed. Insect farmers found themselves in direct competition with Tyson and Cargill for their raw materials, purchasing chicken feed to feed bugs, grinding those bugs into an expensive meal, and then attempting to sell it back to the chicken industry at a multiple that would make a luxury-goods executive blush. One startup founder, displaying a candour that his investors probably found less charming than I do, told the trade press: “It doesn’t really make sense to buy chicken feed to feed insects to feed to chicken.” No. No, it does not. This is not a supply chain. It is a Rube Goldberg machine with a sustainability label, and someone somewhere received a government grant to build it.

The marketing around this was particularly galling. Protix, one of the largest Dutch insect producers, describes its process on its website as turning “organic waste from the industry” into insect feed, calling it “circular food production” that “mirrors nature’s circle of life.” In practice, Protix feeds its insects oat husks and starch, ingredients with an established secondary market in conventional livestock feed. There was nothing circular about it. The insects were eating the same supply chain they were supposed to be replacing.

The geography made it worse, which by this point you may have assumed was impossible. Most of the capital went to Northern Europe (the Netherlands, France, Scandinavia) where insects, being cold-blooded, needed their barns heated to 25 to 30 degrees Celsius through winter. European energy prices roughly doubled between 2021 and 2024. So picture the full absurdity: you are heating a warehouse full of larvae in Rotterdam in January to the temperature of a Southeast Asian jungle, using electricity that costs four times what it costs in the actual Southeast Asian jungle, to produce a product that is ten times more expensive than soybean meal, which Brazil is currently exporting at record volumes and at prices so low they are crushing margins for everyone in the chain. Someone greenlit this in a board meeting. Multiple someones. They brought slides.

A conventional post-mortem would pivot here to the lesson (hubris, exuberance, bubble) and move on. But the more interesting story is what is growing out of the wreckage. Because the biological thesis was always sound. The insects did what insects do: convert waste into protein with a quiet, mindless efficiency that borders on the elegant. It was everything built on top of them that failed.

Production costs for black soldier fly larvae fell 40 percent between 2020 and 2024. From roughly $2,200 per ton down to about $1,400. The number of commercial insect farming facilities worldwide went from 50 to over 300. Feed conversion ratios are extraordinary: roughly 1.7 kilograms of organic waste produces one kilogram of protein, versus eight to ten kilograms for beef. At current cost trajectories, black soldier fly meal is approaching parity with premium-grade fishmeal, which has itself been climbing as overfishing pressure outstrips the planet’s roughly 4.5 million tons of annual fishmeal production. The biology was never the problem. The venture capital model was. Raise enormous sums. Build a $200 million mega-facility in a cold climate. Try to compete head-on with the $100 billion soybean meal complex before proving unit economics. That playbook works in software, where marginal costs approach zero and network effects reward speed. Applied without modification to a sector where marginal costs are stubbornly, irreducibly physical, and where the incumbent has a hundred-year head start and the backing of three of the largest agricultural economies on earth, it is a recipe for burning cash in a heated warehouse. The insect economy did not need disruptive capital. It needed patient capital. It needed, as the CEO of Protix told the trade press in mid-2024, to stop being “pushed to do things too quickly.”

The survivors have absorbed this with the clarity only financial near-death can provide. The model they are building looks nothing like what came before. The pivot is blindingly obvious in hindsight: stop competing with soy.

The EU Deforestation Regulation, which was meant to require deforestation-free provenance for all soybean imports, has been delayed twice and is now not expected to take full effect until late 2026 at the earliest. When it does arrive, it will add an estimated $2 to $5 per metric ton in compliance costs to certified cargoes, and Bunge has already had to geolocate more than 16,000 South American farms and invest in blockchain-enabled batch tracking just to ship 185,000 tons of compliant meal to Thailand. The Cerrado keeps losing native vegetation. The trade war keeps reshuffling who buys from whom. Port infrastructure in Brazil, despite decades of investment, remains 40 percent higher than in the U.S. Gulf, largely due to logistics gaps. Soybean meal’s price may be low, but its true cost, if you account for deforestation, geopolitical fragility, and supply chain risk, is significantly higher than $500 a ton. The problem is that nobody in the feed market is actually paying that true cost. Feed buyers buy on price. And insect meal, at $3,800 to $6,000, was never going to win that argument.

So the surviving insect companies stopped having it. They went to markets where premium pricing is not a death sentence.

The insect-based pet food market hit approximately $1.6 billion in 2025. It is growing at about 10 percent a year. Eighty-three percent of American pet owners say they would feed their animals insect-based protein. The ick factor that proved fatal for human consumption (and it is worth pausing here to note that only about 5 percent of the sector’s funding ever actually targeted human food, a statistic that quietly demolishes the entire “elites want you to eat bugs” narrative that powered a thousand YouTube conspiracies) turns out to not exist when the product is kibble going into a bowl on the floor. Nobody interrogates their Labrador’s protein source with the same existential suspicion they bring to their own plate. Insect protein is hypoallergenic, which makes it genuinely useful for the growing number of pets with sensitivities to chicken and beef. That is a real product advantage, not a sustainability story bolted onto an inferior offering. Nestlé Purina launched an insect protein line in 2023. The customer pays a premium willingly. The unit economics work. An actual business. The previous generation of insect startups treated that concept as largely theoretical.

But the real unlock goes further than pet food. The survivors figured out something the first wave missed entirely: the insect is not a single product. It is a biorefinery. One batch of black soldier fly larvae yields protein meal for pet food and aquaculture. It also yields lipids for cosmetics, animal feed, and biodiesel. Chitin and chitosan for biomedical sutures, wound healing, drug delivery, and agricultural pest defence. And frass, the insect compost, which works as a premium organic fertiliser that triggers plant immune responses through its chitin content. Researchers in Singapore are extracting melanin from black soldier flies for use in electronics. That is the kind of sentence that would have sounded unhinged five years ago and now merely sounds like Tuesday in the bioeconomy. The value was never in the protein alone. It was in the full extraction. The first generation of startups never bothered to look.

There is an irony here worth sitting with. The insect industry’s original pitch was that bugs would save the world from soy’s environmental destruction. They would replace soybean meal in livestock feed, break the link between animal agriculture and Cerrado deforestation, and reduce the carbon footprint of the global feed supply chain. It was a grand narrative, and it attracted $2 billion. The reality is that insects cannot compete with soy on price and probably never will, not for bulk commodity feed. But they can compete in premium niches where soy’s low cost is irrelevant. Pet food. Aquaculture specialty feeds. Biomedical materials. And through the frass, they can produce a fertiliser that competes not with soy directly, but with the synthetic nitrogen fertilisers that soybean farms themselves depend on. The circular economy the pitch decks promised is not the one that is actually emerging. It is stranger, smaller, and more interesting than anyone anticipated.

Stack four or five revenue streams from a single production run. Then consider that operators are often paid a gate fee on the input side (food processors paying to have organic waste hauled away). The economics flip completely. The operational model is shifting, too. Instead of $200 million mega-factories in Rotterdam, companies are going modular and moving to warm climates. Nasekomo, a Bulgarian startup, is licensing turnkey insect farms co-located with food processors. FreezeM, an Israeli company, figured out how to pause the life cycle of black soldier fly neonates and ship them globally as biological seed stock. Effectively, insect genetics as a service. That follows the same disaggregation that poultry farming traced decades ago, when breeding, hatching, growing, and processing each became separate specialised businesses. Entobel opened a facility in Vietnam doing 10,000 tons of protein meal a year. Temperature management there consists of opening a window.

None of this will produce the hundred-billion-dollar commodity revolution that the pitch decks promised. What it will produce, if the surviving operators can execute with even modest discipline, is a $5 to $10 billion niche ecosystem. Real, growing, and appropriately modest in the way that actual agricultural businesses tend to be before venture capital gets involved and inflates them into something they were never meant to become.

And this, I think, is the real thesis. It extends well beyond bugs. The pattern keeps repeating across agricultural technology with a regularity that would be comic if it were not so expensive. Vertical farming followed the same arc. Lab-grown meat is tracing its now, its early exuberance smashing into the stubborn physics of bioreactor costs. Every time, a genuine technology with real merit attracts capital structured for software returns on software timelines.

The core problem is not that these technologies fail. They work fine. The insects convert. The vertical farms grow lettuce. The bioreactors produce meat cells. The problem is that venture capital, as a financial instrument, cannot tell the difference between a technology that will replace a commodity and a technology that will create a specialty. These require different capital, different patience, and a completely different definition of what success looks like. And VC, by its structure, funds the first and destroys the second, every time, with the mechanical reliability of a fate that has read the term sheet and concluded that everyone involved deserves what is about to happen.

