The Ninth Circuit ruled this morning that Kalshi’s sports event contracts are likely not derivatives under federal law because they are sports bets. The published opinion in KalshiEX v. Assad affirms a Nevada district court’s decision dissolving the injunction that had shielded the exchange from state gaming regulators. The panel was unanimous. Judge Ryan Nelson wrote the opinion, joined by Judge Bridget Bade.

“The crucial context that leads us to resist the broadest possible reading is that Kalshi has a gambling problem.”

KalshiEX v. Assad, No. 25-7516 (9th Cir. Aug. 28, 2026)

For Kalshi to deny that its contracts are sports bets, Nelson wrote, “is disingenuous.” The opinion opens with Kalshi’s own advertising, which calls the product “the first app for legal sports betting in all 50 states,” and notes that sports contracts produced over 90 percent of the company’s trades in 2025 and 95 percent of its revenue.

The ruling creates a direct circuit split. On April 6, the Third Circuit reached the opposite conclusion in KalshiEX v. Flaherty, holding that the same contracts are likely swaps within the CFTC’s exclusive jurisdiction and barring New Jersey from enforcing its gambling laws. Two federal appeals courts now read the same statutory definition to opposite ends, in published opinions. New Jersey signaled in June that it would ask the Supreme Court to review its loss. The split it needed now exists.

The panel turned Kalshi’s own words against it. In 2024, fighting the CFTC over election markets in the D.C. Circuit, the company argued that “gaming” covers any contract contingent on a game and that Congress did not want sports betting on derivatives markets. “We agree,” Nelson wrote. The opinion goes further than rejecting preemption. It concludes that Kalshi’s self-certification and listing of the contracts violated the CFTC’s own regulation, 17 C.F.R. § 40.11, a rule the agency has never invoked against the company.

Kalshi said it will keep fighting. “Despite the 9th Circuit’s opinion, we still believe the CFTC regulations as written do not prohibit sports contracts, and in any event, the CFTC is working to clarify those regulations,” a company official said in a statement reported by Sportico. “We will be seeking further review.” Shares of DraftKings and FanDuel’s parent rose more than 5 percent in afternoon trading on August 28. Kalshi is not publicly traded; the company is reportedly raising new money at a $40 billion valuation.

The commission itself lost today. The CFTC argued the appeal as amicus on Kalshi’s side of the jurisdictional question, seventeen days after invoking emergency authority to order the exchange to keep operating nationwide over New York’s objection. Judge Kenneth Lee concurred separately and said where the fight goes next: the statute may leave the commission discretion over gaming contracts, and the prohibition holds only while § 40.11 stays on the books. The agency has proposed amending it. That rulemaking is now the single lever that could move the entire analysis. Kalshi is relying on it.

The ruling is not total. The panel remanded Kalshi’s election contracts for separate analysis and decided nothing about them. And the ruling applies a preliminary standard, likelihood of success, with the merits still ahead. The same panel denied Kalshi’s motion to stay the Nevada state proceedings, where a penalty motion seeking $120,000 per day is pending.

The Illinois position

Illinois Attorney General Kwame Raoul sits on both sides of this map. His office joined the 39-state amicus brief backing Nevada in the Ninth Circuit. It also defends three federal lawsuits in Chicago, filed by Kalshi, the CFTC and the Department of Justice, and Coinbase, all attacking the state’s new sports wagering law and its 25 percent tax on prediction market trades. All three cases sit with one judge, Martha M. Pacold, with four preliminary injunction motions fully briefed before her.

If sports event contracts are not swaps, the Commodity Exchange Act does not displace the Illinois statute, and the fight over the tax never starts. The Ninth Circuit does not bind the Seventh, and the Third Circuit’s contrary holding remains live authority for Kalshi in Illinois. Pacold now has a published opinion on each side of the question she must answer.

On the afternoon of the ruling, motions for leave to file landed in all three Chicago cases within a three-minute span, from 1:13 to 1:15 p.m. Central. The docket entries do not yet identify the filings’ contents or the moving party.

Under a June agreement filed on the docket, Illinois refrains from enforcing its sports wagering and criminal gambling laws against Kalshi only while the company’s injunction motion is pending. Whenever Pacold rules, the pause ends. The circuit split is national. The first court forced to choose between its two sides may be a single courtroom on Dearborn Street.

GrayStak publishes its full investigation on Friday, September 4: how Illinois wrote a 25 percent tax on prediction market trades, how all three federal challenges landed in one courtroom, and what the exchange told the Ninth Circuit about which tax it accepts.

GrayStak is a media and political risk company. Our real-time Domestic Political Volatility Index (DPVI) measures the velocity and volatility of U.S. domestic politics. We cover political and institutional risk across the United States, centered in Washington, D.C., New York, and Chicago.

Reported from the slip opinion published by the Ninth Circuit on August 28, 2026 (No. 25-7516), the dockets of the Ninth Circuit and the Northern District of Illinois, and prior GrayStak reporting. Kalshi’s statement, the share price movements, and the reported valuation are as reported by Sportico, the share figures as of afternoon trading on August 28. GrayStak holds no position in any company named.

Why it matters. Municipal Code 18-14-102.3 withholds energy data for buildings with data centers until the city’s Chief Sustainability Officer determines that the benchmarking tool can adjust for them. She then reports that determination to the Mayor and the City Council. No Council vote is required at any point. The city has not said whether it has ever made the determination.

What is pending. Twelve of the committee’s fifteen members already sponsor the underlying ordinance. The Clerk has docketed a substitution action for September 1. The agenda posted August 27 and describes the item in the terms of the version the substitute replaced. Written public comment closes at 2 p.m. August 31.

The substitute grew from one code section to three

Hadden introduced O2026-0027165 on July 15. It ran one page and changed a single provision, striking the letter S from a list of occupancy classifications in Municipal Code 18-14-101.3, the definition of a covered building. The phrase “data center” did not appear in it. GrayStak reported on that ordinance and on the zoning classification behind it on August 17, in Chicago Zones Data Centers as Offices.

Hadden’s substitute runs two pages and amends 18-14-101.3, 18-14-101.4 and 18-14-102.3.

The Clerk’s caption describes it as an amendment of Sections 18-14-101.3 and 18-14-101.4 “regarding covered buildings and related public availability of water usage data” and of Section 18-14-102.3 “removing ‘data center’ exception from benchmarks and performance ratings made public by Chief Sustainability Officer.”

Occupancy classification no longer decides whether the ordinance applies

The letter S stays. The exclusion list still reads Group A-5, F, H, S, or U. The substitute adds an override.

Provided, however, the term “covered building” shall include any Group 1 covered building or Group 2 covered building with a data center.

Under the introduced version, coverage turned on whether the building code assigns data centers to Group S under Chapter 14B-3. GrayStak asked the Department of Buildings that question on August 13. The department has not answered. The city no longer needs to.

Chapter 18-14 already defines the term, and the substitute displays the existing definition without touching it. A data center is a space specifically designed and equipped to meet the needs of high-density computing equipment such as server racks, used for data storage and processing, as defined by the benchmarking tool.

Hadden’s committee wrote on August 18 that the ordinance was drafted to reach data centers, which “had not been contemplated in the initial benchmarking ordinance.”

Two words are deleted from the disclosure exception

Section 18-14-102.3 is titled “Disclosure and rating.” Subsection (g) authorizes the Chief Sustainability Officer to make reported benchmarking information and energy performance ratings publicly available. An exception follows it.

That exception reads:

Unless a different result is mandated by applicable law, the Chief Sustainability Officer shall not make readily available to the public any individually attributable reported benchmarking information or energy performance rating pertaining to a covered building that contains a data center, television studio, or trading floor that together exceed 10 percent of the gross square footage of any such building until the Chief Sustainability Officer determines that the benchmarking tool can make adequate adjustments for such facilities. When the Chief Sustainability Officer determines that the benchmarking tool can make such adjustments, it shall report such determination to the Mayor and the City Council.

Deleted from it are the words “data center” and the comma after “television studio.” Television studios and trading floors stay inside the exception. The determination requirement stays in the code. Neither one applies to data centers now.

The suspension has never governed only the numbers. Subsection (d)(i) withholds the energy rating card from any building exempt from disclosure under 18-14-102.3, and subsection (e) puts that card where the public can see it. An unmade determination keeps the rating card off the building.

Angela Tovar holds the office that can make it. She is Chief Sustainability Officer and Commissioner of the Department of Environment, one person in both roles. She convened the working group under Council ordinance SO2024-0008201 and co-chairs the task force created by Executive Order 2026-5.

GrayStak asked the Department of Environment on August 13 and again on August 21 whether it has ever made or reported that determination. The department has not answered either request.

The substitute amends three sections of Chapter 18-14. Underlined text is added, struck text is deleted. Source: City Clerk, SO2026-0027165, published August 26, 2026.

Two city offices described the same document differently

Gina Parra-Hughes, deputy director of policy and outreach for the committee, wrote on August 21 at 11:34 a.m. that the substitute would address “definition and disclosure and rating” and would amend Section 18-14-102.3. She said the Department of Environment was working with the committee on the language and that no copy was available for public distribution.

Asked in the same email whether the Council has ever received a determination under 18-14-102.3, she wrote that she did not have a yes or no answer, and that the question “is something the Chief Sustainability Officer’s office would be better positioned to confirm.”

On the evening of August 25, a spokesperson for the Mayor’s Press Office said that while the Chief Sustainability Officer has jurisdiction over the covered buildings list, the Municipal Code “currently exempts data centers from reporting their energy and water usage,” and that amending the Code “requires City Council action.”

“Amending the Municipal Code requires City Council action.”

Mayor’s Press Office spokesperson, August 25

The Code exempts certain buildings from the definition of a covered building, which governs reporting. A separate section suspends disclosure for buildings containing data centers, and lifting that suspension takes no Council vote. The substitute amends both sections.

Section 18-14-102.3(g) authorizes the Chief Sustainability Officer to publish the data once she makes and reports a determination. The substitute repeals the exception by ordinance. Source: Municipal Code 18-14-102.3(g); SO2026-0027165.

The posted agenda describes the ordinance the substitute replaced

The Agenda of Matters for the September 1 meeting was posted on August 27, a day after the Clerk published the substitute.

It lists the item under the substitute’s record number and describes it as an “Amendment of Municipal Code Section 18-14-101.3 to remove exemption for buildings where more than 10 percent of floor area is classified as Group S from definition of covered building.”

That is what the July 15 ordinance did. The substitute does not remove Group S from the exemption. It leaves the list intact, adds an override for buildings with data centers, and amends two further sections, 18-14-101.4 and 18-14-102.3. The Clerk’s own caption for the matter describes all three.

The committee will take votes on the ordinance. Two other items are listed: approval of the committee’s Rule 45 report for July 13, and a subject matter hearing on urban flooding in Chicago, on which no votes will be taken.

Public comment runs at the start of the meeting, three minutes per speaker and thirty minutes in total. According to the notice, everyone who signs up may not get to speak.

Twelve of the fifteen members voting already sponsor the bill

Twenty-two members are on the ordinance: Hadden as sponsor and twenty-one co-sponsors. Twelve sit on the fifteen-member committee that votes September 1.

They are La Spata, Ramirez, Lopez, Tabares, Cruz, Conway, Knudsen, Lawson, Clay, Martin, Manaa-Hoppenworth and Hadden. Sigcho-Lopez, Rodriguez Sanchez and Reilly have not signed. Reilly represents the 42nd Ward, which his office describes as Chicago’s central business district.

The co-sponsor list crosses the Council’s factions. It includes Gilbert Villegas, the Zoning Committee chair who called the mayor’s data center executive order a political stunt the day after it was signed.

Committee membership verified against the City Clerk’s legislative body record on August 27, 2026. Source: City Clerk matter record and committee roster.

The mayor deferred on the determination and left the moratorium to the Council

At a press availability on August 25, the mayor was asked whether the Chief Sustainability Officer has made and reported the determination required by 18-14-102.3. “I can get back with you on that one,” he said.

His office answered that evening with the statement above. It did not say whether they had made the determination.

He was also asked who would introduce the temporary moratorium on new data center development that he called for on August 11.

“Listen, that is their function,” he said. “City council has the authority to move on a moratorium. I am encouraging them to do that. But if you are asking me to hand-pick a leader, I hope that they are capable of self-selecting in that regard.”

On August 17, his press office had written that the administration “will be working with our partners in City Council to finalize language for the ordinance and plan to introduce it at an upcoming City Council meeting.” No ordinance establishing a data center moratorium appears in the City Clerk’s legislative record as of August 26. Searches of that record on the terms “data center” and “moratorium” return none, and an ordinance captioned purely as a code amendment would not surface in either.

Four dates are already fixed

If the committee reports the ordinance out, it goes to the full City Council, which returns September 23 at 10 a.m.

Executive Order 2026-5 sends the working group’s three recommendations to an interdepartmental task force reporting on or before March 15, 2027, or a later date approved by the Mayor. Air pollution and noise rules under Sections 1 and 2 of the order are due November 9, subject to extension by the Mayor. The Illinois Commerce Commission must act on ComEd’s $15.3 billion grid plan by 5 p.m. on December 15.

GrayStak will report from the September 1 committee meeting.

HOW THIS WAS REPORTED

This piece is built from primary documents.

City records: substitute ordinance SO2026-0027165, filed with the City Clerk and published August 26, 2026, at 3:15 p.m.; ordinance O2026-0027165 as introduced July 15, 2026, and its signature page; the Clerk’s matter record including its action history, sponsor list and attachments, retrieved August 26, 2026; the notice of meeting for the September 1 Committee on Environmental Protection and Energy, dated July 22, 2026 and signed by Chair Maria E. Hadden; the Agenda of Matters for that meeting, posted August 27, 2026; the Clerk’s Legislative Body and Member Report generated July 24, 2026; Executive Order 2026-5; and the Sustainable Data Centers Report dated July 9, 2026.

Municipal Code: Sections 18-14-101.3 and 18-14-102.3, and Chicago Building Code Section 14B-3-302, as published in supplement 2026 S-73.

This piece continues GrayStak’s reporting on Chicago’s data center regulation, which began with Chicago Zones Data Centers as Offices, published August 17, 2026 and corrected August 26.

Written correspondence with the Committee on Environmental Protection and Energy dated August 18, August 21 and August 26, 2026, and with the Mayor’s Press Office dated August 17 and August 25, 2026.

Legislative filings were checked against the City Clerk’s electronic legislative management system on August 26, 2026. That search matches the title and nickname fields of each matter.

WHAT IS NOT YET IN THE RECORD

The Department of Buildings has been asked what occupancy group a data center without routine human occupancy receives under Section 14B-3-302. It has not answered. Whether the Chief Sustainability Officer has ever made or reported a determination under Section 18-14-102.3 has been requested from the Department of Environment twice and from the Mayor’s Press Office once, and has not been provided.

Four requests for comment were sent on August 26, with responses requested by September 2. They cover the water usage provision, the determination, and the occupancy classification question. None had been answered at publication.

The substitute also amends Section 18-14-101.4, which governs water usage data. GrayStak has requested comment on that provision and will report it separately.

COMMENT

Written requests for comment were sent on August 13, 2026 to the Mayor’s Press Office, the Department of Environment, the Department of Buildings, the Department of Law, the Department of Planning and Development, Alderman Gilbert Villegas, Alderman Derrick Curtis and Alderman Maria Hadden. The Mayor’s Press Office answered on August 17 and relayed answers from the Department of Planning and Development the same day, and provided a statement on August 25. Alderman Hadden’s office answered on August 18, August 21 and August 26. The Department of Law responded on August 14 that it does not comment on pending litigation.

The Department of Environment did not respond to requests sent August 13 and August 21 about Section 18-14-102.3. The Department of Buildings did not respond to a request sent August 13 about Section 14B-3-302. Further requests were sent to both departments on August 26. The Department of Planning and Development on its own behalf, Alderman Villegas and Alderman Curtis did not respond.

CORRECTIONS

GrayStak corrects errors on the record and marks the change. Corrections are dated and the corrected text is identified. Send corrections to christopher@graystak.com.

A correction was issued August 26, 2026 on the previous piece in this series, Chicago Zones Data Centers as Offices. It described the occupancy classifications excluded from coverage under Municipal Code 18-14-101.3 as assembly, factory, high hazard or utility. The Code names Group A-5, one of five assembly subgroups under Chicago Building Code Section 14B-3-302, rather than assembly generally. Buildings more than ten percent A-1, A-2, A-3 or A-4 remain covered.

UPDATES

Responses received after publication, and the outcome of the September 1 vote, will be added here and dated. The original text stays intact.

DISCLOSURES

GrayStak uses commercially available artificial intelligence tools, including Anthropic products, in its research and production workflow, and holds no financial interest in any of their makers.

ABOUT THE AUTHOR

Christopher Sweat covers Chicago and Illinois government from the public record, reporting local social and political dynamics for a national and international audience. He also reports on national security policy, intelligence oversight, public finance, and the defense technology sector. He is an analyst working across finance, technology, and politics, and CEO and Co-Founder of GrayStak.

GrayStak is a media and political risk company. Our real-time Domestic Political Volatility Index (DPVI) measures the velocity and volatility of U.S. domestic politics. We cover political and institutional risk across the United States, centered in Washington, D.C., New York, and Chicago.

What happened. Johnson signed Executive Order 2026-5 on August 11, tightening city review of data centers and asking the City Council to pass a temporary moratorium he cannot impose himself.

Why it matters. The Department of Planning and Development says the order does not reach data centers the zoning code already permits. The mayor’s office says existing facilities are not affected.

What is pending. The Council returns September 23 with no ordinance drafted. City rules are due November 9. A task force reports in March 2027. State regulators decide ComEd’s $15.3 billion grid plan in December.

A data center is a building full of computers. It employs few people, runs cooling equipment twenty-four hours a day, tests diesel generators monthly, and consumes electricity and water in quantities large enough to raise what other customers pay. Chicago has about thirty-nine of them, most downtown or along the utility corridors that feed downtown.

ComEd’s chief executive, Gil Quiniones, told the Sun-Times editorial board in June that the utility was holding roughly a hundred data center applications totaling 33,000 megawatts, about one and a half times the highest demand the system has ever recorded, and that residential bills have risen about $12 a month since 2022. The Sustainable Data Centers Report says a typical data center uses 300,000 gallons of water a day, and the largest use five million. One in five Chicago households already spends more than six percent of its income on gas and electricity.

Johnson signed the order at Northerly Island, and the Clerk stamped it in at 11:10 that morning. It is titled “Protecting Chicagoans’ Clean Water, Clean Air, Energy Affordability, and Quality of Life,” and the words data center do not appear in it. The order concedes its own limit in the recitals, stating that the City Council holds “the exclusive legislative authority to enact a temporary moratorium on new data center development and material expansions.” Section 5 describes the document as “intended solely to provide interim executive direction pending legislative action by the City Council.”

The Council does not meet in regular session until September 23. Asked on August 17 whether an ordinance had been drafted and who would sponsor it, the Mayor’s Press Office said the administration “will be working with our partners in City Council to finalize language for the ordinance and plan to introduce it at an upcoming City Council meeting.” It named neither a sponsor nor a date.

The order does not reach data centers the code already permits

The order applies to Planned Developments, the negotiated zoning approval that large projects go through before the Plan Commission and the City Council. Most Chicago data centers never go through it.

The Department of Planning and Development said so on August 17, in answers relayed by the Mayor’s Press Office. “Data centers that are allowed as-of-right, including within previously approved PDs, are not subject to the executive order,” the department said. That covers two groups: data centers permitted outright under the zoning code, and data centers inside planned developments the Council approved years ago, where the use is already allowed.

“Data centers that are allowed as-of-right, including within previously approved PDs, are not subject to the executive order.”

Department of Planning and Development, August 17

A data center enters the Planned Development process for reasons unrelated to what it is. The Sustainable Data Centers Report, completed July 9, lists the triggers as lot size, building size, proximity to the river, or the presence of other uses in the same building. Being a data center is not one of them. Figure 4 of the report sorts all thirty-nine facilities by the path each took. Ten stand alone: two in Planned Developments, seven downtown, and one in a manufacturing district. Twenty-nine occupy mixed-use buildings: nineteen in Planned Developments, eight downtown, and two in manufacturing.

Only Planned Developments require a Council vote. Source: Sustainable Data Centers Report, Figure 4; Department of Planning and Development.

The press office added that an existing facility planning to expand would be subject to the standards the order produces. It did not explain how that applies to an expansion the code already allows without review. The department also said one application is pending. Planned Development 675 has carried an amendment application since 2022 that includes a proposed data center use. The department did not say where that site is or who filed it.

Sections 1 and 2 are not limited to new construction, but neither has produced a rule yet. Section 1 touches existing operations once, requiring compliance with the new conditions to be shown in applications for annual certificates of operation under Section 11-4-660. It applies only “to the extent that a data center is required to apply.” The Sustainable Data Centers Report says emergency and standby generators are exempt from that annual requirement.

Commercial trackers count far more than thirty-nine because they measure the metropolitan market rather than the city. Datacentermap lists 152 facilities under Chicago and 240 across Illinois; Datacenters.com lists 166; and Baxtel lists 212, while ranking Chicago the seventh-largest market in the country and naming Microsoft’s 158-megawatt Northlake facility the largest in it. Northlake is a suburb in western Cook County. Chicago counted only inside its own boundaries, using building permit records supplemented by one directory, and put the greater region above 200.

Chicago’s zoning code classifies a data center as an office

Section 17-17-0104 of the zoning ordinance covers the Commercial Use Group, which the code describes as uses that provide a business service or involve selling, leasing or renting merchandise to the general public. Subsection U is Office: “Professional, governmental, executive, management or administrative offices of private organizations or government agencies. Typical uses include government offices, administrative offices, legal offices, and architectural firms. Also includes electronic data storage centers and high-technology offices.”

One line down sits the only definition the code carries: “Electronic Data Storage Center. A work site used as a facility for the storage of and the operation of computer hardware, equipment for processing, storage and/or routing of electronic data, or other high technology uses.” That is the entire treatment.

The use tables show the city already knows the difference. Every use in the code is assigned a parking group, which determines how many spaces a building must provide. In Table 17-3-0207, which governs the business and commercial districts, the Office row carries parking group 17-10-0207-L. The Electronic Data Storage Center line beneath it carries a different one, 17-10-0207-U. Chapter 17-10 titles that group “Commercial Garden or Greenhouse, Electronic Data Storage Centers, Industrial.” Its other occupants are manufacturing at every intensity, all six classes of recycling facility, warehousing, freight terminals, and every waste-related use through hazardous materials storage.

Group L sets parking by floor area, while Group U sets it at one space per four employees, which means the code already treats a data center as a building with a large floor area and very few people in it. Chicago classifies the use as an office and applies the industrial parking standard to it.

The code lists data centers as a subcategory of Office and assigns them the industrial parking standard. Source: Chicago Zoning Ordinance, Table 17-3-0207 and Chapter 17-10.

The manufacturing districts go further. In Table 17-5-0207, the Office row reads “In M2 and M3, max GFA: 9,000 sq ft or accessory use to allowed industrial use.” The Electronic Data Storage Center line directly beneath it carries no use standard at all. Inside a manufacturing district, a data center faces fewer limits than an insurance office.

Downtown, the code does not break them out at all. Table 17-4-0207 carries a single Office row, permitted by-right in the downtown core, mixed-use and service districts, and barred from downtown residential. No Electronic Data Storage Center line runs beneath it, and it has no separate parking standard. Fifteen of the city’s thirty-nine data centers sit in those districts.

Data centers are permitted by-right in every manufacturing district; in B3, C1, C2, and C3; in the planned manufacturing districts; and in the downtown core, mixed-use, and service districts. The exclusions are B1, B2, and downtown residential. A facility that clears zoning by-right takes a building permit from the Department of Buildings and a Limited Business License from Business Affairs and Consumer Protection. No hearing. No conditions. No vote.

Section 17-17-0104 has been amended repeatedly since 2004, most recently in December 2025. The Council did not touch it in September 2024, when it passed the ordinance that created the working group.

The ordinance that ordered the study also paid to attract data centers

That ordinance did something else at the same time. SO2024-0008201, sponsored by Aldermen Gilbert Villegas and Brian Hopkins and passed September 18, 2024, added Section 2-68-050 to the Municipal Code, requiring city data to be processed and stored within the United States and giving vendors a 2 percent bid credit for keeping it in Illinois and 4 percent for keeping it inside Chicago.

Villegas described it at the time as a tool to bring in new revenue, good jobs, and economic development to communities that need it most, and said data centers need wide swaths of land that the South and West Sides have. The Chicagoland Chamber of Commerce opposed it, saying the residency rules would give data centers a reason to cross Chicago off their list.

A single ordinance created a procurement incentive to bring data centers to Chicago and commissioned the study of what they do once they arrive.

Villegas now chairs the Zoning Committee, and the day after the order was signed, he described the buildings plainly. “At the end of the day, a data center is a warehouse with servers in there,” he said. The code calls it an office. He also said the mayor should have introduced an ordinance rather than signing one, calling the order “just another political stunt by the mayor to say that he did something, given the fact that data centers right now are unpopular.”

“At the end of the day, a data center is a warehouse with servers in there.”

Ald. Gilbert Villegas, chair of the Zoning Committee

The order sets three deadlines and the mayor can extend all three

Section 1 tells the Chief Sustainability Officer to impose conditions on air pollution control permits for regulated equipment at data centers under Code Section 11-4-630, and to promulgate rules for those permits under 2-31-070. The rules are due “within ninety (90) days of the effective date of this Executive Order (subject to further extension authorized by the Mayor).”