The soybean complex will remain the dominant protein source in global animal feed for the foreseeable future. It will continue to expand in the Cerrado. It will keep getting reshuffled by trade wars. Brazilian ports will keep bottlenecking during the harvest season. The EU will keep postponing its deforestation regulation. The system is enormous, flawed, and entrenched. It is not going to be disrupted by a startup, no matter how many slides that startup has about feed conversion ratios.

But around the margins of that system, in warm climates, premium niches, and modular facilities co-located with food waste, the insects are doing fine. They always were. Two billion dollars bought us that particular lesson. The insects converted their feed beautifully. They grew fast. They produced protein, fat, chitin, and fertiliser from waste. They did exactly what the biology said they would, because insects have been doing this for roughly 400 million years and are not easily deterred by the opinions of a venture partner in San Francisco who believed he had found a better way. It was the spreadsheets that were delusional. And the pitch decks. And the term sheets. And the entire apparatus of capital that looked at one of the planet’s most ancient food sources and decided what it really needed was a Series C, a mega-factory in the Netherlands, and a Y-umlaut. The insects were fine. They were always fine. It was everything around them that failed.

House music just had the biggest year any genre has had in a decade. According to the Splice x MIDiA Research “Sounds of 2026” report, house climbed from the fifth most downloaded genre among producers in 2023 to second in 2025, driven by a 778% explosion in Afro house alone. Fred again.. is selling out stadiums. Peggy Gou is a global fashion icon who happens to DJ. John Summit turned tech house into a lifestyle brand. Beyoncé made an entire album about it. Drake tried. The culture is no longer underground. It’s in the group chat.

But here’s what almost nobody in this massive new audience has thought about: every time they walk into a club, there’s a setup behind the DJ booth that costs more than most people’s car, and there is essentially one company on earth that makes it. The new fans don’t know this. Most casual clubgoers don’t know this. And the economics of it affect everything from how much your cover charge costs to whether your favorite venue survives another year.

This is the story of AlphaTheta Corporation, formerly Pioneer DJ, and how one company quietly built itself into the invisible infrastructure of an entire global culture.

The numbers behind the booth

Let’s start with the raw market position, because it’s startling once you see it laid out.

AlphaTheta, operating primarily under the Pioneer DJ brand, controls approximately 70% of the global DJ equipment market. That figure comes from multiple industry reports and has been cited in regulatory filings. But even that number understates the situation. In professional club installations specifically, the venues where touring DJs perform, the places that define nightlife culture in cities like Berlin, Ibiza, Tokyo, New York, and London, Pioneer’s market share approaches 100%.

Walk into any professional DJ booth on earth and you’ll find the same setup: two to four CDJ-3000 media players and a DJM-A9 mixer, all bearing the Pioneer DJ logo. A full rig, four players and a mixer, runs well into five figures. Multiply that across thousands of clubs worldwide, and you begin to see the scale of the infrastructure investment that underpins this single-brand dominance.

On the software side, the picture is equally concentrated. The Digital DJ Tips Global DJ Census, the largest annual survey of DJs in the world, with over 22,000 respondents in 2026, shows Rekordbox (Pioneer’s proprietary software) as the most-used DJ platform at 34%, with Serato at 27% and VirtualDJ at roughly 17%. As recently as 2019, Serato held the top spot with over 30%. Rekordbox has steadily pulled ahead, and the gap is widening. The same company that dominates the hardware now dominates the software too.

If we combine Rekordbox and Serato’s market positions, those two platforms together account for roughly 90% of professional DJ software usage. This matters enormously because of what almost happened in 2023, but we’ll get to that.

How one company locked down an entire culture

Pioneer DJ’s dominance didn’t happen overnight, and it didn’t happen through villainy. It happened through good timing, smart product design, and a self-reinforcing economic loop that, once established, became nearly impossible to break.

The origin story starts in 2001. Pioneer released the CDJ-1000, a digital media player with a touch-sensitive platter that mimicked the feel of vinyl, hot cue buttons, and a layout deliberately similar to the Technics SL-1200 turntables that were already the club standard. It wasn’t the first CD player designed for DJs, but it was the first one that felt right to DJs who had grown up on vinyl. The transition from physical records to digital was already inevitable; Pioneer just built the bridge that most DJs were willing to walk across.

Crucially, Pioneer also did something that competitors either couldn’t or didn’t: they got their gear into key clubs early. By placing equipment in influential venues and building relationships with major promoters and DJs, Pioneer established a beachhead that created a cascade effect.

Once a critical mass of clubs had Pioneer gear installed, the self-reinforcing loop began. It works like this, and understanding it is the key to understanding everything about this market:

Step one: Venues buy Pioneer gear. A professional DJ setup costs five figures. Venues don’t make this purchase lightly, and they don’t make it often. Once installed, this equipment represents a sunk cost that the venue needs to amortize over years of use.

Step two: DJs learn Pioneer gear. Because that’s what’s in the clubs. If you’re an aspiring DJ with any ambition to play professionally, you learn Rekordbox. You practice on CDJ layouts. You build your entire music library, thousands of hours of cue points, beat grids, playlists, and analysis, inside Pioneer’s ecosystem.

Step three: Headliners demand Pioneer on their technical riders. A headlining DJ’s performance contract includes a “technical rider”, a legally binding document that specifies the exact equipment required for their show. For decades, that rider has almost universally specified Pioneer CDJs and a DJM mixer. This isn’t a preference; it’s a contractual requirement.

Step four: The cycle completes. Venues buy Pioneer gear to attract headlining DJs. New DJs learn Pioneer because that’s what’s in the clubs. When those new DJs become headliners, they put Pioneer on their own riders. The loop perpetuates itself across generations of DJs and venue ownership changes.

Step five: Competitors can’t break in. Even if a competitor builds a technically superior product at a lower price, and several have, they can’t overcome the network effect. Clubs won’t install non-Pioneer gear because DJs don’t request it. DJs won’t request non-Pioneer gear because clubs don’t have it.

This is what economists call a natural monopoly reinforced by network effects. It’s the same structural dynamic that made Windows the default operating system, or that keeps QWERTY keyboards on every laptop despite more efficient layouts existing for decades. The product doesn’t have to be the best. It just has to be ubiquitous enough that switching costs exceed the benefits of any alternative.

The Rekordbox lock-in

The hardware loop is powerful on its own, but the real lock-in is software.

Every professional DJ builds their entire library inside Rekordbox. This isn’t just a playlist, it’s thousands of hours of meticulous work. Every track is analyzed for BPM and key. Cue points are set at specific moments. Hot cues mark build-ups, drops, and vocal entries. Beat grids are adjusted for tracks with tempo variations. Playlists are organized by genre, energy level, time slot, and event type.

All of this, a DJ’s entire professional infrastructure, their career in a database, gets exported to a USB stick. That USB stick is optimized to work with Pioneer hardware. Technically, you can convert libraries between platforms using third-party tools like Lexicon, DJCU, or Mixo, but the process is imperfect, time-consuming, and risky. For a touring DJ who might play a different city every night, even a small risk of library corruption or lost cue points is unacceptable.

This is where the comparison to other technology monopolies becomes most apt. Microsoft didn’t just sell operating systems; it created an ecosystem of file formats, enterprise tools, and developer dependencies that made switching to alternatives prohibitively expensive even when those alternatives were technically superior. Pioneer has done the same thing with Rekordbox. Your career lives inside their ecosystem, and the cost of leaving is measured not in dollars but in years of work you’d have to recreate.

The result is a chicken-and-egg problem that has trapped every competitor that’s tried to challenge Pioneer’s position. Denon DJ, owned by InMusic, has consistently produced hardware with superior specifications, larger touchscreens, built-in Wi-Fi, dual-layer playback from a single deck, and often at a lower price point. Industry observers and DJ forums routinely acknowledge that Denon’s feature set outpaces Pioneer’s. But features don’t matter if no club has your gear and no DJ can risk showing up to a gig without certainty that the equipment will work with their library.

The Serato saga: when regulators noticed

If there was any doubt about whether AlphaTheta’s market position constitutes a meaningful concentration of power, the Serato acquisition saga put it to rest. Not because the company did anything wrong, but because multiple sovereign governments independently concluded that letting it go further would harm competition.

In July 2023, AlphaTheta announced its intention to acquire Serato Audio Research, the New Zealand-based company behind Serato DJ, the second most popular DJ software platform in the world. The deal was valued at approximately $59 million. If completed, AlphaTheta would have owned both Rekordbox and Serato, giving it control of roughly 90% of the professional DJ software market in addition to its existing hardware dominance.

The response was swift and consequential.

InMusic, the parent company of Denon DJ, Rane, Numark, and other DJ brands, launched a public campaign against the acquisition. InMusic CEO Jack O’Donnell articulated the competitive threat bluntly: when InMusic works with Serato, they share product designs up to a year in advance so Serato can build software integration. Handing that information pipeline to AlphaTheta would mean handing it directly to their dominant competitor.