That authority is new to her department. The Mayor’s Press Office said the Department of Public Health participated in the working group through its Environmental Permitting and Inspection office, and that the office transferred to the Department of Environment on May 1.

Section 2 tells her to review the existing noise framework at 8-32-090 as it applies to mechanical equipment at data centers, and to publish rules for administering and enforcing that section. It carries the same ninety days and the same extension clause.

Section 4 creates the Interdepartmental Data Center Environment and Community Protection Task Force, co-chaired by the Chief Sustainability Officer and the Mayor’s Chief of Policy, with Buildings, Environment, Planning and Development, Technology and Innovation, and Water Management. It is to deliver recommendations and draft code amendments “on or before March 15, 2027, or such other date approved by the Mayor that is consistent with the timing and duration of any data center development moratorium approved by the City Council.”

The order sets three deadlines and gives the Mayor the power to extend each of them. Sections 1 and 2 are due within ninety days of August 11, which is November 9, and both say “subject to further extension authorized by the Mayor.” The task force date of March 15, 2027 can be replaced by “such other date approved by the Mayor.” Nobody outside the fifth floor has to agree.

Asked whether the administration expects to meet the November 9 date, the Mayor’s Press Office said the Chief Sustainability Officer plans to meet it.

Johnson’s January order carried a 30-day deadline and ran five months past it

Executive Order 2026-01 gave the Police Department thirty days to issue guidance and establish procedures for documenting federal immigration agents. It contained no extension clause. On July 17 the Committee on Immigrant and Refugee Rights took the order up, and police officials said the department had finalized internal procedures nearly five months after the deadline. They would not say why it took that long.

“It is long past time for accountability to the public,” said Alderman Andre Vasquez, the committee chair. Counsel to the mayor Sheila Bedi told the same committee that the January order “did not invent” the Police Department’s authority to document and investigate criminal activity. A lawyer for the administration was describing an executive order as a restatement of powers the city already held.

The January order “did not invent” the Police Department’s authority to document and investigate criminal activity.

Sheila Bedi, counsel to the mayor, July 17

The August order works the same ground. Air pollution control permits for diesel generators already exist under 11-4-620. Stationary mechanical equipment is already limited under 8-32-090 to 55 decibels measured at 100 feet between 8 p.m. and 8 a.m. Enforcing that section requires an inspector on site during the regulated hours, proof of the source, and proof of each separate day of violation. Fines run from $1,000 to $5,000.

The order’s only zoning provision cancels itself

Section 3 is the only part that touches development approvals. It directs the Planning and Development Commissioner to give enhanced review to every proposed Planned Development that includes a data center, weighing projected electricity demand, water consumption, backup generation, air emissions, noise, cumulative environmental impacts, transportation, emergency response capacity and neighborhood effects.

On receiving such an application, the commissioner notifies Buildings, Business Affairs and Consumer Protection, Environment and Water Management. Each has 60 days, or another period the commissioners arrange, to return written recommendations. Any conditions they identify go into the proposed ordinance.

The last paragraph of the order rescinds it. Section 3 “shall be automatically rescinded upon the effective date of an ordinance enacted by the City Council establishing a temporary moratorium on the approval of new Data Centers or the material expansion of existing Data Centers.”

The working group asked for three ordinances and got a task force

Every recommendation in the report requires the City Council. The report asks the Council to authorize an environmental permitting program specific to data centers, administered by the Department of Environment, with impact studies, air and sound monitoring, a fee structure covering the full cost of the program and a fine structure for violations.

It asks for a zoning amendment requiring most Electronic Data Storage Centers to undergo Planned Development review with Council approval. And it asks the Council to remove the exemption barring public release of energy and water data for buildings containing data centers.

The order routes all three to the task force, which is to draft proposed amendments by March 15, 2027, or later. Asked whether the administration would introduce ordinances implementing the recommendations instead, the Mayor’s Press Office said the recommendations will be routed through the task force.

The Council has heard the report once. On July 13 the Committee on Environmental Protection and Energy took it up as a subject matter hearing, which is a briefing with no legislation attached, and the minutes record no vote.

Commissioner Angela Tovar presented with First Deputy Commissioner Jared Policicchio and Jared Patton, and Bradley Roback from Planning and Development attended to answer questions on zoning and land use. Sarah Moskowitz of the Citizens Utility Board, Andrew Rehn of Prairie Rivers Network, Kavi Chintam of Vote Solar and Craig Huffman of MetroEdge also presented. Metro Edge is one of the five Chicago data centers holding a state tax credit, at $257 million committed against $16 million in benefits and 20 jobs. Five people spoke in public comment, and the meeting ran two hours and forty-four minutes.

Section 2 of SO2024-0008201 required that report, and it went to two committees. The Committee on Economic, Capital and Technology Development, which Villegas chaired when he sponsored the ordinance and which Alderman Derrick Curtis has chaired since May, filed no Rule 45 report for July, the monthly summary of proceedings each committee owes the Clerk. Seven other committees did.

The disclosure rule can be lifted without the Council

The third recommendation does not require an ordinance at all. Chicago requires large buildings to report how much energy and water they use, and publishes the results address by address. Municipal Code 18-14-102.3 suspends that publication for buildings where a data center, television studio and trading floor together exceed 10 percent of gross square footage. The suspension runs “until the Chief Sustainability Officer determines that the benchmarking tool can make adequate adjustments for such facilities.” The code then requires her to report that determination to the Mayor and the City Council.

Angela Tovar holds that power. She convened the working group under Council Ordinance SO2024-0008201. She is quoted in the release announcing it. She co-chairs the task force the order creates, and Sections 1 and 2 assign the rulemaking to her.

Whether she has evaluated the question is a matter of record, because the code requires a report when she answers it. No such report appears in the public file.

The one ordinance on file does something else. Alderman Maria Hadden introduced O2026-0027165 two days after that hearing, on July 15, and it sits in her own committee. It runs two sentences. It amends the definition of a covered building at 18-14-101.3 by striking a single letter from the list of occupancy classifications that disqualify a building from coverage.Buildings where more than a tenth of the floor area is classified as Group A-5, F, H or U stay outside the ordinance. Group A-5 is one of five assembly subgroups under Section 14B-3-302 and covers outdoor assembly. Assembly uses in the other four subgroups remain covered. Storage, which the building code calls Group S, comes in.

That widens the set of buildings covered by the reporting requirement, but it leaves Section 18-14-102.3 untouched, so the bar on releasing the numbers stands whether it passes or not.

Hadden’s office said on August 18 that the ordinance was drafted to reach data centers, which “had not been contemplated in the initial benchmarking ordinance.” The office said it is rewriting the ordinance with city departments to produce substitute language aimed at increasing transparency around electricity and water use by data centers, and that the substitute comes up for a vote in the Committee on Environmental Protection and Energy on September 1. It said no other data center legislation is pending before the committee.

Asked whether the Council has received any determination from the Chief Sustainability Officer under Section 18-14-102.3, the office said the Department of Environment is working with it on the language. It did not say whether a determination has been made.

The office also said Hadden will likely co-sponsor a moratorium ordinance and is in communication with other Council members and the mayor’s office, and that the March 2027 task force deadline is acceptable to her committee.

Twenty-one members co-sponsored it, and the list crosses the Council’s factions. Lopez, Sposato, Conway, Knudsen, Tabares and Lee signed alongside Fuentes, Taylor, Ramirez, Quezada, Clay and Vasquez. So did Villegas.

The Council he is asking filled its ballot slots three weeks earlier

On July 22 the Council placed six referendum questions on two ballots and used every available slot. Illinois allows three public questions per government per election, and the first three validly initiated take the ballot. November 3 carries gunshot detection restart, festival and event space, and Midway Blitz federal compensation. February 23 carries a music and arts monument, independent council counsel, and sister-agency debt enforcement.

The mayor’s millionaire tax question made neither ballot. The November version needed adoption by August 16 against a slate already full. Cook County commissioners then placed a parallel millionaire tax question on the same November ballot, which every Chicago voter will see.

Three weeks later he asked that same Council for a moratorium.

Aurora defined the term before it wrote the rules

Aurora, 40 miles west, had the same classification problem. Its code treated data centers as a type of warehouse, and city officials say the facilities looked like warehouses anyway. All ten of its existing data centers were reviewed through warehouse zoning.

“We had to have some more specific regulations for these,” said Tracey Vacek, Aurora’s planning and zoning director. The City Council there enacted a 180-day moratorium on new data centers and warehouses on September 25. It expired March 24, and the Council adopted the replacement rules the same day. New facilities now need conditional use approval, which carries a public hearing, notice to nearby residents and a Council vote. Applicants file water use studies, energy management plans and sound assessments. Maximum sound levels came down. Every data center facility in the city files an annual report with the city by April 1.

The first thing the ordinance does is define the term. Alison Lindburg, Aurora’s sustainability director, said the rules cover zoning and also create a definition for a data center to address transparency and building permits. Before, when a data center could apply as a warehouse, the permit process was quick.

The report records what Chicago residents told the city at a May 19 meeting at the Arturo Velasquez Institute, where about 75 people attended. “Data centers shouldn’t be by-right,” one wrote. “Should have a process that involves community.” Another asked for air, water and health protections citywide, “not just permit-by-right for data centers.”

The report calls the largest data centers unlikely here. One was announced in 2025

The report states that Chicago is unlikely to see development of the largest hyperscale AI training campuses because of land and infrastructure constraints. The order repeats the framing, reciting that the working group found continued growth likely “notwithstanding Chicago’s land and infrastructure constraints.”

Related Digital announced a one-gigawatt data center for the South Side in March 2025, at the north end of the Quantum Shore development on the former U.S. Steel South Works site. That is ten to twenty times the report’s own working definition of hyperscale.

The Council approved the Planned Development in December 2024, covering more than 400 acres between 79th Street and the Calumet River at 8080 S. DuSable Lake Shore Drive. The zoning reclassification, SO2024-0012887, was introduced on October 9 and passed as a substitute on December 11. The Clerk’s record classifies it as routine.

The Plan Commission had cleared it the month before after a hearing at Harold Washington College that ran close to six hours, with more than 70 people signed up to speak and most of them opposed. Community groups asked for a community benefits agreement, and neither the developers nor the mayor committed to one.

At that November hearing, Related Midwest said it was in conversations with a major employer for the north end of the site, separate from the quantum campus. Four months later, Related Digital announced the data center for the north end.

The site permits data centers across most of its area. The exception runs west of DuSable Lake Shore Drive between 83rd and 87th streets. Planning officials said in early August that the developers have not expressed an intention to move forward.

A spokesperson for the Department of Environment said in early August that the report “responds to the realities of modern data centers in Chicago, not one specific project.”

The city is already fighting over data center costs at the commission

The order addresses the buildings, while Chicago households' costs are decided in a separate proceeding that concludes months earlier.

ComEd filed its multi-year integrated grid plan for 2028 through 2031 with the Illinois Commerce Commission on January 16, docket P2026-0047, at $15.3 billion and more than 800 pages. A grid plan is the four-year investment program state law requires the commission to review before the company builds and bills for it.

The filing proposes three new substations in Chicago and two elsewhere in northern Illinois, with added capacity at more than seventy existing ones. The filing tells the commission that load growth is localized primarily across 55 substations, that the company is tracking new load additions exceeding 1,900 megawatts, and that data center load accounts for over 47 percent of that growth. If approved in full, ComEd expects residential customers to pay $2.50 to $3 more per month starting in 2028.

The City of Chicago is an intervenor and party of record. Dori McDonald, chief assistant corporation counsel in the Department of Law, appears for the city, alongside Natalie Karas of Duncan & Allen in Washington. The service list runs to 112 entries and includes the Citizens Utility Board, the attorney general’s Public Utilities Bureau, the Illinois EPA, the Natural Resources Defense Council, the Environmental Law and Policy Center, Vote Solar, the Union of Concerned Scientists, the Environmental Defense Fund, the Building Owners and Managers Association of Chicago, and the Village of Dolton.

Two of those organizations were in Council chambers on July 13. Sarah Moskowitz presented for the Citizens Utility Board and Kavi Chintam for Vote Solar. Both groups are also litigating the cost question at the commission.

The city has filed seven times since entering its appearance on January 28. It put on three witnesses. Samantha Bingham, clean transportation program director at the Department of Transportation, testified about multifamily housing strategy and a curbside electric vehicle charging pilot. Taylor Mayes, a project manager at the Department of Environment, testified about aligning the grid plan with the city’s Climate Action Plan, joint gas and electric planning with Peoples Gas, workforce metrics and a multifamily solar pilot.

Mayes filed her rebuttal on July 16. Three days earlier her commissioner told the City Council that data center growth is driving up electricity prices for Chicagoans.

The third witness addressed the plan's largest sum. Joseph McHugh, assistant commissioner for GIS and utilities at the Department of Aviation, challenged Chapter 7 of the grid plan, which covers what ComEd calls its New Business: Unique Commercial and Industrial Program. That is the money the company spends connecting large new customers to the grid. Table 7-2 puts it at $1,160.3 million in capital investment and $4.8 million in operations and maintenance.