New Zealand’s Commerce Commission declined to grant clearance for the deal in July 2024, the first merger clearance application the Commission had rejected in over six years. The Commission’s chairman stated that the evidence indicated Serato and Rekordbox compete closely in the DJ software market, and that remaining competitors would not be sufficient to replace the lost competition. More pointedly, the Commission expressed concern that the merger could give AlphaTheta the means to eliminate or worsen rival hardware makers’ ability to integrate with Serato.

Simultaneously, the UK’s Competition and Markets Authority launched its own investigation, finding that the deal could create a supplier nearly double the size of its nearest rival and potentially impact prices, quality, and innovation. The CMA’s probe was ultimately cancelled after the deal was abandoned.

The deal was dead.

But here’s the epilogue that makes the story even more interesting. Serato’s founders still wanted to sell, the desire for a financial exit was real regardless of who the buyer was. In April 2025, Canadian investment firm Tiny purchased a 66% stake in Serato at a valuation of $175 million, nearly three times what AlphaTheta had offered. Serato’s value had nearly tripled in two years, partly because the regulatory fight had demonstrated just how strategically important independent DJ software is to the competitive health of the entire market.

A few months later, AlphaTheta successfully purchased Serato.

The paradox: demand explodes, venues collapse

Here is where the story moves from inside-baseball industry dynamics to something that affects anyone who’s ever gone out to hear music.

The nightlife economy is in a paradox. Cultural demand for dance music is exploding, house is the fastest-growing genre among producers, festival attendance is surging, and electronic music is more mainstream than it’s been since the late 1990s. But the venues that deliver this culture to audiences are under extraordinary pressure.

According to NIQ’s Night Time Economy Market Monitor, the UK’s late-night economy has declined 28% since 2020. An estimated 74,000 nightlife jobs were lost between early 2024 and autumn 2025. The Night Time Industries Association reports that while consumer spending in the broader night-time cultural economy grew to £153 billion, businesses face shrinking margins due to rising operational costs and inflation. Similar patterns are playing out globally. The US bars and nightclubs industry, valued at roughly $39 billion, saw revenue decline by 0.5% in 2026 even as the broader market is projected for long-term growth.

Venues are being squeezed on every margin simultaneously: rent, insurance, staffing, licensing fees, energy costs. And buried inside that cost stack is a non-negotiable line item: the DJ equipment that allows a venue to function as a club.

A full Pioneer DJ rig, the minimum equipment required to attract touring DJs and be taken seriously as a professional venue, costs upward of $10,000 and often significantly more. This isn’t a choice driven by product comparison. It’s a tax on participation in nightlife culture, imposed by the network effect rather than by any deliberate pricing strategy. A venue owner can’t choose to save money by installing Denon gear, because doing so would mean headlining DJs either won’t play there or will demand the venue rent Pioneer equipment for their specific show, adding cost and logistical complexity.

For large, established venues in major cities, this cost is manageable. For the small clubs, independent venues, and emerging-market operators that are the lifeblood of music culture, the places where genres are born, where new DJs develop, where communities form, the equipment cost is a meaningful barrier to entry that compounds every other financial pressure they face.

This isn’t Pioneer’s fault in any moral sense. The company built a good product, earned its position through legitimate competitive success, and maintains its standard through reliability and consistency that venues genuinely value. But the structural outcome is that a single company’s pricing decisions ripple through the entire global nightlife economy in ways that neither the company nor its customers fully control.

What could change

If this is a problem, and reasonable people can disagree about whether it is, what would a solution look like? The answer isn’t obvious, because this isn’t a traditional monopoly that can be resolved through a simple antitrust action. AlphaTheta hasn’t engaged in predatory pricing, exclusive dealing arrangements, or the kinds of anticompetitive behavior that typically trigger regulatory intervention. The dominance is structural, born from network effects in a market that may simply be too small to support multiple competitors at the professional tier.

But there are structural interventions that could change the dynamics without requiring anyone to be the villain.

Interoperability standards for DJ libraries. The single most impactful change would be an open standard for DJ library formats, cue points, beat grids, playlists, and analysis data, that works across all hardware and software platforms. This is analogous to what the EU achieved by mandating USB-C as a universal charging standard. If a DJ’s Rekordbox library worked seamlessly on Denon hardware, or if Serato playlists transferred perfectly to Engine DJ, the software lock-in that reinforces hardware dominance would weaken substantially. Third-party tools like Lexicon and DJCU already attempt this conversion, but they’re imperfect workarounds rather than systemic solutions. An industry-backed or regulator-encouraged open standard would fundamentally change the competitive landscape.

Rider culture disruption. The technical rider is the contractual mechanism that perpetuates the hardware loop. If major artists, agencies, or festival promoters began specifying “professional-grade media players” rather than specific Pioneer models, the cascading effect would be significant. This is a cultural change rather than a policy one, and it would need to start with high-profile DJs willing to publicly champion hardware-agnostic performance. A handful of major artists normalizing Denon or other alternatives on festival mainstages could shift norms faster than any regulation.

Equipment-as-a-service models. The current model requires venues to make large upfront capital investments in equipment that locks them into a single ecosystem for years. An equipment leasing or subscription model, potentially operated by a third party rather than by equipment manufacturers, could lower switching costs and allow venues to rotate between brands based on evolving needs. This wouldn’t eliminate the network effect entirely, but it would reduce the sunk-cost dynamic that prevents venues from experimenting with alternatives.

Software independence. The Serato/Tiny deal is actually a positive development from a competition standpoint. Keeping the second-largest DJ software platform independent from the dominant hardware manufacturer preserves a competitive pressure point. If Serato continues to innovate independently and maintains broad hardware compatibility, it serves as a counterweight to Rekordbox’s growing dominance.

Why this matters beyond the booth

The AlphaTheta/Pioneer story isn’t just about DJ equipment. It’s a case study in how network effects can create structural monopolies in small, specialized markets, and how those monopolies can have outsized impacts on the cultures that depend on them.

The nightlife economy isn’t a niche curiosity. In the UK alone, the night-time cultural economy generates £43 billion in annual spending and employs over 2 million people. Globally, the bars and nightclubs market is projected to reach well over $100 billion within the next decade. Music tourism, festivals, destination clubbing, music-themed travel, is growing at nearly 7% annually. These are real economic sectors with real jobs, real businesses, and real communities.

When one company controls the standard tooling for an entire performance industry, it shapes the economics of that industry in ways that extend far beyond the price of a media player. It influences who can afford to open a venue. It determines what the barrier to entry is for aspiring DJs. It affects how much margin venues have to work with. It contributes, however indirectly, to the cost structures that determine whether your favorite club survives or closes.

None of this makes AlphaTheta a bad company. They built a reliable product, they earned their position, and many DJs genuinely prefer their equipment. The CDJ layout’s consistency across generations, you can step up to a brand-new DJM-A9 with muscle memory from a DJM-600 made two decades ago, is a legitimate engineering achievement that serves touring professionals well.

But “not bad” isn’t the same as “not worth examining.” The millions of people who just discovered they love house music, who are packing into clubs and festivals in numbers not seen in a generation, have a stake in the health of the infrastructure that delivers that experience to them. And right now, that infrastructure runs through one company most of them have never heard of, a company that recently renamed itself after brain waves.

Whether that’s a problem worth solving, or just how small markets work, is the question. But at minimum, it’s a question worth asking out loud. Because the music is too good, and the moment is too big, for the plumbing to be invisible.

Subscribe now

Grace Hopper wanted to put programmers out of a job.

That’s documented intent. When the original COBOL specification was drafted in 1959, the explicit goal was a language so readable, so close to plain English, that business executives could write their own programs.

It did not go that way.

What COBOL actually produced, over six decades of compounding institutional dependency, is the most consequential priesthood in the history of modern finance: a small, aging, irreplaceable class of specialists who are the only living interface between trillion-dollar institutions and the logic those institutions run on. There are about 24,000 of them left in the United States. Their average age is 55. Most of them have already been offered the retirement package once.

Hopper would either find this hilarious or deeply depressing. Probably both, in the way that only people who’ve watched their life’s work metastasize into something unrecognizable can fully appreciate.

Here’s what the conventional telling of the COBOL story gets wrong: it treats this as a technology problem. Old language, new world, needs updating. File under “legacy debt,” assign to IT, follow up in Q3.

That framing is wrong in a way that actually matters. This is not a technology story. It’s a story about knowledge, specifically, about what happens when an entire civilization encodes its institutional knowledge into a system and then forgets how to read it.

The system keeps running. That’s the part that makes this so easy to ignore. Your ATM works. Your paycheck clears. The Social Security payment arrives. COBOL does its job with a reliability that genuinely embarrasses modern software, because it was built on a technical foundation that modern languages quietly abandoned for the sake of speed. Fixed-point arithmetic doesn’t round. Floating-point, what Python, Java, C, and every other contemporary language use, is fast and flexible and introduces tiny errors that compound in ways that don’t matter at all until they matter enormously. In banking, an accumulated rounding error isn’t a curiosity. It’s a federal compliance violation. So the language that never rounds, never left.