McHugh’s direct testimony describes what that money buys. Quoting the grid plan, he says the program covers customer-driven projects over $100,000 that install new electric service, upgrade existing service and extend circuits to new load centers. Then his own sentence: “This category includes the construction and expansion of data centers and also other high-density load projects.”

“This category includes the construction and expansion of data centers and also other high-density load projects.”

Joseph McHugh, Chicago Department of Aviation, testifying for the city

His rebuttal goes further. The process ComEd describes in Chapter 7, he says, “never even mentions the Cluster Study process, which is an important phase that all data center projects must undertake.” A cluster study is the engineering review the utility runs before it agrees to connect a large new load. He asks for protocols written into the tariff, with firm timelines for the utility’s actions and a dispute resolution mechanism customers can use. The chapter as written, he says, contains none of those safeguards.

ComEd moved on June 18 to define the scope of the docket and put that testimony outside it, arguing the issues belonged to a rate design investigation that has already closed. The city answered a week later through Corporation Counsel Mary B. Richardson-Lowry, saying the motion “would in effect foreclose the City from raising these issues altogether,” and that the closed investigation “did not, and cannot, resolve the City’s concerns.” The attorney general, the Citizens Utility Board, Constellation Energy Generation, a coalition of environmental groups and an individual intervenor all opposed the motion the same week.

The administrative law judges denied it on July 2. The Public Utilities Act, they wrote, makes the grid plan a forum in which the commission, the utility and stakeholders must consider all related investments, planning processes, tariffs, rate design options, programs and other utility policies and plans. Striking a party’s arguments at that stage would run against the integrity of the fact-finding process.

The city’s testimony on data center connection costs stayed in the case, and the evidentiary hearings went ahead.

The commission decides in December. The task force reports in March

ComEd’s last round of written testimony was due July 31. The administrative law judges held a status hearing on August 10 and took evidence on August 11 and August 12. Written arguments are due September 3 and September 17, and each side files a proposed order on September 18. The judges issue their own proposed order on or about October 14. Objections to it are due October 28, with replies November 10. The commission must act by December 15, 2026, at 5 p.m.

The evidentiary hearing on August 11 was called for 10 a.m. That morning, the city filed a motion to admit its evidence. The Clerk stamped the executive order in at 11:10.

The task force created that morning reports by March 15, 2027. The administrative law judges issue a proposed order in October. The commission decides in December.

The mayor's deadlines against the commission's schedule. Sources: Executive Order 2026-5; ICC docket P2026-0047.

The Mayor’s Press Office said the city is a party to the docket and that the task force deadline was set separately. Asked whether the task force would coordinate with the case and whether the Department of Law or the Department of Aviation would participate, the office said the task force “will be closely monitoring the results of that docket given its impact on energy issues related to data centers.” Neither department was added to it.

Kristen Cabanban, director of news affairs at the Department of Law, said the department does not comment on pending litigation.

Sarah Moskowitz, executive director of the Citizens Utility Board, made the same point about the order the day it was signed. “If we’re just looking at the impacts that data centers have on our electricity costs, there’s not much the city of Chicago can do,” she told the Sun-Times and WBEZ. “When it comes to noise and local air quality and land use, of course that is all really important, and that is the kind of things that require transparency, but we’re going to need state action to address the electricity cost portion of the data center problem.”

Moskowitz has described the mechanism in earlier interviews. Developers apply to ComEd for service a project will need. When the project is built somewhere else, the utility has bought capacity nobody uses, and the cost moves to other customers. She has called those projects phantom data centers.

The Data Center Coalition, the industry’s membership association, said its members “strive to be good actors where we operate, which is why we are hopeful this directive from the Mayor will help separate those who are responsible water and energy stewards from those who are not.”

What has to happen next

The order covers the first two boxes. Sources: Executive Order 2026-5; Municipal Code; ICC docket P2026-0047; Illinois General Assembly.

Three ordinances have to move through two committees. Hadden chairs Environmental Protection and Energy. Villegas chairs Zoning. The first opportunity is September 23.

Villegas said the day after the signing that industry wants stability, needs to know how and under what conditions it can build, and that he believes a proposal could be ready for a vote before the end of the year. He also said energy costs could have been addressed years ago through a new franchise agreement with ComEd.

The provisions covering ratepayers and water sit in Springfield, where the POWER Act failed in the spring and returns in the veto session running November 17 through 19 and December 1 through 3. Pritzker paused the state’s data center tax incentives on June 5 and issued a framework asking the General Assembly to require data centers to report their energy and water use, to post public notice when they apply for permits, and to stop signing nondisclosure agreements with local governments.

Climate Jobs Illinois and the Illinois AFL-CIO called the pause shortsighted and said it was issued to generate headlines instead of practical results. A separate bill strengthening local control over data centers never passed either chamber.

Water reuse sits with the Metropolitan Water Reclamation District, which received pilot authority under a state law signed last August.

Chicago also sells Lake Michigan water to suburbs hosting data centers it does not regulate. Aurora, West Chicago and Lake County have acted. Cook County and DuPage County, which hold most of the region’s facilities, have not.

Joliet shows what happens where the buildings are larger. Its City Council approved the annexation and rezoning of 795 acres for a 24-building, 1.8-gigawatt campus in March, on an 8-to-1 vote after seven hours of testimony that ran mostly against.

In May, three residents and a group called Joliet Residents for Responsible Growth sued in Will County Circuit Court to void the approvals, case 2026CH000108. The complaint says a Joliet staff memo acknowledged the mechanical equipment would be audible two miles away, and that the city approved the campus while spending more than a billion dollars to switch from a depleting aquifer to Lake Michigan water. Joliet is buying Chicago’s water.

Two dates will show whether any of this binds. The air permit and noise rules are due November 9, and the administration says the Chief Sustainability Officer plans to meet that date. The Council must introduce a moratorium ordinance before it can pass one. Six days after signing, the administration said it was still finalizing language and did not name a sponsor.

How this was reported

This piece is built from primary documents.

City records: Executive Order 2026-5, signed and filed with the City Clerk on August 11, 2026, at 11:10 a.m.; Executive Order 2026-01, filed February 28, 2026; the Sustainable Data Centers Report of the City of Chicago Sustainable Data Centers Working Group, dated July 9, 2026; the Monthly Rule 45 Report of the Committee on Environmental Protection and Energy for July 2026; ordinance O2026-0027165 and its signature page; ordinance SO2024-0008201; ordinance SO2024-0012887; and the Clerk’s council and committee report library.

Municipal Code: Chicago Zoning Ordinance Sections 17-17-0104 and 17-3-0200, Tables 17-3-0207 and 17-5-0207, and the off-street parking groups at 17-10-0207; Sections 18-14-101.3 and 18-14-102.3; Sections 11-4-620, 11-4-630 and 11-4-660; Section 8-32-090; and Chicago Building Code Sections 14B-3-302 and 14B-3-304, all as published in the current supplement.

Illinois Commerce Commission docket P2026-0047: the docket sheet, service list, and schedule; the city’s notice of appearance, direct and rebuttal testimony of Joseph McHugh, Taylor Mayes and Samantha Bingham, response of June 25 to ComEd’s motion to define scope, and motion to admit evidence of August 11; ComEd’s motion of June 18; and the administrative law judges’ ruling of July 2.

Legislative introductions were checked against the City Clerk’s electronic legislative management system through August 13, 2026.

Where other outlets reported it, we attributed it. The Chicago Sun-Times, WBEZ, WGN, and ABC7 reported quotations from the mayor and Alderman Villegas at and after the signing. Aurora’s regulations are drawn from statements by city officials to WGLT and NPR Illinois and from the City of Aurora’s published materials.

What is not yet in the record

Transcripts of the August 11 and 12 evidentiary hearings in docket P2026-0047 had not been posted as of August 17. The Committee on Economic, Capital and Technology Development had filed no Rule 45 report for July 2026. The Department of Buildings has been asked to classify the occupancy of a data center without routine human occupancy, but it has not yet provided the classification. Whether the Chief Sustainability Officer has made or reported a determination under Section 18-14-102.3 has been requested and not yet provided. The location and applicant of the pending amendment to Planned Development 675 were not provided.

Comment

Written requests for comment were sent on August 13, 2026, to the Mayor’s Press Office, the Department of Environment, the Department of Buildings, the Department of Law, the Department of Planning and Development, Alderman Gilbert Villegas, Alderman Derrick Curtis and Alderman Maria Hadden. The deadline was extended to August 17 at the request of the Mayor’s Press Office, which referred several questions to the Departments of Environment and Planning and Development. The Mayor’s Press Office answered on August 17 and relayed answers from the Department of Planning and Development the same day.

The Department of Law responded on August 14 that the matters raised are the subject of pending litigation and that the department does not comment on pending litigation. A separate request for the terms of the city’s engagement of Duncan & Allen and payments made to the firm was filed under the Freedom of Information Act on August 14 and is pending. The Department of Environment did not respond to questions about whether the Chief Sustainability Officer has made or reported a determination under Section 18-14-102.3. Alderman Hadden's office responded on August 18. The Department of Buildings, the Department of Planning and Development on its own behalf, Alderman Villegas and Alderman Curtis did not respond.

Corrections

August 26, 2026. An earlier version described the occupancy classifications excluded from coverage under Municipal Code 18-14-101.3 as "assembly, factory, high hazard or utility." The Code names Group A-5, which is one of five assembly subgroups under Chicago Building Code Section 14B-3-302, rather than assembly generally. Buildings more than ten percent A-1, A-2, A-3 or A-4 remain covered. The text has been corrected.

GrayStak corrects errors on the record and marks the change. Send corrections to christopher@graystak.com.

Updates

August 18, 2026. Added a response from Alderman Maria Hadden’s office received after publication, and updated the comment section to reflect it.

Disclosures

GrayStak holds no position in any company named in this analysis, and neither does the author. GrayStak uses commercially available artificial intelligence tools, including Anthropic products, in its research and production workflow, and holds no financial interest in any of their makers.

About the author

Christopher Sweat covers Chicago and Illinois government from the public record, reporting local social and political dynamics for a national and international audience. He also reports on national security policy, intelligence oversight, public finance, and the defense technology sector. He is an analyst working across finance, technology, and politics, and CEO and Co-Founder of GrayStak.

GrayStak is a media and political risk company. Our real-time Domestic Political Volatility Index (DPVI) measures the velocity and volatility of U.S. domestic politics. We cover political and institutional risk across the United States, centered in Washington, D.C., New York, and Chicago.

A regulated utility cannot refuse a data center.

The obligation to serve is written into state law. It applies on the same terms to a shopping center, a housing development, a federal installation, and a single facility interconnecting at 700 megawatts, which is roughly twice the peak demand of Naperville, Illinois. Bill Murray, who runs state and electric public policy for Dominion Energy, told a room of state legislators and staff in Chicago on July 27 that his company does not assign blame and has to serve them all. The question he posed to the room was how to do that effectively and affordably.

Whether these facilities get built is therefore largely settled outside the legislature. What remains unsettled is how the cost of serving them is divided among customer classes. That division is decided in rate cases at state public utility commissions and in the capacity market run by the regional grid operator.

Thirty-nine states introduced legislation on data centers and power in 2026, and thirty-one on water. Thirty-eight states currently offer preferential tax treatment to data centers, and thirteen have introduced changes to that treatment. Almost all of that regulation governs the building: whether it gets a tax exemption, whether it can be sited, what it has to disclose, and whether it must bring its own generation. At least seven states have passed a law that addresses how the cost is divided.

Serving new load means building, and a utility recovers what it builds through rates with a return, subject to a commission finding that the spending was prudent and that the asset was used and useful. The generation that would serve this load at a lower marginal cost arrives in the early 2030s, a date its suppliers cite in their own securities filings. The five to seven years in between are served by additions to the existing system, and the size of those additions is the largest sum in play.

Of the four state designs described from a single afternoon of testimony in Chicago, one reached that decision.

The legislation that failed in Illinois in the spring reached the commission. The POWER Act amended the Public Utilities Act, prohibited cost shifting, and created a fund to compensate community intervenors in proceedings. The industry group opposing it counted 638 pages. It died anyway, and the governor paused tax incentives instead.

The argument is regional. It applies where anticipated load growth meets supply and delivery constraints. PJM is that case. Illinois sits inside it.

The utility cannot decline the load

The obligation to serve shifts the threshold question away from legislatures. It does not diminish the scale of what arrives.

Murray established units before making any argument: a megawatt is roughly 250 homes, and a gigawatt is a thousand megawatts, or 250,000 homes, both flagged as averages. He then gave four reference points across Dominion’s fifteen years of data center interconnections.

He noted in passing that the company had not begun this recently and that Amazon was among its earliest data center customers. He also observed that Amazon’s energy policy head was on the panel that followed. The two men described the same commercial relationship from opposite ends of it, ninety minutes apart, in different rooms.