Three trillion dollars a day. 95% of ATM transactions. The IRS. Social Security. Veterans Affairs. 43% of the core banking systems in the United States. The back-end of 40% of online banking. More than 220 billion lines of it in active production right now, being maintained by that shrinking priesthood, and here’s the number nobody puts in the headline: 1.5 billion new lines being written every year. Not maintained. Written. New. Because the institutions that depend on it have no credible exit, and they know it, and so they keep building deeper into the hole.

The design had a flaw in it from the beginning. Jean Sammet, one of COBOL’s original architects, later admitted it herself: no parameterization. The modules that made up a COBOL program couldn’t pass data to each other in any clean way they had to share global state, a single pool of variables that everything touched and nothing fully owned. Change something in one module and you could corrupt data in a program three systems away. Silently. Without any error. You’d find out weeks later when a number was wrong somewhere it shouldn’t be, or a payment didn’t process, or someone’s benefits stopped arriving.

Layer on top of that the GO TO statement, an unconditional branch that could fire execution from any line of code to any other line, anywhere in the program, and you end up with what developers call spaghetti code. Logic that winds through a codebase like a fire drill through a skyscraper, touching everything, following rules no one fully wrote down, executing paths that only reveal themselves when something goes wrong in production.

Edsger Dijkstra, one of the most important computer scientists who ever lived, said that “the use of COBOL cripples the mind” and that teaching it should be considered a criminal offense. Which is a very strong thing to put in an academic paper and also possibly correct. The GO TO statement specifically drove him to write a famous letter arguing that structured programming, the approach every modern language now uses, was the only sane way to write code. COBOL was, in part, the thing he was arguing against.

He wasn’t wrong about the mechanism. He was just too late. By the time the argument was settled, the code was already everywhere.

Now take that structurally compromised, globally-stateful, GO TO-laced codebase and patch it continuously for 40 years. Not by the people who wrote it, they left in the eighties. By the people who inherited it from the people who inherited it, none of whom have documentation because COBOL’s self-documenting promise only held if the original programmer was disciplined enough to write it well, and Fred Gruenberger put it perfectly in the 1960s: “COBOL in the hands of a master is a beautiful tool. COBOL as it’s going to be handled by a low-grade clerk somewhere will be a miserable mess.”

Guess which one predominated at scale.

What you’re left with, across most major financial and government institutions worldwide, is a codebase where the code itself is the only surviving record of what the system was designed to do. Not the documentation, there isn’t any, or what exists is decades out of date and describes a system that’s been patched beyond recognition. Not the engineers, they’re retired or dead. The code. Millions of lines of spaghetti logic that contains, embedded somewhere in its execution paths, the actual business rules of institutions that process more money in a day than most countries produce in a year.

This is what “institutional knowledge” means when it calcifies inside a technical system and the human layer that understood it walks out the door.

So why hasn’t it been replaced? This is the question everyone asks, and the answer people reach for is money. It’s too expensive. It is true that the Commonwealth Bank of Australia replaced its core system in 2012, a decision that took five years and cost $750 million for one bank, but it’s not the real answer. The real answer is more uncomfortable.

Replacing it would require admitting, in detail, in a boardroom and eventually in public, that you don’t fully understand your own system.

Think about what that actually means for the decision-maker sitting across from that choice. You’re a CTO at a major bank. You can authorize $65 million a year forever to keep the current system alive, stable, and out of the headlines. Or you can authorize a multi-year, billion-dollar project to replace infrastructure that currently works, during which any number of things can go catastrophically wrong in ways that will be visible to regulators, shareholders, and customers. The $65 million is a known cost. The replacement is an open-ended liability in a regulated industry where the failure modes are public and the tolerance for error is essentially zero.

Every rational actor in that room chooses the maintenance budget. Quarter after quarter. Year after year. “If the back end’s not broke, don’t fix it” became the dominant institutional strategy not because the people running these organizations are stupid or reckless, but because the incentive structure of public companies and government agencies systematically rewards deferral over confrontation. The cost of fixing it now is always higher than the cost of not fixing it until it isn’t, and by then the options have narrowed in ways that make the original decision look even worse in retrospect.

The result is that the leadership of most major financial institutions and government agencies is, right now, responsible for infrastructure they cannot fully interrogate. They know this. The people one level below them know this. The consultants billing $200 an hour to maintain it know this. Nobody says it plainly, because saying it plainly would require a follow-up sentence about what they’re going to do about it, and that sentence has no good answer.

The thing that should crystallize this for you is New Jersey in 2020. Not because it was a disaster, it wasn’t really, in the end, but because of what it revealed. When COVID crashed hundreds of thousands of unemployment claims into the state’s system simultaneously, the governor went on television and asked for COBOL volunteers. Not contractors. Volunteers. Because there was no one on staff who could update the code fast enough, and the crisis didn’t care.

That’s not a funding failure. That’s a knowledge failure. The money was there. The people weren’t. And this wasn’t some obscure state system; this was unemployment insurance, the most demand-sensitive government function imaginable, and it buckled the moment it was actually needed because the institutional capacity to modify it had already evaporated.

New Jersey eventually built a new system. And then ran it on a mainframe. With COBOL on the back end. Because there was no other viable path in the timeframe that mattered.

The GAO has identified 69 critical federal systems right now with what it describes as expanding security and operational risk. Social Security. The IRS. The Department of Veterans Affairs. Systems that millions of people depend on for money that determines whether they can pay rent. Running on code that predates the internet, maintained by a workforce that is aging out of existence with no succession plan, no documentation, and no institutional will to confront what happens when the last person who understands a given module decides they’ve had enough.

What makes this structurally different from every other tech debt story, and there are many, is the absence of a forcing function.

Y2K is the template. It was a genuine near-miss that got resolved, expensively and imperfectly, because it had a hard deadline built into the problem itself. January 1, 2000. You either fixed it or you failed publicly, visibly, on a specific date, with your name attached. No deferral. Every institution with a Y2K problem had to decide: fix it now or face the consequences on a known date. That structure produced action.

COBOL has no such clock. The workforce retires gradually. The knowledge evaporates incrementally. The maintenance cost rises in ways that are visible in a budget line but not in a headline. There is no date on which the failure becomes unavoidable and public. There are only individual retirements, individual systems becoming slightly harder to maintain, individual moments where someone discovers that the person who understood a particular piece of code left two years ago and nobody flagged it.

This is why the deferral continues. Not ignorance. Not incompetence. The rational calculation, made over and over by decision-makers who know exactly what the problem is, is that deferral is always cheaper than confrontation right up until the moment it catastrophically isn’t.

Here’s the take I keep waiting for someone to make, and no one does, because it requires sitting with something genuinely unsettling:

COBOL is not the last time we will do this.

The pattern builds critical infrastructure on something expedient, fails to plan for succession, allows institutional knowledge to calcify in a shrinking specialist population, and defers replacement because the cost is always higher than the cost of deferring. This pattern is not unique to a 65-year-old programming language. It’s the natural behavior of institutions that are optimized for short-term stability over long-term legibility.

We are doing it right now with other systems. We will do it again with systems that don’t exist yet. The lesson of COBOL is not “don’t use old technology.” It’s that institutions are structurally incapable of prioritizing the preservation of their own understanding over the preservation of their own stability. The two look the same from the outside until they don’t, and by the time they don’t, the people who made the original deferral decisions have retired.

Grace Hopper wanted to build a language anyone could read. She helped build one that only a priesthood can. The priesthood is retiring. The language is still running. The institutions that depend on it are, with a few exceptions, hoping the math works out.

It might. The three curves, shrinking workforce, rising maintenance cost, and evaporating institutional knowledge, haven’t crossed yet. There’s still time, technically, for a coordinated modernization effort that treats this the way Y2K was treated: as a national infrastructure problem requiring public coordination, shared risk, and a forcing function that doesn’t wait for a crisis to provide one.

But that would require someone in a position of authority to admit, plainly, that they have built the global financial system on a foundation they can no longer read, that fixing it will be expensive, risky, and disruptive, and that the alternative is worse.

Nobody has said that yet. Not clearly. Not in a way that sticks.

The system is still running. For now, that’s apparently enough.

Subscribe now

Beneath every AI chip, beneath the substrate, beneath the packaging, beneath several layers of material that you have also never heard of - there is a sheet of woven glass cloth thinner than a human hair. One Japanese company makes the version sophisticated enough to matter. It was founded as a cotton spinner in 1923. Apple, Nvidia, Google, Amazon, and AMD are currently fighting over their output, with mixed results.