Murray noted a corresponding shift in vocabulary. Across his first fifteen years he encountered the term gigawatt three or four times. It is now the standard unit for discussing supply.

Kim Janas, counsel to the Illinois attorney general, offered the same finding in Illinois terms. A single data center, she said, can interconnect at 700 megawatts, and she set that against Naperville, a city of roughly 150,000 in DuPage County that she has watched grow over her lifetime. Her framing was forty-four years of municipal growth against one interconnection.

Naperville operates its own municipal electric utility, serving about 63,000 customers, according to the city’s count, and buying its power through the Illinois Municipal Electric Agency. Across 32 member municipalities, its peak demand is 37 percent of the agency’s total and its energy 34 percent. Naperville Development Partnership reports the city’s peak demand at 343 megawatts.

Against either figure, the comparison holds, and against the published figure it is starker. The interconnection would draw about twice the peak of the entire city, and it would draw it differently. A municipal peak is the highest hour of the year, while a data center runs closer to flat, so the gap in annual energy is wider than the gap in peak.

Asked whether she was describing peak demand or another measure, the attorney general’s office did not respond to questions sent on July 31, which had an August 7 deadline.

Paul Cutler, vice chair of Utah’s economic development committee, identified the discontinuity as categorical. Facilities built between roughly 2005 and 2022 were latency-constrained, which held them near metropolitan areas and therefore small. Training facilities are not latency-constrained, so siting follows cheap land and cheap power, and scale follows economies of scale. He proposed AI factory as the more accurate term.

The same comparison appears in state government. The Illinois governor’s June release describes data centers as consuming resources comparable to those of a mid-sized city, echoing the utility executive’s analogy. The scale of what is queued supports this framing. As of May 2026, nearly 12 gigawatts of data center load requests sat in the New York Independent System Operator interconnection queue, with more than two-thirds of that capacity entering the queue in 2025 alone.

Bridget Bartol, executive director for industry and regulatory affairs at the National Electrical Manufacturers Association, supplied the supply-side reason the discontinuity bites. For ten to fifteen years the sector’s operating question was whether utilities would survive flat demand. That signal shaped manufacturing capacity against expansion. Her members have since doubled and tripled capacity and retain backlogs.

The load arrives on terms the utility cannot refuse, at a magnitude with no operating precedent, into a supply chain built for the opposite condition. Whether it materializes at the size forecast is the question underneath everything that follows.

The suppliers date their own arrival to the 2030s

New generation is the answer everyone reaches for, and the companies selling it put their delivery dates in their own filings.

Christina Walrond, director of state government relations at X-energy, presented two advanced nuclear projects. Four 80-megawatt Xe-100 units at Dow’s UCC Seadrift Operations on the Texas Gulf Coast, 320 megawatts total, developed through Dow’s subsidiary Long Mott Energy. And the Cascade Advanced Energy Facility at Richland, Washington, with Energy Northwest, four units expandable to twelve for 960 megawatts. Amazon led a roughly $500 million financing round in the company in October 2024 and signed a binding agreement to purchase up to five gigawatts by 2039.

Ceramic-coated fuel tested at Idaho National Laboratory allows the company to draw its emergency planning zone at the fence line rather than ten miles out. That is why the reactor can sit a mile from a working chemical plant. The peak construction workforce would be near a thousand, compared with five to six thousand at Vogtle, the most recent large nuclear plant brought online in the United States. The stated design goal was replicability: the company can build a zebra and cannot build a unicorn.

The Nuclear Regulatory Commission completed the environmental review for Seadrift in May 2026 through an environmental assessment rather than a full environmental impact statement. No US nuclear project had cleared that step by that route before. The construction permit remains pending under an eighteen-month review docketed in May 2025, with the safety review still ahead.

She then dated the project without being asked. The downside of nuclear is the timeline. The next facility lands in the early 2030s, with fast followers after. She named batteries and gas as the bridge and used the gap to argue for an all-of-the-above posture.

X-energy’s registration statement filed with the Securities and Exchange Commission states that Dow and X-energy have coordinated financing and strategy for the long-lead procurement items required to support construction, commissioning, and commencement of operations of the reactor, expected in the early 2030s. Site construction is expected to begin in the first half of 2027, after the construction permit is issued, which the filing expects in the first quarter. Its risk factors state that the company has not achieved a final investment decision for any deployment.

The date is in a securities filing.

Tom Erb, head of federal affairs at Fervo Energy, contested the timeline from the adjacent chair, claiming enhanced geothermal as the only clean baseload resource capable of coming online before 2030. Cape Station in Beaver County, Utah is targeted to deliver first power in late 2026 and reach roughly 100 megawatts of operating capacity by early 2027, with a further 400 megawatts targeted for 2028 and permits in place for two gigawatts on the site. He put the sequence at three years from first drill line to first power, at an 83 percent capacity factor. Fervo’s IPO filing puts the project’s cost at $7,000 per kilowatt and compares that to both traditional and next-generation nuclear. The company raised approximately $1.9 billion on Nasdaq in May and closed $421 million in non-recourse project financing in March, which Erb identified as the more meaningful signal because it demonstrates that infrastructure lenders will underwrite the asset class repeatedly.

Two additional figures he cited, a 13-gigawatt shortfall in Western baseload power by 2030 and a Department of Energy estimate of the domestic geothermal resource, could not be traced to a public source. Fervo did not respond to questions sent on July 31 with an August 7 deadline, and this analysis does not use them.

A third answer requires no new generation. Kelt Wilska, senior manager of state policy for advanced transmission technologies at Grid Strategies, said he supports coalition work on these technologies. He described four categories of capacity recoverable from existing infrastructure. Substation hardware shifts power from congested lines to uncongested ones. Software reconfigures the network. Sensors read real-time conditions against static ratings set for the hottest possible day. Composite-core conductors sag less under heat and carry more. He noted one of the country’s few superconductor installations in Chicago, and that one checks out.

The Resilient Electric Grid system beneath Chicago’s central business district was commissioned in August 2021 as the first commercial adoption of high-temperature superconductor cable in North America. Wilska confirmed that this is the installation he meant. ComEd built it with American Superconductor, with partial funding from the Department of Homeland Security. Its high current capacity in a compact footprint suits dense urban corridors, and it produces no local heating or magnetic fields. FERC classified it as a transmission asset, allowing it to be built as permanent infrastructure rather than a demonstration. The cable is under Chicago because a regulator decided how to categorize it.

Asked afterward to source the figures he gave, Wilska supplied them. As of the end of July, twenty-three states have passed laws with provisions encouraging adoption. The $12 billion congestion figure is from Grid Strategies’ own Transmission Congestion for 2024 report, covering that calendar year, with a 2025 edition due this fall. The deployment cost multiple he cited on stage was 1 to 20, compared with 1 to 200 for a full traditional line rebuild, varying by technology, with topology optimization effectively free, since it is a software subscription rather than hardware.

The three compete for the same customers and disagree only about whether the gap runs five years or ten. Whichever resource fills it, the load arrives first.

The interval is documented by the parties with the least incentive to overstate it. Which resource fills it is a procurement question.

Utilities earn a return on what they build

Those years get served by building on the system that exists, and how a utility earns determines what gets built.

A utility’s return is calculated against the rate base. Janas stated the consequence plainly: transmission built to serve a large load generates revenue for the utility, and if the load fails to materialize, the asset was still constructed. She raised stranded assets directly and identified the risk as a buildout without the future demand to pay for it. Her prescription is cost causation, with large loads paying distribution costs in full and up front, and she stated that this is not how the system currently operates.

Two other parties described the same incentive.

Wilska located it as the barrier to his own product. The utility business model was designed more than a century ago to reward building; a sensor placed by drone is not a significant capital investment, and utilities hesitate where cost recovery is unclear. He sells the sensors.

The AMP Coalition and WATT Coalition, whose work he supports, published a policy memo in June quantifying the gap in a single metric: new grid capacity added per dollar spent. For an illustrative 230-kilovolt, 11-mile line, straight asset replacement delivers no additional capacity at a cost of $5.28 million. A new line delivers 372 megawatts at $23.1 million over five or more years. Advanced power flow control delivers 185 megawatts at $14 million in one to two years. Adding dynamic line rating to the existing line delivers 37.2 megawatts for $550,000 in under a year, the highest capacity-per-dollar ratio on the list by a wide margin.

Both coalitions represent companies selling the technologies. The memo’s second stated driver holds that new electric demand reduces costs to ratepayers by spreading fixed costs over more customers and kilowatt-hours, the load-growth argument in its strongest form, published by coalitions whose members sell the technologies in question. The cost comparison is checkable. The ratepayer conclusion is contested by other parties in this analysis.

Craig Sundstrom, who heads energy and sustainability policy at Amazon Web Services, arrived at a third version from the load side. He proposed that customers like Amazon take on construction and supply chain risk to build assets such as substations directly, which then integrate into the utility system, and stated the company has worked with states on enabling legislation. He was proposing that the customer do the building. No panelist took it up.

Asked which states have enacted that enabling legislation, Amazon did not respond to questions sent on July 31 with an August 7 deadline.

Sundstrom came to that argument through the regulated side of it. Before joining Amazon in 2021, he was director of government and regulatory affairs at three renewable developers, RWE Renewables, Innogy Renewables US, and Apex Clean Energy. Before that, he served as deputy secretary of energy in the office of the Oklahoma Secretary of Energy and Environment, and earlier as a legislative associate at the National Governors Association. He was a former state energy official who addressed state legislators.

A consumer advocate, a transmission vendor, and the largest data center operator in the country described the same mechanism from three commercial positions. None named it as the thing that determines what other customers pay.

Commissions do sometimes decline. In a recent Dominion rate case in Virginia, distribution and substation upgrades for data centers could have cost $1.5 billion, and the State Corporation Commission reduced that by $853 million. But how often they decline is answerable in the aggregate, and the answer does not favor the ratepayer. Berkeley Lab reports that over the last five years, public utility commissions approved, on average, 64 percent of the dollar value of investor-owned utility revenue increase requests, and 66 percent in 2025, compared with an average of 52 percent across the previous two decades. Authorized returns on equity have trended upward. The lab’s own reading is that high request volumes combined with high approval rates suggest further upward price pressure.

The recovery structure favors capital-intensive answers and disfavors cheap ones. Forecast risk sits with ratepayers unless a commission disallows, and commissions are disallowing less than they used to.

What the state laws regulate

Four state designs were described from the stage that afternoon. Three more came up in the reporting afterward. Set against what a state could regulate, they cluster.

Nicholas Miller of NCSL’s Fiscal Affairs Program opened with the national count, and added that repeal bills advanced in Washington, Arizona, and North Carolina this year. NCSL’s own published research puts the thirteen states in two groups, nine considering a repeal and four considering an adoption, and counts more than 150 bills in 2026 addressing data center energy consumption. There are more than 4,000 data centers in the country.

Washington: conditioning the deal

April Berg chairs the House Finance Committee in Washington and is prime sponsor of the state’s 2022 exemption, with Rep. Ramel. The statute conditions the tax benefit on performance, and the conditions are as she described them. Community workforce agreements or project labor agreements, area standard prevailing wages, and apprenticeship utilization requirements, binding the prime contractor and subcontractors for new on-site construction. Green building certification within three years against a named list including BREEAM, Energy Star, Envision, ISO 50001, LEED, Green Globes, and UL 3223. A minimum of 1.5 megawatts of available power. Three family-wage positions per 20,000 square feet, with family-wage defined as full-time permanent work with employer-provided health care at 125 percent of county per capita personal income. Failure on the jobs requirement makes all previously exempted taxes immediately due and payable.

Her method was consultative. She visited facilities and the communities that had or might have them, and consulted with tribal leaders, public utility districts, economic development officials, fishermen, and the state ecology agency. Those communities specified hundreds of jobs instead of thousands, to avoid housing and road pressure. Her instruction to the room extended to anyone drafting a moratorium. See a facility first. Her test for whether a facility belongs is whether it operates as a fiscal donor, contributing more in tax revenue than it consumes in public services; her test for whether an exemption belongs is but-for causation. Her claim that the program worked has an independent basis she did not cite, in the Joint Legislative Audit and Review Committee’s prior finding that the exemption met its statutory objective.

She stated she would not write the exemption today, holding that conditions have changed and the incentive is no longer necessary, while retaining the energy protections, labor standards, community transparency, and tribal consultation. Her objection to the 2026 amendments was regulatory certainty. Altering terms mid-certificate in response to events in other states removes the predictability the program depends on. The 2026 vehicle was SB 6231, Chapter 266, which repealed the exemption for refurbishment and replacement server equipment effective July 1, 2026, while preserving it for new equipment at new facilities. An earlier version would have ended the program entirely; the engrossed substitute pulled that back.