The blindspot

For two years, the approved narrative about AI infrastructure has centred on chips. TSMC’s chairman stood at an industry awards ceremony in November and declared advanced-node capacity “not enough, not enough, still not enough” - roughly three times short of demand. (1) The four largest hyperscalers doubled their combined annual capex to around $600 billion. (2) Jensen Huang made pilgrimages to Seoul. SK Hynix sold its entire 2026 HBM output before 2025 had the decency to end. Consumer DRAM prices rose 4 times over two months. (3) Every major publication produced some version of the same piece: the chip shortage never really ended; it merely changed shape.

All of this is accurate. It is also, as it turns out, a description of the top three layers of a much longer supply chain. While the industry was staring upward at fabs and wafer counts and Jensen’s leather jacket, a separate shortage was forming quietly four tiers below - in a material with no ticker symbol, no earnings call, and until approximately six months ago, no press coverage whatsoever. Nobody, it seems, was watching what the chip sits on.

The material

T-glass, or low-CTE glass cloth, sits directly beneath the chip in the substrate. Its purpose is to keep the package from warping when processors heat to around 100°C, which they do constantly and with some enthusiasm. As AI chips have grown larger and added more stacking layers for advanced packaging such as CoWoS, manufacturers have had to double the T-glass usage per substrate. IC substrates expanded again in 2026. Demand roughly doubled. Supply did not, having been at full capacity since the second half of 2023. Customers who visit the factory are offered extended waiting periods and, presumably, tea. Orders are already backlogged into the second quarter of this year.(4)

The company

The company at the centre of this is Nitto Boseki - Nittobo - which posted net income of $37 million last year. Apple’s most recent quarterly profit was $24 billion. In autumn 2024, Apple flew staff to Japan and stationed them at Mitsubishi Gas Chemical, one of Nittobo’s downstream customers, to lobby for allocation. Finding that insufficient, Apple also asked the Japanese government to intervene.(7) Nvidia, showing rather more foresight, had already signed a mid-to-long-term exclusive supply agreement with Nittobo before the shortage became a topic of polite conversation in supply chain circles, which had the predictable effect of making everyone else’s position worse.(8) AMD and Nvidia dispatched senior executives directly to the factory floor. Nittobo’s stock subsequently rose more than 300%.(9)

Nittobo has announced ¥150 billion in capex to expand capacity, targeting a 3x increase by 2028. Its operating profit last year was a record $104 million. Meaningful new supply arrives at the earliest in late 2027.(10)

“No additional capacity is no additional capacity, even if you pressure Nittobo.” - executive at a substrate supplier serving AMD, Nvidia, and Apple, January 2026(7)

The price cascade

On August 1st, 2025, Nittobo issued a blanket 20% price increase on all premium glass fiber materials.(11) A procurement director at a major substrate manufacturer reportedly called their Taiwanese supplier the same afternoon. Standard E-glass sells for $3-5 per kilogram. High-end grades were $40-50. T-glass is now trading in the $80-100 range, a price trajectory Citigroup analysts expect to continue, with a further increase of 25% or more projected for this year.(12) BT substrate prices began climbing in July. Resonac raised prices on certain semiconductor packaging products by more than 30% in January.(13) Those costs transfer downstream with the implacable logic of water flowing downhill - to device makers, to retailers, and ultimately to consumers, none of whom have been consulted about any of this.

Why does society let this happen?

This shortage did not arrive unexpectedly. It arrived exactly as the structure of the global semiconductor industry guaranteed it would, given enough time and sufficiently rapid demand growth.

The semiconductor supply chain vertically disintegrated in the 1990s. Chip design went to America. Fabrication was done in Taiwan and South Korea. Specialty materials went to Japan, where companies had been quietly refining their craft-glass chemistry, weaving tolerances, furnace recipes - for decades, largely unobserved by their eventual customers. The logic was efficiency: let each region do what it does best, connect the parts via container ships and long-term contracts, and call it a supply chain. The OECD has catalogued the result: the semiconductor industry is simultaneously the most upstream and the most geographically concentrated industry in its dataset. A chip can cross international borders up to 70 times before reaching a finished device. (14) Each crossing is a handoff. The further upstream a material sits, the less any downstream company has ever thought to inquire about it. Chipmakers watch substrate makers. Substrate makers watch glass cloth makers. Glass cloth makers, being at the bottom, watch nobody in particular.

There was, it should be noted, a preview. In 2021, a shortage of BT resin substrates - upstream, unglamorous, entirely unmonitored by the companies depending on them - caused finished silicon wafers to sit completed but uninstallable for months. Volkswagen cut production of hundreds of thousands of vehicles over a component its procurement team had never audited. (15) The industry commissioned reports, held conferences, and pledged $630 billion in new semiconductor investment, almost all of it directed at fabs and memory. The materials several tiers upstream received no such programme. The lesson, in other words, was learned and quietly set aside, as lessons of this kind tend to be until the next crisis makes them unavoidable again.

Who gets hurt

The allocation mechanism is straightforward, if not exactly fair. AI companies arrive with longer contracts, deeper pockets, and in Nvidia’s case a pre-signed exclusive arrangement. They get first access. BT substrate makers - the ones supplying chips for iPhones, cars, routers, and medical devices - receive what remains, which Goldman Sachs projects will be short by double digits for multiple consecutive quarters. (16) Citigroup has said, with a precision that presumably gave its consumer electronics clients some discomfort, that shortages will concentrate in consumer products specifically because those customers cannot outbid hyperscalers .(12) Aspeed Technology, which makes chips for server management controllers, has already revised its 2026 growth forecast downward on account of substrate constraints.(17) That is not an analyst projection. It is a reported earnings revision. The shortage is not coming. It is here, and it is being rationed by price.

The alternatives problem

Alternatives are being pursued with the urgency appropriate to a $3 trillion industry that has recently discovered it depends on a cotton mill. Taiwan Glass became the third company globally to produce low-Dk glass cloth - a development worth noting, not least because the existence of only three such companies illustrates rather than resolves the problem. (18) Nittobo has partnered with Nan Ya Plastics, which should be able to weave about 20% of Nittobo’s specialty fiber output by 2027, which is helpful in the way that an umbrella is helpful in a flood. (19) China’s Grace Fabric Technology is being coached on quality standards by Mitsubishi Gas Chemical at Apple’s specific request, a sentence that would have seemed implausible at an Apple product launch several years ago. (7) Qualcomm has visited Unitika, a smaller Japanese supplier. A Chinese quartz-glass variant has reportedly cleared Nvidia’s certification.(20) None of it adds up to enough, for a reason that resists the usual solutions: every glass fiber must be perfectly round with zero internal bubbles, or the chip substrate fails in ways that cannot be addressed after assembly. The qualification process runs through three stages - CCL maker, substrate producer, and end customer - each taking 1 to 2 years. Nittobo has been refining this for decades. That lead exists not as a patent but as accumulated knowledge distributed across furnace operators and production floors, and it does not transfer on a purchase order.

Nittobo’s position

Nittobo, for its part, is approaching the situation with equanimity that borders on the philosophical. The ¥150 billion capacity investment is real. The Nan Ya partnership will add weaving capacity by 2027. The company plans to triple output by 2028. But the CEO has said publicly that Nittobo will not expand at the same pace as the AI market, will prioritise quality over volume, accepts that it will lose some market share to competitors, and has added - with what the Nikkei Asia journalist covering the interview diplomatically described as satisfaction - that electronics companies have “finally recognised glass cloth as a key material.”(21) The company recorded its highest-ever operating profit last year. It raised prices 20% in August. Citigroup analysts project another 25% or more increase in 2026. (12) Nittobo’s stock is up more than 300%. (9) A century-old silk-weaver has achieved, in the space of eighteen months, the pricing power that most semiconductor firms spend decades trying to cultivate.

The full weight of the AI buildout - $630 billion in announced US semiconductor investment, TSMC raising its capex to $52-56 billion this year alone,(22) hyperscalers booking wafer capacity in tranches through the end of the decade - rests, at one remove, on the judgment of a former silk weaver in Tokyo with 500 employees, who has publicly stated that he does not expect AI demand growth to be sustained, and who has arranged his capacity expansion timeline accordingly.

The chips are fine. The cloth is the problem.

My take: five things the T-glass story reveals that the T-glass story isn’t really about

01 · Nvidia’s supply chain team deserves more credit than Jensen gets

The most revealing detail in this story is not the shortage. It is the timing. NVIDIA signed a mid-to-long-term exclusive supply agreement with Nittobo before the shortage became a matter of public discussion. (8) That means someone, somewhere in Nvidia’s procurement organisation, identified T-glass as a binding constraint and moved to lock in allocation while Apple was still managing its Nittobo relationship through its substrate supplier’s substrate supplier. This is not glamorous work. It does not appear in keynotes. Jensen Huang does not present slides about glass cloth weaving tolerances at GTC. But it is, arguably, one of the more consequential competitive advantages Nvidia currently holds over its rivals - not a better chip, not a better software stack, but a better answer to the question: have you actually mapped your dependencies four tiers deep? The companies now flying executives to Tokyo already lost the round that mattered.