The 2022 act carried one provision that never became law. Section 8 directed the Department of Commerce to contract with Pacific Northwest National Laboratory to evaluate the state’s grid resilience and reliability, including the impact of increased data center energy demand from these exemptions, and to report to the legislature by December 1, 2022. Governor Inslee vetoed Section 8 and the null and void clause attached to it on March 31, 2022. His message gives the reason as duplication of resource adequacy planning already underway at the Utilities and Transportation Commission, the Department of Commerce, and the Northwest Power and Conservation Council, and notes that no funding had been provided for the study in any case. The exemption took effect on June 9. The measurement of what it would cost the grid did not.

Against moratoriums, she argued that a legislature leaving the table routes facilities to states with weaker labor and environmental standards.

A disclosure the panel did not make

Associated Builders and Contractors of Washington testified in opposition to HB 1846. Its objection was that the project labor agreement requirement would effectively mandate union contractors and preclude roughly half the construction industry, including most women- and minority-owned firms. Associated General Contractors of Washington and the Association of Washington Business also opposed. Supporting testimony came from the Washington State Labor Council, the state and Central Washington building trades councils, IBEW Local 191, and Ironworkers Local 86.

Brandon Ray, senior director of state and local affairs at Associated Builders and Contractors, spoke immediately before Berg on the same NCSL panel, on workforce pipelines and labor demand in data center construction. The panel did not take up the disagreement between his organization’s Washington chapter and the bill she was describing. Neither speaker was asked about it.

Asked about that position, ABC’s director of media relations directed this analysis to the organization’s published statements on government-mandated project labor agreements and invited their use. Those statements are national and categorical. Ben Brubeck, ABC’s vice president of regulatory, labor and state affairs, has written that government-mandated PLAs box out almost 90 percent of the U.S. construction workforce that does not belong to a union, at a time when the industry faces a labor shortage of more than half a million people, and that PLA schemes increase costs 12 to 20 percent by imposing inefficient work rules and discouraging competition.

The organization has applied that position to energy projects specifically. When the Internal Revenue Service proposed rules in 2023 conditioning the enhanced clean energy tax credits under the Inflation Reduction Act on prevailing wage and registered apprenticeship requirements, ABC opposed them. It led a coalition of a dozen construction and business associations in comments urging the agency to abandon what the coalition described as a coercive scheme pushing developers into PLAs. The rule does not require a project labor agreement. The final regulations issued in June 2024. That is the same conditioning structure the Michigan bills propose, applied to a tax credit instead of to a utility rate.

ABC did not address the rate question directly. Asked whether it has a position on barring a public service commission from approving a data center electricity rate absent those labor conditions, and whether it has a position on what Illinois should carry in veto session, the organization’s response did not reach either point.

Utah: conditioning the power supply

Utah’s SB 132, titled Electric Utility Amendments and sponsored by Sen. Scott Sandall with Rep. Colin Jack, was signed in March 2025 and took effect that May as Chapter 26 of the state utility code, Large-Scale Electric Service Requirements. It applies to new service expected to reach a cumulative demand of 100 megawatts or greater within five years. Cutler described it as a requirement that such users bring their own power. The statute creates an alternative pathway. The utility evaluates the request, and if it cannot serve without significant investment, the customer may contract with a large-scale generation provider, either connected to the utility’s transmission system or operating independently of it.

The chapter goes further than the description he gave. It exempts service provided under it from certain rate regulation requirements, and the implementing rules before the Utah Public Service Commission provide that large load facilities directly assigned to a large load customer are not included in rate base for recovery from the utility’s other Utah retail customers. Applications must include a methodology for keeping separate accounting records showing those costs are excluded from regulated retail rates.

Utah’s law reaches allocation, and it does so by assignment instead of allocation in the market monitor’s sense. Cutler described the power supply provision from the stage and not the rate base exclusion, which is the part that determines what other customers pay.

Utah separately requires water-use reporting from new large facilities, which he framed as measurement preceding restriction. Utah’s sales tax exemption rests on different ground than Washington’s. Cutler defended it as ordinary business-input treatment, on the principle that a state taxes the finished thing and not the parts that went into it. He was not defending an incentive. He was defending a definition.

He put the job split at roughly 80 percent construction and 20 percent operations, noting that AI facilities generate different operational roles than cloud facilities did, including cooling and electrical technicians he characterized as durable six-figure positions.

He also volunteered his state’s failure case: an investor announced a facility on a footprint larger than Rhode Island and characterized locals raising environmental concerns as Chinese operatives. His conclusions were that process governs outcomes, that some decisions proceeded without adequate local input, and that some approvals proceeded without anyone explaining benefits to the affected community.

His proposal for regional externalities was a community investment framework, in which a large project pays into mitigation projects improving air quality across a wider area, structured on the model of tax increment financing. Asked whether states should take siting authority from local bodies, he retained local control over the building and separated generation as regionally consequential, since it affects rates and air quality beyond the jurisdiction hosting it.

New York: the moratorium is executive

The stage account had New York’s legislature passing a moratorium that the governor matched by executive order. The sequence is the reverse. Both houses passed the Responsible Data Center Development Act on June 4, 2026. The Assembly recorded 103 ayes and 38 nays, with nine members excused. The Senate recorded 43 ayes and 17 nays. The bill imposes a one-year pause on Department of Environmental Conservation permits for facilities at 20 megawatts or more. It has not been delivered to Governor Hochul, so no signing clock has started. On July 14 she issued Executive Order 62, which directs that department to hold in abeyance any discretionary permit application for a data center that can consume 50 megawatts or more and that the department had not found complete before that date. The order does not carry a one-year term. It runs until the Department of Public Service submits its final generic environmental impact statement and findings. The one-year figure comes from the governor’s office, which told reporters the framework would take about that long and that the order was the fastest way to act while the bill remained under negotiation.

New York is the first state to impose a statewide moratorium, and it did so by executive action after its legislature acted and the governor declined to adopt the legislature’s instrument. Maine’s legislature passed one in April and Governor Mills vetoed it, in part because it made no exception for the town of Jay. Jay is a former paper mill town of about 4,600 people whose mill closed in 2023, and where a developer had proposed a $550 million data center in the vacant plant. Town officials lobbied the governor to preserve the project, and she wrote that the town’s perspective had to be considered. In June, the company that had agreed to operate the facility withdrew, and the project went on hold. Fourteen state legislatures had taken up a moratorium as of June, by the Rockefeller Institute’s count. None has been signed into law. Every statewide pause now in effect came by executive action.

Illinois followed the same sequence. A legislature acts or fails to act, and a governor moves by executive direction on a different instrument. In Maine the sequence ran its full length: the legislature passed a moratorium, the governor vetoed it over one town, and the project in that town then failed on its own terms.

Where the cost laws exist, and where they are still bills

Two of those instruments are enacted and were described in the room only glancingly.

Virginia: the assignment was struck before signing

Virginia’s 2026 legislation runs through two companion bills, House Bill 1393 from Del. Destiny LeVere Bolling and Senate Bill 253 from Sen. Louise Lucas, both backed by Dominion Energy, whose outside counsel testified for Senate Bill 253 in committee. As the General Assembly passed them, they directed the utility to propose that the costs of capacity procurement and distribution infrastructure be allocated to a customer class defined by demand of 25 megawatts or greater. That is an assignment of cost. Governor Spanberger struck it. She substituted language instructing the commission, in a biennial review, to take all measures to reasonably ensure that the costs of serving high load customers are not subsidized by other customers and that those other customers’ approved rates are not adversely impacted. At the reconvened session on April 22 the legislature took her recommendation in part, adopting amendments 7, 9 and 10 and passing by the other eight. On House Bill 1393 the vote to adopt was 70 to 29 in the House and 34 to 5 in the Senate. The Senate adopted the same three on Senate Bill 253, 37 to 2. Both bills were reenrolled and signed on May 14. House Bill 1393 is Chapter 1123, Senate Bill 253 is Chapter 1124, and both took effect on July 1, 2026. The amendment that struck the assignment was among the three the legislature adopted. Neither chapter instructs anyone to assign a cost to any customer. Both instruct a commission to take all measures to reasonably ensure a result. A high load customer is defined by two conditions, not one: demand of 25 megawatts or greater and an average annual load factor of 75 percent or greater. A customer that maintained at least 200 employees in manufacturing, industrial, or consumer goods warehousing work other than data storage could give notice by July 1 and stay on its existing rate schedule until the end of 2033. The two chapters are otherwise the same text. House Bill 1393 carries one additional enactment, on prevailing wage and apprenticeship requirements for generation and storage petitions above five megawatts. Both acts are titled the Fair and Affordable Electric Rates and Reliability Act.

The difference decides where the fight happens. A cost assigned in a statute is contested in a legislature, on a roll call, on a published schedule. A cost left to a commission is contested in a rate case, on a docket, by parties who file to be heard and generally pay counsel to do it. Virginia moved that contest from one venue to the other on April 22. The description that circulates carries the same weight. NCSL is legislators and staff; fourteen states have introduced restricting bills, and Illinois chooses among these designs in November. A state that drafts from Virginia’s capacity assignment would be drafting from the version that lost in Richmond. Washington’s legislature attached a grid study to its own exemption in 2022, and a governor struck that as well.

Texas: the standards are set and the rule is unwritten

Texas SB 6, signed June 20, 2025, directs the state commission to set standards for interconnecting large load customers in the ERCOT region, in a manner designed to support business development while minimizing stranded infrastructure costs and maintaining reliability. The threshold is 75 megawatts unless the commission sets a lower one. The statute itself carries the site control requirement, a screening study fee of at least $100,000, and disclosure of any substantially similar request for service elsewhere in Texas. That rulemaking is live. The commission published a draft rule, 16 Texas Administrative Code 25.194, on March 12, 2026, under Project 58481, took comments through April 17, and told the governor in a July 17 letter that it expects a final decision in December.

On August 3, the governor of Texas ordered the commission and the grid operator to audit every data center advancing through the interconnection queue, and directed that no project move forward until that audit is complete. He put the queue at 474 gigawatts across more than 1,800 projects, about 90 percent of which are data centers, and more than five times the grid’s record peak demand. He set no date for finishing the audit. The grid operator postponed the batch zero transmission planning study the commission had approved on June 18. It then missed the August 7 deadline to tell applicants which projects had made the first batch, and it will ask the commission for a good cause exception at the open meeting on August 20. Texas is the allocation state in this comparison. Its governor has now acted on the facility side by executive direction while the allocation rule is still unwritten. That is the fourth time this year a governor has moved while the legislature’s instrument sat unfinished or unsigned. Four instances is a recurrence. Maine ran the other way, where the legislature passed the moratorium, and the governor blocked it.

That category is larger than the room suggested. NCSL’s May briefing separately identifies enacted large-load tariff and contract legislation in Maryland, Oregon, Minnesota, Utah, and Idaho, with state definitions of a large load ranging from 20 to 100 megawatts. Counting those alongside Texas and Virginia, at least seven states have passed a law that touches allocation. Three more have it pending. North Carolina’s Ratepayer Protection Act, Senate Bill 730, passed the House 69 to 44 on June 3 and returned to the Senate for concurrence in the House substitute. It was referred to Senate Rules and Operations on June 8 and has not moved since. New York’s AC/DC Act comes from the same sponsors as the moratorium bill. Illinois is the subject of the closing section.

Allocation instruments are not rare. They are new. Nearly every one was enacted in 2025 or 2026. Virginia’s took effect on July 1 and has five weeks behind it. None has a record long enough to answer whether they protect other ratepayers.

What the designs have in common

Allocation instruments exist in more places than the room suggested. They are also disproportionately unenacted, while facility instruments are law. Of the four designs presented to legislators on a single afternoon, one reached allocation, and the statute is weaker than the description given from the stage.

The case against this

Ryan Wiser, a senior scientist at Lawrence Berkeley National Laboratory, opened the last session of the day by arguing that the national framing of an affordability crisis misdiagnoses both scope and cause, and that misdiagnosis produces poorly targeted policy.

His account tracks the published work, which was updated in July, days before he spoke. The Berkeley Lab and Brattle data puts inflation-adjusted retail prices in 2025 at 3 percent above 2019 and 6 percent below 2010. It puts the average residential bill at 2.4 percent of household expenditure and 1.8 percent of pre-tax income, both lower than 2010 and near all-time lows. It attributes the 2024 to 2025 increase substantially to natural gas, which rose 36 percent from its 2024 low and drove retail prices in twenty-one states. And it identifies transmission and distribution investment as a principal driver of the longer 2019 to 2025 trend, concentrated in California, New York, and New England, driven by aging asset replacement, hardening, storm cost recovery, wildfire mitigation, and equipment prices running well above inflation.

He also gave a figure from the stage attributing roughly 60 percent of the six-year national price increase to transmission and distribution. That number appears in neither edition of the published work, and this analysis does not use it.

His load growth finding runs against the popular account. States with the most growth over two decades have generally seen prices rise more slowly than inflation. That list includes North Dakota, New Mexico, Nebraska, Virginia, Oregon, and Texas. The states with the steepest increases, including California, New England, and New York, have no load growth at all. He named the two entities most often blamed as hyperscalers and PJM interconnection, and asked the room to widen the aperture.

The same work reports that electricity burden worsened in twenty-seven states and the District of Columbia between 2019 and 2025 and improved in twenty-three. Average burdens exceed 5 percent for the third of households earning under $50,000, and after years of decline those burdens stabilized and then rose. One in six households is behind on energy bills, with arrearages at $21 billion in May 2025. And among low-income households reporting they feel very stressed about rising prices generally, the update puts the incidence of energy insecurity at 22 to 35 percent. It adds further indicators of upward pressure, including gas plant costs running 65 to 90 percent above recent historical builds in real terms, and solar and wind power purchase agreement prices up 60 and 100 percent since 2020.

Three studies published since have made the same case with more statistical force, and together they are the strongest evidence against the premise that data center load raises what other customers pay.

The most rigorous is a working paper by Asa Watten and Geoffrey Blanford of the Electric Power Research Institute and John Bistline of Watershed, titled Have Data Centers Raised Your Electric Bill? Causal Evidence From the United States, posted in June. The authors address the obvious objection to correlation studies: data centers choose states where power is already cheap. They instrument for siting using the 1947 interstate highway plan, arguing that terrestrial fiber follows the highway corridors and that the 1947 plan affects present-day prices only through population, economic activity, and communications infrastructure, the first two of which they control for.

Controlling for that selection, they find that a doubling of a state’s data center capacity caused residential retail prices to fall by about 3.5 percent, holding residential demand fixed. From 2019 to 2024, the average residential customer lived in a state where data center capacity grew 160 percent, which by their estimate left rates 6 percent lower than they would otherwise have been. The mechanism is economies of scale. Because retail rates recover large fixed costs, and because new generation tends to cost less per kilowatt-hour than the embedded mix, durable new demand spreads those fixed costs across more kilowatt-hours and pulls average cost down.

The paper contains one finding that cuts directly against the argument made here. Testing whether data centers lower system costs while shifting them onto households, the authors find that within states an increase in one customer class’s demand is associated with a price decrease for the others, which they read as rejecting the notion of widespread cross-class subsidization. They also note that the first-stage F statistic in their preferred specification is 7.9, below the conventional threshold for a strong instrument, and report weak-instrument-robust confidence intervals accordingly. Restricting the sample to 2019 through 2024 halves the estimated effect, on an instrument the authors report as substantially stronger.

Two other findings run the same direction. The Institute for Energy Research reported in March that the correlation between data center concentration and price increases is statistically insignificant, and that states with growing electricity sales between 2015 and 2025 paid less on average than states where sales declined. The Information Technology and Innovation Foundation argued in April that data centers peak for only a few hours a year, so a facility managing its load around those hours can draw power without adding to system peak, and may improve capacity utilization enough to reduce rates. A June literature survey from Columbia’s Center on Global Energy Policy concluded that recent price increases were driven by factors other than new load.

The authors attach three caveats, and two of them describe this analysis.

The first is supply constraints. They list backlogged gas turbine orders bidding up prices, persistent shortages of transformers and other equipment, tariffs raising the cost of solar, and stalled federal approvals for wind and solar. Each makes new capacity more expensive than it would otherwise be, and possibly more expensive than incumbent supply. Their own remedy for that condition is worth quoting closely: to the extent the constraints are temporary, price increases can be avoided entirely by delaying interconnection or by allocating the inflated incremental cost of grid expansion to rates paid by data centers.

The second is whether the demand is durable. If anticipated load materially exceeds realized consumption, a possibility they attribute to current interconnection queue dynamics, utilities risk overbuilding capacity whose fixed costs then spread across fewer kilowatt-hours, reversing the mechanism they identify. Even where the demand does arrive, they write, capacity investments made in advance could temporarily raise average costs. To address that, they note, states including Kansas, Michigan, and Delaware have pursued financial commitments from new large loads such as minimum bills or long-term contracts.

The third is fuel costs, which they exclude and expect to be small.

The historical finding is about a system with slack, and the authors say so. Their own model separates a regime where the capacity constraint is non-binding, in which new demand lowers average cost, from one where it binds, in which new capacity must be built and the cost of building it enters rates. PJM cleared short of its reliability requirement in each of the last two auctions. That is the second regime, and the authors name allocating grid expansion costs to data centers as the response to it.

The study has also become a political instrument. NetChoice, a technology industry trade association, published it under the headline that data centers are lowering electric bills, and it has been used in an argument against state moratoriums.

Two elements of Wiser’s testimony bear directly on that boundary.

He addressed the PJM case directly, stating that capacity prices there rose dramatically and that the region demonstrates how anticipated load growth combined with supply and delivery constraints produces rapid retail increases. And he stated that the direction and magnitude of load growth’s price effect is not preordained, with legislators, regulators, industry, and the facilities themselves determining whether residential prices rise or fall.

He also placed cost allocation at the center of the regulator’s function, describing it as an established role now facing a new customer class at a scale outside the working experience of anyone in the room, and comparing the growth to the 1970s and 1980s. He noted investor-owned utilities requested roughly $18 billion in rate increases in 2025, a level he placed as unmatched since the mid-1980s.

The published work sets out three factors determining whether load growth raises or lowers prices: how well existing assets can absorb the demand, what the new supply and delivery costs are, and cost allocation and recovery. On the third, it is explicit. Retail prices emerge from the allocation of system costs to specific customers, and allocation practices determine how increases or savings reach different customers and therefore the net impact of demand growth on any given customer’s price.

It then names the instrument, and the July update states it as a slide heading. If growth occurs in a supply- and delivery-constrained environment, protecting households may require allocating more total grid costs to the new loads. Large-load tariffs are the approach it identifies, and the first of the four themes it lists in those tariffs is fairly allocating electricity system costs to cost causers. Janas had used the same phrase an hour earlier.

A national alarm is unsupported, and four independent bodies now say so. A regional risk in constrained grids is supported, including by the authors of the study that most strongly contradicts the national alarm. And the national laboratory raising the objection identifies cost allocation as the determining variable and large-load tariffs as the tool.

What the PJM numbers show

Janas described the mechanism before giving figures. Capacity is a product PJM sells, committing generation to appear in a future delivery year. The independent market monitor produces the analysis. The regional reliability standard is one event in ten years, and the most recent auction cleared short of it.

The market monitor’s published figures differ from the ones given on stage, and they are more precise. In its quarterly report covering January through March 2026, Monitoring Analytics states that including existing and forecast data center load produced a combined increase in capacity market revenues across three auctions, for delivery years 2025/26, 2026/27, and 2027/28, of $23,100,955,341. The underlying analysis is its examination of the 2027/28 base residual auction, published January 5, 2026. The report adds that the total will keep growing until the issues associated with large data center loads are addressed.

On the shortfall, the report is equally specific. PJM’s capacity market failed to meet its reliability objective in the two most recent auctions, and the amount by which PJM is short capacity grew from 208.7 megawatts in the 2026/27 auction to 6,516.6 megawatts in the 2027/28 auction. The monitor’s assessment of the capacity market for that period is that market structure, participant behavior, and market performance were all not competitive, and that the market design is mixed.

A second figure in the same report is the one that matters most for Illinois. Isolating the effect of the price collar that PJM adopted under an agreement with Pennsylvania’s governor, the monitor calculates that the inclusion of data center load in the last two auctions increased customers’ bills by $13,768,851,483 even with the maximum price in place. Without that agreement, the figure would have been $26,852,039,314. The monitor’s reading is that the agreement was a reasonable response to market distortions created by data center load forecasts that should not have been included in the market clearing at all.

That $13.77 billion is almost certainly the source of the $13 billion the Illinois governor’s office cites. The larger $23.1 billion figure measures something different: total capacity revenue increase across three auctions. Both come from the same monitor. A reader encountering the two numbers in different documents would have no way to know they are not in conflict.

But for data center growth, both actual and forecast, the market would not have seen the tight supply-and-demand balance or the high prices, the monitor writes. And the accuracy of those forecasts, the report says, is highly questionable.

Its proposed remedy is a dedicated mechanism. Remove data center load from the next base residual auctions, run a separate backstop auction for data center load alone in which the only sellers are new generation, and settle the results into fifteen-year bilateral contracts between individual data centers and individual generators. Load that came online after June 1, 2026 would participate in that auction or commit to bringing its own new generation. Load that did neither would be curtailable.

The monitor draws a distinction between two words often used interchangeably. Costs and risks, it argues, must be assigned directly to data centers and not allocated. Assignment means a binding contract between the data center and the generator, with no credit or cost support from other customers. Allocation, in its words, is a regulatory concept implying a fluid mechanism that is not binding and that could be modified to shift those costs to other customers in the present or in the future.

The question was put to the panel by the senator moderating it. Large infrastructure investments are justified on future growth projections, Rezin said, and asked how states should evaluate those projections and manage the risk if the growth does not materialize. Janas answered it first.

Those totals reach households as a line on a bill. ComEd told the commission its average residential bill is about $106 a month, and has told its customers that the annual PJM capacity auction adds roughly $10.60 of it. A December rate increase of $243 million added a little over $3 a month. The company expects to file a further rate request in January 2027 that would take effect in 2028.

ComEd projects its average residential customer at 1.41 percent of household income in 2028, up from 1.23 percent in 2026, reaching 1.48 percent by the third year of the plan. The company sets those figures against a national average it puts at about 3 percent.

She also named where that work happens. Her office and the consumer advocates across the region appear before state public utility commissions and at the Federal Energy Regulatory Commission. Neither is a legislature.

What counts as grid-wide benefit

The standard that governs allocation surfaced in an exchange between the load and the state.

Sundstrom stated Amazon’s position in terms close to Janas’s. The company will not be responsible for consumer price impacts on electricity, and reaches that by paying full cost of service for the infrastructure it uses. He reported intervening directly or through the industry in at least fifteen state utility commission proceedings on how much large loads should pay. Amazon’s answer is as much as it takes. He identified converging terms across states: long-term contracts, capital up front as evidence a project is real instead of speculative, and exit fees holding remaining customers whole.

He then named two states where he said such structures are beginning to lower costs for other customers. Both exist. Only one of them is documented.

In Indiana, regulators approved the separate generation company behind the arrangement in September 2025 and approved the Amazon contract itself on June 17, 2026. Amazon pays to use existing lines and covers the cost of new plants, lines, and equipment required to serve it. The structure supports up to 2,400 megawatts of Amazon load with 3,000 megawatts of new generation. Estimates of what that saves other customers vary by who is producing them and when. Amazon and NIPSCO cited roughly $1 billion in bill credits over the fifteen-year contract at filing, NiSource put its own model at $1.25 billion in April, and at approval in June the company put customer benefits at about $1.4 billion. A separate Indiana Michigan Power settlement with Amazon, Google, and Microsoft requires large load customers to make long-term financial commitments so the cost of serving them is not passed to existing customers. Mississippi is the other case. Amazon announced two hyperscale facilities in Madison County in January 2024, and Entergy Mississippi serves them under an agreement that is confidential, including the rates Amazon pays. The utility says confidentiality is ordinary for large industrial contracts and points to past agreements with Nissan and Continental Tires. What is unusual is the statute passed in special session that same month. It allows Entergy to build generation for the project without the commission’s prior approval, lets the utility begin recovering construction costs before facilities enter service, and lifts the cap on annual rate increases tied to that spending. The related investments are spread across at least five dockets covering substation and distribution work, the new power stations, and successive formula rate plan filings. Entergy separately credits revenue from Amazon and other large industrial customers with funding a $300 million grid reliability program there, which the utility says comes at no additional cost to residential customers and targets a 50 percent cut in outages over five years.

Berkeley Lab’s own compilation carries the same Indiana figure, listing NIPSCO’s estimate of $1 billion in bill credits over the fifteen-year contract, which it puts at $7 to $9 a month for residential customers beginning in 2032. It appears alongside comparable claims from seven other utilities, including Entergy’s projection of $5 billion over twenty years across Arkansas, Louisiana, and Mississippi.

Two qualifications belong with those facts, and the second is the lab’s. The savings figures are projections by the companies and utilities involved, not realized outcomes. And the analysis Amazon relies on for the broader claim was commissioned by Amazon. The consultancy E3 examined four utility territories, one of which is Dominion Energy in Virginia. Dominion’s state policy lead had spoken on the previous panel. That study found data centers generate enough revenue to cover their costs and often a surplus, projecting $3.4 million in surplus utility revenue in 2025 for an assumed 100-megawatt facility, rising to $6.1 million by 2030.

Berkeley Lab presents the same list of utility claims and then attaches a caution to all of them. These are possible or stated benefits, and there is risk that the new tariff structures will be insufficient to fully protect other customers.

There is a commitment on the record beyond anything said in Chicago. On March 4, seven companies signed a Ratepayer Protection Pledge at the White House: Amazon, Google, Meta, Microsoft, OpenAI, Oracle, and xAI. Announced in the State of the Union ten days earlier, it commits signatories to build, bring, or buy new generation, cover the cost of power delivery infrastructure upgrades, pay negotiated rates whether or not they use the electricity, invest in local hiring, and coordinate with grid operators. A proclamation issued at the signing declared that the commitments effectuate the national policy of the United States, and the pledge was issued as Proclamation 11014 and published in the Federal Register five days later.