02 · The industry has now failed the same exam twice

The T-glass shortage is the 2021 chip shortage, one abstraction layer further down. In 2021, the global auto industry discovered that it had no visibility whatsoever into its tier-3 semiconductor suppliers, and that this was a problem that cost Volkswagen hundreds of thousands of vehicles and Ford roughly $1 billion in lost revenue. (15) The industry’s response was to invest massively in the layer it could see - fabs - and nothing in the layers it couldn’t. The CHIPS Act funds wafer fabrication. There is no upstream materials provision. No one announced a programme to auditthe glass cloth supply. No one asked what the substrate’s substrate was made of. The next shortage is already embedded somewhere in the stack. It may be drill bits - PCB drilling machinery, also in emerging shortage, and also unglamorous. (7) It may be something else. The safest assumption is that it is whatever nobody is currently writing about.

03 · Japan’s upstream position is stranger and more durable than it appears

Nittobo makes glass cloth. Ajinomoto - best known internationally for inventing monosodium glutamate - makes ABF film, without which Nvidia’s packaging does not exist. Shin-Etsu makes silicon wafers. JSR makes photoresists. Collectively, Japanese companies control an extraordinary share of the most upstream, least substitutable materials in the semiconductor stack, almost none of which have Western equivalents and almost none of which attract Western media coverage until they run short. What makes these positions durable is not patents - it is production knowledge that exists in the hands of specific furnace operators at specific facilities. (23) Taiwan Glass becoming the third company to produce low-Dk glass cloth is, by normal market logic, a competitive development. It is also, when you note that only three companies in the world have managed it despite years of demand, a fairly vivid illustration of how high the barrier actually is. This is not a supply chain risk. It is a structural feature of advanced materials science that the technology industry has spent thirty years choosing not to think about very hard.

04 · The CEO of Nittobo is the most interesting figure in this story and receives roughly 2% of the coverage

Hiroyuki Tada runs a company that, by the standards of its customers, is tiny. He has watched his factory receive delegations from Nvidia, AMD, Apple, and - apparently - the Japanese government, all of whom would like more material than he is able or willing to produce. His public response has been to note that he doesn’t expect AI growth to be sustained at its current pace, that Nittobo will not overextend itself chasing demand forecasts it finds implausible, and that it is gratifying that the industry has finally noticed glass cloth.(21) This is not timidity. It is the institutional memory of a company that has been in business for over a century and has watched several generations of Western technology customers describe their requirements as permanent before quietly revising them downward. The asymmetry here is worth sitting with: the constraint on the AI buildout is being managed by someone whose reference frame includes the last three times Silicon Valley was wrong about its own demand projections. Jensen Huang operates on a product-cycle timescale. Hiroyuki Tada operates on a timescale of furnace investments. These are not the same timescale, and at the moment, his is the one that matters.

05 · The person paying for the AI buildout is not who you think

The allocation mechanism in a shortage is price. Companies that pay more get the supply first. Companies that pay less get what remains. The companies paying more for T-glass are hyperscalers building AI infrastructure. The companies getting what remains are the ones making chips for iPhones, laptops, and the various connected devices that constitute most people’s experience of consumer electronics. Citigroup said this explicitly: shortages will concentrate in consumer products.(12) The inflationary pressure from the T-glass shortage - the 20% Nittobo increase, the 30% Resonac increase, the rising BT substrate prices, the downstream device cost increases - is not landing on the data centers. It is landing on the person buying a phone. We have spent two years discussing the democratisation of AI as if the technology's benefits were flowing broadly while its costs were being absorbed by well-capitalised institutions. The supply chain suggests a different arrangement: the infrastructure is being built on capital that flows up, and the costs are transferred down through a series of price increases on materials that most consumers have never heard of and never will.

Thank you for reading :)
Please consider subscribing if you made it this far!

Footnotes

1. TSMC Chairman C.C. Wei at the Semiconductor Industry Association Awards, San Jose, November 2025. Reported by Tom’s Hardware, Nov 25 2025.

2. Tom’s Hardware / Bajarin analysis, Jan 2026. Covers Amazon, Google, Microsoft, Meta combined capex.

3. Deloitte 2026 Semiconductor Industry Outlook, published Feb 2026. Consumer DRAM price trajectory Sep-Nov 2025.

4. Digitimes Asia, “IC substrate shortage to persist in 2026,” Nov 20 2025.

5. Digitimes Asia, Nov 20 2025. Taiwan IC substrate makers’ internal projections for T-glass supply gap.

6. TrendForce and Meritz Securities (analyst Yang Seung-soo) via @Jukanlosreve, Jun 2025. Some estimates cite 80% when including PPG; 90% reflects production-grade T-glass for advanced packaging specifically.

7. Nikkei Asia, “Apple and Qualcomm face supply woes over Japan’s glass cloth,” Jan 14 2026. Reported by Seeking Alpha, Jan 14 2026.

8. Meritz Securities analyst note (Yang Seung-soo), reported via @Jukanlosreve on X, Jun 2025.

9. StockCrock Substack, “The Shifting Monopolies of AI,” Feb 2026. Nittobo (3110.T) ~54x trailing P/E as of early 2026.

10. UGPCB industry analysis, Jan 2026. Nittobo ¥150B capex figure. Operating profit $104M from Nikkei Asia, Dec 2025.

11. Commercial Times via TrendForce, “BT Substrate, Fiberglass Prices Eye 20% Hike,” Jul 22 2025. Price notice issued Aug 1 2025.

12. Citigroup analyst note cited in Nikkei Asia reporting, Jan 2026. 25%+ further increase projection for 2026.

13. Resonac price increase: Nikkei Asia / WSJ reporting, Jan 2026.

14. OECD, “Vulnerabilities in the Semiconductor Supply Chain,” Jun 2023.

15. Multiple semiconductor supply chain journals; Rand Technology supply chain analysis, Aug 2025. Ford $1B figure from industry estimates widely cited in 2021-22 chip shortage coverage.

16. Goldman Sachs report cited in TechNews/TrendForce, Nov 2025. Double-digit percentage T-glass shortfall projected for BT substrate makers over multiple quarters.

17. Nikkei Asia, Jan 2026. Aspeed Technology chairman Chris Lin: “We do see very strong demand and good growth momentum for 2026 but we do have to adjust down our forecast because of the constraints in material supplies.”

18. TrendForce, “Taiwan Glass enters AI server supply chain,” Nov 2025.

19. TrendForce, “Nittobo Expands Glass-Fiber Output with Nan Ya,” Nov 28 2025. Nan Ya Plastics to handle ~20% of specialty fiber weaving by 2027.

20. Tom’s Hardware, Jan 16 2026. Chinese quartz-glass certification referenced in Nikkei Asia reporting.

21. Nikkei Asia, “Why a key Japanese Nvidia supplier is betting on a premium niche strategy,” Dec 22 2025. Interview with CEO Hiroyuki Tada.

22. TSMC Q4 2025 earnings guidance. Capex guidance $52-56B for 2026 reported by AI Magazine / TSMC investor relations, Jan 2026.

23. Techovedas, “Nvidia’s New AI Bottleneck: Japan’s Nittobo Glass Cloth,” Jul 2025. Semiconductor supply chain expert characterisation of embedded process knowledge.

This week’s edition comes with a companion piece: The Probabilistic Expected Value Modeling Workbook - a step-by-step framework for evaluating conflict, disruption, and legal risk in stocks, with AI prompt templates and a worked example. Download link at the bottom.

THIS WEEK’S THROUGHLINE

We confused the map for the territory and then optimized for the map.

Five stories this week. Five completely different industries: wildlife management, geopolitical manufacturing, financial theory, healthcare disruption, and brand strategy. And yet when you lay them next to each other, a single, uncomfortable thesis rises to the surface: somewhere along the way, we stopped measuring reality and started measuring our measurement of it. We optimized the instrument until it broke the thing it was tracking. This week, every story is a case study in what happens when the map becomes more important than the territory, and nobody notices until the damage is done.

STORY 01 | WILDLIFE / ECONOMICS

The Wolf Policy Has a Problem

Start with what looks like a boring regulatory story. The U.S. Fish & Wildlife Service and various state agencies have been wrestling with how to manage wolf populations in states like Wisconsin and Montana for decades. The political pressure from ranching communities is intense and legitimate. Wolves kill livestock, wolves cost money, and wolves are genuinely disruptive to a business model that already operates on razor-thin margins. So policymakers reached for the obvious lever: controlled killing permits. The logic is straight out of Economics 101. If you create a legal supply channel for something people want, in this case, the ability to kill wolves, the illegal market for that thing should collapse. Legalize some, and the pressure from the black market dissipates.

The kill permits didn’t reduce illegal poaching. They caused a 34% spike in it.

That number comes from radio-telemetry data; wolves that didn’t die naturally or from permitted kills simply vanished from the tracking data. Cryptic poaching, researchers call it. The legal killings appeared to accelerate the illegal ones. The metric that was supposed to measure tolerance was instead performing tolerance, signaling to the fringes that the wolf was a pest, a problem, something to be managed rather than protected. It validated the exact culture of resentment it was trying to diffuse.

Montana provides the even more devastating data point. In 2024, the state sanctioned the removal of 308 wolves, nearly a third of the entire population, under the assumption that fewer predators would mean fewer livestock deaths. The Integrated Patch Occupancy Model (iPOM) tracked results across millions of acres. The livestock conflict numbers didn’t fall. They plateaued. And the reason is biological, not economic: wolves operate within rigid hierarchical structures. When you eliminate an alpha, you don’t subtract one unit of predator from the ecosystem. You shatter the pack’s governance. What you’re left with is a fractured group of inexperienced juveniles who have no idea how to hunt elk, so they pivot to the slow, easy calories standing in the pasture. You made things worse by making them look better on paper.

The $162,000 problem nobody wants to pay

The financial reality is genuinely brutal. A UC Davis analysis put the impact of a single wolf on a ranching operation at up to $162,000, and that’s not just a dead calf. That’s the ‘landscape of fear’ effect, a real biological phenomenon where the mere presence of wolves patrolling a perimeter sends the entire herd into chronic stress. Grazing efficiency drops. Cortisol burns calories that the cattle should be storing. Immune systems weaken. Pregnancy rates fall. For a family running a cattle operation in a rural Montana county, this isn’t an abstract environmental debate; it’s an existential financial threat.

But here’s the thing that should puncture the entire culture war narrative around this topic: when ranchers were surveyed anonymously, away from Twitter discourse and identity pressure, 79% of farmers and ranchers expressed support for Endangered Species Act protections. The war you’ve been watching on the news is a fabrication. The real majority opinion is nuanced, financially stressed, and completely drowned out by the loud fringes on both sides, who benefit from the conflict continuing.

The instrument was measuring the wrong thing. Kill permits were designed to measure and build tolerance. Instead they measured and broadcast hostility. The map said ‘progress.’ The territory said ‘we made it worse.’

What’s actually needed isn’t more permits or more biologists. It’s an economic mechanism that compensates the rancher for privately subsidizing a public asset, the wolf, which the urban public romanticizes on calendars, while the rural rancher absorbs the full cost. That’s a policy design problem, not a species management problem. We keep solving the wrong equation because we fell in love with the equation we already had.

STORY 02 | GEOPOLITICS / MANUFACTURING

The Chinese Auto Industry Has a Problem (And So Do We)

BYD sold three million electric vehicles last year. More than Tesla and Volkswagen combined in some quarters. Chinese automakers are paying 8 cents per kilowatt-hour for electricity, roughly half what American factories pay. That cost differential doesn’t just save money at the outlet. It cascades. Cheaper power means cheaper smelting. Cheaper smelting means cheaper steel. Cheaper steel means cheaper battery casings. Cheaper batteries mean a four-thousand-dollar electric car that somehow doesn’t feel disposable. Ford’s CEO visited China’s manufacturing infrastructure and described the experience as ‘humbling.’ That’s corporate diplomacy for: we genuinely cannot build like this anymore, and we have no short-term plan to change it.

The market says China is building cars. The territory says China is building infrastructure that doubles as military logistics.

But the deeper story isn’t the cost structure. It’s what the cost structure is hiding. Chinese automakers haven’t just gotten efficient; they’ve built what economists call vertical integration at a national scale. Lithium to chassis to battery, all within national borders, all subsidized by state-directed capital. The closed industrial loop exists not just to make cars cheaper, but to make the country’s industrial base impossible to disrupt from outside.

And underneath even that is something more alarming. The artificial economic equilibrium is propping up all of this. China’s property boom financed the construction of entire cities without residents, ghost cities, manufactured demand, and GDP growth conjured from concrete and political will rather than genuine consumer need. The economy, on the dashboard, looked healthy. The territory, vast quantities of empty apartment buildings, overextended developers, and debt that will eventually have to be reckoned with, told a different story.

The conversion problem the West isn’t having

In World War II, American automobile manufacturers converted their production lines to produce bombers within months. The industrial agility of civilian manufacturing was the country’s strategic weapon; the ability to pivot a Buick factory to wartime production was not incidental to the Allied victory; it was central to it. That capacity has quietly eroded over the past four decades of offshoring and quarterly-earnings optimization. The factories that once taught America how to adapt are largely gone, and the institutional knowledge that went with them is gone too.

China is rebuilding that playbook in reverse. The car factories, battery plants, and cargo shipyards being built today with civilian investment are also being designed with military convertibility in mind. Even if the property boom collapses, and it may, the foundries, ports, and refineries remain. Real assets, convertible at short notice from commerce to a command economy. The West, by contrast, is rich in capital and genuinely poor in capacity. Our GDP looks fine. Our industrial flexibility, the actual strategic substrate of deterrence, has been quietly hollowed out while we were watching the stock market.

The West looked at the metric (GDP growth, consumer demand, car sales) and assumed the territory was healthy. The territory is a dual-use military-industrial complex dressed in consumer packaging.

Tariffs in this framing aren’t petty protectionism. They’re an attempt to buy time for an industrial base that no longer knows how to think about itself as a strategic asset rather than a quarterly cost center. The question is whether we have enough time left to rebuild the thing we spent forty years dismantling.

STORY 03 | MARKETS / THEORY

George Soros Has a Problem And He Wants You to Have One Too

George Soros made one of the most important and most ignored contributions to economic theory in the twentieth century, and he named it in the driest possible way: the General Theory of Reflexivity. The idea is simple and devastating. Standard economic theory, the kind you learn in any introductory course, assumes that markets are information-discovery mechanisms. Prices are supposed to converge toward fundamental value over time as more information becomes available. The market is a thermometer, and reality is the temperature.

Soros looked at this and said: No. ” The thermometer changes the temperature. Markets don’t just passively read fundamental value; they actively construct it. When enough participants believe a stock is worth a certain price and act on that belief, their collective behavior changes the underlying conditions that determine whether the price is justified. The belief produces the reality that proves the belief. It’s a recursive loop, and once you see it, you cannot unsee it.

We aren’t trading fundamentals. We are trading on the basis of how much the crowd is currently willing to dissociate from reality in service of a compelling narrative.

The 2026 study on inaccurate metaperceptions that keeps showing up in behavioral research adds a layer to this: we don’t just participate in feedback loops, we actively misread where other people are in them. We assume everyone else is more committed to the narrative than they are. We assume the consensus is more robust than it is. And so we stay long on stories that a calmer read of the room would tell us are already starting to crack.

The AI boom as a reflexive super-cycle

Apply this lens to the current AI investment cycle, and it quickly becomes uncomfortable. The thesis goes: AI will transform economic productivity at a scale not seen since electrification. This widely held thesis forces every CEO with a public market obligation to divert capital expenditure toward Nvidia and the infrastructure stack, or face the penalty of appearing strategically behind. The capital flows in. The capital flows produce quarterly results that validate the thesis. The validation attracts more capital. The map draws the territory.

The problem Soros has spent his career exploiting is the gap. The gap between the narrative and the underlying utility. The gap between what the chart implies and what the balance sheets actually say. Valuation multiples have detached from anything you could call traditional cash flow analysis, not because investors have abandoned rationality, but because, in a reflexive market, pricing in a future state that the investment itself is helping to create isn’t irrational. It’s just that when the gap becomes too wide, the snapback is violent. The mean reversion that bankrupts the narratives and sometimes the nations built on them.

The AI market is not tracking AI utility. It is tracking the collective belief in AI utility, and that belief is self-reinforcing until suddenly it isn’t. Soros doesn’t think you’re stupid for believing the hype. He’s just positioning to sell you the top.

STORY 04 — HEALTHCARE / MARKETS

HIMS Has a Problem. Or Does It?

HIMS stock dropped 30% in a single session. Novo Nordisk filed a patent lawsuit. The DOJ issued a referral. Retail investors, trained to read major legal action as a signal of fundamental distress, sold immediately. The story the market told itself was: existential threat, exit now. The story the underlying business was telling, at exactly the same moment, was: our top-line revenue has expanded 14 times since the IPO, and our free cash flow per share is rising and to the right.

The disconnect here is a masterclass in what we might call the diagnosis problem, the tendency to mistake a symptom for a root cause, a lagging indicator for a leading one. Novo Nordisk dropping legal bombs is what an incumbent does when it’s scared. It’s not what a healthy monopoly does when a scrappy startup is failing. Patent lawsuits are among the most expensive and slow-moving business strategies. Companies do not deploy them unless they’ve concluded that the competitive threat is real and cannot be addressed through product or price competition alone.

Novo Nordisk is telling you HIMS is winning. The market heard ‘HIMS is losing.’ Both statements can be true and refer to the same lawsuit.

The Napster analogy that actually holds

The structural play HIMS is making is genuinely interesting and worth understanding properly. They’re operating under the FDA’s 503A compounding exemption, a regulatory pathway that allows pharmacies to synthesize tailored dosages of drugs when the commercially available version isn’t clinically optimal for the patient. They’re using this to provide compounded semaglutide (the active ingredient in Ozempic and Wegovy) at substantially lower cost than the branded product. This is what Napster did to record labels: it didn’t copy the music industry’s product; it copied the music and, in doing so, revealed that the industry’s real business was in the distribution monopoly, not the music itself.

But the deeper game HIMS is playing is even more interesting than the compounding arbitrage. Their Trojan horse is longitudinal biomarker data. By making diagnostics dramatically cheaper, think of HIMS Labs as the Duolingo owl for your endocrine system; they’re building a network effect around patient data that inverts the traditional value chain. Once a patient has comprehensive data on their hormonal and metabolic functions, the decentralized telehealth network becomes more valuable than the centralized physician visit. The model flips from reactive sick-care to proactive, personalized peptide therapy. That’s not a story about a compounding pharmacy. That’s a healthcare infrastructure story.

How sophisticated investors actually price this

The smart money isn’t treating this as a binary pass/fail event. The methodology is probabilistic expected-value modeling: building a spectrum of scenarios (regulatory extinction, costly setback, favorable precedent) and assigning weighted probabilities to each, then comparing the expected value across those scenarios to the current market price. When you run that exercise on HIMS, the current free cash flow trajectory in the non-extinction scenarios tends to produce a positive expected value even at meaningful probability weights for the downside. The market panic was an emotional response to a legal filing, not an analytical assessment of the business.

The market’s instrument for reading HIMS was a binary legal-risk meter. The actual territory is a multi-scenario probability distribution across a genuinely disruptive business model. The map said catastrophe. The territory said: opportunity for the patient.

STORY 05 | BRAND STRATEGY / MARKETING

Duolingo’s Brand Has a Problem

Duolingo had its worst trading day in recent memory after reporting slower user growth and signaling a shift in brand strategy. On the surface, it looked like a monetization story. Below the surface, it was something more structurally interesting, a company that built one of the most effective audience acquisition engines in consumer technology history, and then discovered that the engine had quietly become more powerful than the company itself.

The unhinged green owl was not an accident of internet virality. It was a deliberate and sophisticated brand strategy built around the insight that extreme, unpredictable personality is the fastest way to hack algorithmic visibility on platforms that reward emotional engagement. The TikTok posts threatening users who hadn’t done their lessons, the LinkedIn backlash at a human employee’s death, the absurdist humor that read as genuinely unhinged rather than corporate-approved quirky, it worked. User growth hit 40% in some quarters. The acquisition funnel was an engine.

The brand built a distribution system so powerful that the distribution system became the brand. Then the company tried to evolve — and the system fought back.

The double bind

Here’s the structural trap that makes this more than a marketing story. When Duolingo paused the unhinged content following a LinkedIn controversy, user growth decelerated almost immediately. The engagement algorithm, which had been trained on years of chaotic, high-reaction content, recalibrated. Posts that didn’t match the established persona received fewer reactions, which meant the algorithm served them to smaller audiences, which in turn led to even fewer reactions. The feedback loop that had been driving growth inverted, and it started punishing deviations.

The brand had created what you might call an identity cage. The audience had assigned Duolingo a character, the manic owl, and the distribution infrastructure now only rewarded performance of that character. When Duolingo tried to act like an AI education company with serious aspirations about transforming how humans learn language, the audience punished them for breaking the role. The product ambition (world’s most advanced AI tutor) and the brand identity (unhinged prank account) had become mutually incompatible — and fixing one meant breaking the mechanism that had built the other.

The path out isn’t obvious. Restarting the unhinged posts will recover some growth, but not at the same velocity. Algorithms reward consistency over time, not resurrection. The more durable fix requires something most companies find genuinely difficult: intentionally diluting the thing that made you successful to build something that can scale to the next level. Diversifying the brand voice. Building campaigns across multiple tones. Accepting a period of slower distribution in exchange for a more robust long-term identity.

The engagement metric was tracking brand health. What it was actually tracking was brand brittleness, how completely the audience had been trained to expect one specific thing, and how costly deviation from that thing would be.

THE THROUGHLINE

We Optimized the Map Until It Ate the Territory

Five stories. Five industries. One underlying failure mode.

In every case this week, the story wasn’t about bad intentions or missing information. It was about a measurement instrument that had drifted from the thing it was built to track, and a system that kept optimizing for the measurement long after the measurement stopped reflecting reality.

Wolf kill permits were built to measure tolerance. They started producing tolerance on paper while breeding hostility on the ground. China’s GDP figures were built to measure economic health. They started measuring the velocity of constructed demand in empty cities while the real output, dual-use military infrastructure, wasn’t on any dashboard the West was watching. The AI market was built to price future utility. It started pricing the momentum of collective belief in future utility, a subtly but catastrophically different thing. The HIMS stock price was built to reflect the business's health. It began reflecting retail investor panic in a legal document, decoupled from the underlying company's free cash flow trajectory. And Duolingo’s engagement metrics were built to measure brand resonance. They ended up measuring the depth of the audience’s expectation of one specific persona, a brittle, fragile thing masquerading as strength.

The map looked perfect. The territory was quietly falling apart.

There’s a philosophical concept here worth naming: Goodhart’s Law. When a measure becomes a target, it ceases to be a good measure. Originally formulated in the context of monetary policy, it has quietly become one of the most universal laws of complex systems. It applies to wildlife management. It applies to geopolitical industrial strategy. It applies to financial markets. It applies to brand-building algorithms. It applies, frankly, to almost every domain where humans have decided that a number is a reliable proxy for reality, and then have proceeded to optimize for the number rather than the reality.

What makes this week’s set of stories particularly interesting is that none of the failures are stupid. The logic in each case is internally consistent. According to economic theory, kill permits should reduce poaching. Chinese factory output should, by GDP accounting, reflect genuine growth. Narrative momentum, according to market theory, should eventually connect back to fundamentals. A 30% stock drop should, by conventional wisdom, signal distress. Brand engagement, by distribution logic, should indicate audience health. Each of these is a reasonable proposition. The problem isn’t the logic. The problem is that the logic was applied to an instrument that had already drifted from what it was measuring, and nobody caught the drift until the system was deep inside it.

What to actually do with this

The practical question this throughline raises is: how do you know when your instrument has drifted? The honest answer is that it’s very difficult from inside a system to see that the ’s become self-referential. This is precisely why Soros’s reflexivity framework is so useful; it provides a mental model for asking not just ‘what does the metric say?’ but ‘what is the metric actually measuring, and has it drifted from what I need it to measure?’

For investors, this means building probabilistic scenario models rather than treating any single legal event or earnings miss as a definitive signal. For policymakers, it means periodically stress-testing whether the measurement instrument is still tracking the underlying social or biological reality you care about. For brand strategists, it means distinguishing between engagement that reflects genuine audience resonance and engagement that reflects algorithmic conditioning. And for anyone trying to understand geopolitical competition, it means asking not just what the economic numbers say, but what the physical assets being built could be repurposed for.

The territory always reasserts itself eventually. The wolves don’t care about the kill permit policy. The empty cities don’t care about GDP growth targets. The market doesn’t care about how compelling the AI narrative is. The customer doesn’t care about how many times they’ve laughed at the unhinged owl. Reality has a way of punching through the map, and when it does, the people who were still navigating by the map tend to be the most surprised.

The instrument has a problem. The instrument is us.

MY PEV FRAMEWORK DOWNLOADABLE

One of the five stories this week - the HIMS collapse and what the market got wrong about it - illustrates a framework that is worth keeping in your toolkit permanently. It is called Probabilistic Expected Value Modeling, and the short version is this: stop treating conflict events as binary pass/fail verdicts and start treating them as probability distributions. I built a free companion workbook that walks through the full five-step process, includes ready-to-paste AI prompt templates for each step, and gives you a partially completed worked example to finish yourself. If the HIMS section of this newsletter made sense to you and you want the actual tool to apply that thinking to any stock you are watching, the workbook is attached to this email. Download it, work through it once on a real situation you care about, and it will change how you read every lawsuit, regulatory probe, and earnings shock you encounter after that.

DOWNLOAD/VIEW HERE

has a problem newsletter · @maxxrosenblum · weekly

If this resonated, share it with someone who would disagree.
And most importantly, don’t forget to subscribe to yourself and/or a friend :)