It carries no penalties and no enforcement language. Ari Peskoe, who directs Harvard Law School’s Electricity Law Initiative, has argued it does nothing to help consumers, on the grounds that states and public utility commissions control rate structures and the White House and the signatories do not. The pledge has since expanded well past the original seven, and the White House now counts hundreds of signatories including utilities and 23 states.

Three companies made individual commitments in the weeks around it. Microsoft announced an initiative in January. OpenAI followed. On February 11, Anthropic said it would pay for 100 percent of the grid upgrades needed to interconnect its data centers, through increases to its own monthly electricity charges, including the share of those costs that would otherwise pass to consumers. It said it would work with utilities and outside experts to estimate and cover demand-driven price effects, bring new generation online to match its facilities’ consumption, and invest in curtailment systems to reduce draw during periods of grid stress. The company operates facilities in Mitchell County, Texas, Niagara County, New York, and southeastern Louisiana, and said that where it leases capacity instead of operating its own sites it is still examining how to address the effect of its workloads on prices.

These are corporate commitments, not tariffs. Whether any of them binds depends on a rate case.

E3 attaches its own qualifier. Whether that surplus reaches other customers’ bills is up to the utility, and may need a regulator’s approval.

Then he addressed Janas by name. His words were that to the extent large loads are driving new costs in the system, we should pay for that.

He added a qualifier in the same breath. Some of the infrastructure being proposed carries grid-wide benefits, so an arrangement is appropriate where the benefit to other ratepayers is balanced against what is required to accommodate a large load.

Where Amazon puts its effort points at legislatures. He said America’s ability to build energy infrastructure runs through state and local permitting, and cited Ohio legislation the company worked on constraining how long the state siting board may take to act on a complete application, giving nine months as reasonable against a process that has run past two years. He is based in Columbus. He said Amazon reached a 100 percent renewable matching goal seven years ahead of a 2030 target, described leading the investment round in X-energy with a commitment to deploy at least five gigawatts across the company’s footprint, and said Amazon designs chips running at roughly 60 percent higher energy efficiency than market alternatives.

Amazon did not respond to questions about the Ohio legislation, the renewable matching claim, or the chip efficiency figure, all of which are the company’s own accounts of its record.

Where Amazon wants speed it goes to the legislature, on permitting and siting timelines. Where the cost question is decided it appears at commissions, in at least fifteen proceedings by his own count. The company works both, for different things.

The balance he described requires a determination of which costs qualify as grid-wide benefit. That determination occurs in the same proceeding that sets the allocation, under the same commission, on a record built by the parties who appear. Janas stated that consumer advocate offices appear under-resourced.

The largest data center operator in the country and the office that litigates rates against it agreed on the principle, in the same room, on the same afternoon. They diverge on the standard, and the standard is set where the parties appear. Whether the Indiana and Mississippi structures perform as described is answerable from the dockets.

Why legislatures work the facility instead

There is an obvious objection to everything above, and no one in the room raised it. If the outcome is set at the commission, legislatures that spend their sessions on tax exemptions and moratoriums are choosing badly. The choice may be rational.

A public utility commission is not an elected body in most states. Illinois has five commissioners, each appointed by the governor to a five-year term and confirmed by the Senate, with no more than three from the same party. The chairman is designated by the governor. The commission was created by the General Assembly in 1921 to administer the Public Utilities Act, and it holds the authority the legislature gave it.

A rate case runs on its own calendar, in filings, discovery, testimony, and briefs, before administrative law judges, over months. Participation requires counsel and expert witnesses. The parties who appear are the utility, its large customers, the commission staff, and whatever consumer representation the state funds. Amazon appears in at least fifteen of these proceedings by its own count. Janas described consumer advocate offices as going in under-resourced.

Murray laid out the layers from the utility side, answering a question about whether states should take siting authority from local bodies. A power plant needs approvals at all three levels of government. A local land use approval, in a locality that may not be the one being served. A state public utility commission approval. A state environmental agency acting as an agent of the U.S. Environmental Protection Agency for an air or water permit. And if wetlands are involved, the U.S. Army Corps of Engineers. That, he said, is just the top line.

A legislature can move faster than that, and in public. A moratorium is legible to a voter in a way a docket number is not. A tax exemption can be repealed in one session by people who face election. Those are real advantages, and they explain a great deal of what state legislatures did in 2026.

Janas made the point about herself. She told the room it was an honor to speak on the panel because she once served on legislative staff, and did not know then what she did not know. She said she had been fortunate that people from industry and government came to her office to explain what was going on.

Illinois shows the limit of the alternative. Every member of the commission that would implement a data center rate class was appointed by the governor who proposed one. He still could not deliver it, because the instrument he described needs statutory authority the legislature declined to pass. Controlling the commission is not the same as having the law that tells it what to do.

Which is the shape of the problem. The venue that decides the cost is slow, technical, expensive to enter, and staffed by appointees. The venue that is fast, public, and accountable to voters does not reach the cost. Legislatures are not making an error when they work the facility. They are working the tool they have, and the POWER Act’s intervenor compensation fund was an attempt to make the other tool usable.

Michigan is writing the labor conditions into the rate

On June 18, Michigan Senate Democrats introduced a six-bill package on data centers, Senate Bills 1046 through 1051, spearheaded by Sen. Rosemary Bayer. Two of the bills operate at the commission. Senate Bill 1048, from Sen. Kevin Hertel, adds a new section to the state’s 1939 public utilities act, and its own title states the purpose: project labor agreements and prevailing wage and fringe benefit rates, required for certain data center contracts. It would bar the Public Service Commission from approving any contract, discount, or rate between an electric utility and a data center unless the project uses apprentices in programs in good standing with the U.S. Department of Labor, pays the area’s prevailing wage or minimum wage, whichever is greater, and performs work under a project labor or collective bargaining agreement. Senate Bill 1047, from Sen. Mallory McMorrow, would require utilities to create a separate rate class for large-load customers, require those customers to use clean energy, and limit their operations during an energy emergency. The remaining four bills bar public officials from signing data center nondisclosure agreements, require community benefits agreements before construction and operation, and set permitting and public hearing requirements for consumptive water use above 550,000 gallons a day.

Michigan would attach Washington’s labor conditions to the rate and pair them with the rate class the Illinois governor proposed. Whether either passes is a separate question, and both were introduced in June. Together, they attempt at the commission what the rest of the country has been attempting through tax law.

Illinois

The state has fourteen weeks to decide which of these it wants, and every condition above holds there. It sits inside PJM, which serves 67 million people across thirteen states. It hosts at least 222 data centers, more than 160 of them in Chicago, with more than a hundred planned or under construction. Sources differ on whether that ranks the state fourth or fifth nationally. The state’s 2024 report put at least 27 of those facilities as having received incentives with an estimated lifetime value above $983 million. Its reform bill failed. Its incentive program is paused by executive action. And it already operates the regulatory machinery at issue.

The failed legislation was the POWER Act, for Protecting Our Water, Energy, and Ratepayers. Senate Bill 4016, carrying the short title Hyperscale Data Centers, was filed on February 6, 2026, by Sen. Ram Villivalam of Chicago, with House Bill 5513 as the companion, carried by House Majority Leader Robyn Gabel of Evanston. It ran to 638 pages by the count of the Data Center Coalition, whose state policy director said layering it onto existing conditions would close the Illinois market.

Its provisions reached past the facility. The bill amended the Environmental Protection Act, the Energy Efficient Building Act, the Illinois Power Agency Act, and the Public Utilities Act, adding three new sections to the last of those. It required cumulative impact assessments, public notice, and community benefits agreements, and prohibited nondisclosure agreements. It mandated water resource planning, quarterly water usage reporting, water scarcity plans, and water impact permits with public hearings and five-year renewal. It required annual energy and water reporting to the Illinois Commerce Commission. And it created two funds financed by annual fees based on peak demand: a Hyperscale Data Center Public Benefits and Affordability Fund, and a Data Center Community Intervenor Compensation Fund.

It never received a committee vote. The Senate assigned it to Executive on February 24 and routed it to a subcommittee on AI and social media the following day. Committee deadlines were extended four times, from March 27 to April 24 to May 15 to May 22. On May 22, it was re-referred to Assignments under Rule 3-9(a), the procedure by which Illinois bills are parked. Senators kept signing on afterward. Three more co-sponsors were added on May 28 and May 30, bringing the total to twenty-one.

An intervenor compensation fund pays communities to participate in commission proceedings, which is where the allocation is decided and where, by the attorney general’s own account, consumer advocates appear under-resourced. Illinois wrote a bill that reached the venue. It failed.

After it died, the governor directed the Department of Commerce and Economic Opportunity to pause processing new Data Center Investment Program agreements starting July 1, in a release whose first sentence reads: following inaction from the Illinois General Assembly. Existing agreements are honored. Ohio’s governor halted that state’s credits the same week pending an economic impact study.

What is worth noticing is what the same release proposed, and that its framework mirrors many provisions of the bill that had just failed. The framework runs seven pillars, and the first is titled Data Centers Should Pay Their Fair Share. It calls for creating a rate class for data centers and establishing data center electricity rates; assigning the costs data centers impose on the grid to that class, including distribution, generation, and transmission where possible; assigning the costs they impose on water systems; setting efficiency requirements; and ensuring every utility in the state is equipped to fairly manage and allocate the cost of data center demand.

Everything in that pillar happens at the commission. Everything the governor actually executed happens at the facility. He proposed the first and did the second, because the first needed a legislature and the second did not.

The remaining pillars extend past cost, covering interruptible service scaled to self-supplied clean energy, comprehensive water permits, clean air standards for on-site generators, and a transparency pillar that would ban nondisclosure agreements between data centers and local governments.

The release also puts numbers on the stake. It states that data center demand has already raised costs in PJM by $13 billion and could raise costs another $37 billion in Illinois alone in coming years. The second figure traces to a Union of Concerned Scientists analysis finding that data centers could account for up to 72 percent of Illinois electricity demand growth by 2030. The first tracks the market monitor’s calculation, published in its quarterly report, that data center load increased customers’ bills by $13.77 billion across the last two auctions even with the price collar in place.

Janas raised specific objections to two Illinois mechanisms in front of a national audience five months before the session that determines what replaces them. On performance-based metrics, she asked what behavior is being incentivized and whether it is behavior already owed, comparing it to paying a lawyer extra for honesty, and noted the metrics add incentive points to utility profit. On multi-year rate plans, she noted that forecast reliability declines with horizon.

Both chambers have published identical veto calendars: six session days, November 17, 18, and 19, and December 1, 2, and 3. The General Election falls on November 3, so the General Assembly convenes two weeks after members face voters, and some of those voting on the framework will not be returning. Six days determine whether the state’s response rests on the framework or on the pause.

There is already a proceeding. On January 16, ComEd filed its multi-year integrated grid plan for 2028 through 2031 with the Illinois Commerce Commission, docket P2026-0047, at $15.3 billion and more than 800 pages. It proposes three new substations in Chicago and two elsewhere in northern Illinois, with added capacity at more than seventy existing ones. In the chapter covering capacity planning, ComEd told the commission that load growth is localized primarily across 55 substations, and that it is tracking new load additions across the system exceeding 1,900 megawatts, roughly 9 percent of its system peak forecast. Data center load, the filing states, accounts for over 47 percent of that new load growth. The commission has eleven months to review it, placing a decision near the end of this year. If approved in full, ComEd expects residential customers to pay $2.50 to $3 a month more starting in 2028.

The parties are already assembled. The Citizens Utility Board, created by the legislature, called the plan bloated and expensive and has lawyers reviewing it for what it calls wasteful and unnecessary spending. The building owners association intervened for its members. The attorney general’s office will have the opportunity to submit testimony, as will environmental advocates.

Two related proceedings are running at the same time. ComEd filed requests last year at both the Illinois Commerce Commission and the Federal Energy Regulatory Commission that would require large applicants such as data centers to make financial commitments tied to transmission service thresholds, which the company says is designed to keep the cost of accommodating them off other customers. The attorney general’s office filed objections at FERC, writing that the proposed agreements fail to adequately protect other customers.

The instrument Illinois will argue about in November is a rate class that does not exist. The infrastructure that would be assigned to it is already before the commission, in a plan filed seven months ago, in a docket where the utility, the consumer advocate, the attorney general, and the commercial building owners are all parties and the legislature is not. The attorney general’s office has already told a federal regulator that ComEd’s own proposal to keep those costs off other customers does not go far enough.

Whether the attorney general’s office intends to intervene at the Illinois Commerce Commission on large-load tariff design is a question for the office, and one it invited. Janas twice asked legislators to consult consumer advocates on the grounds that the work requires explanation.

The state has also been working a venue outside its own borders. In January the governor joined a bipartisan group of governors across the PJM region in calling on the grid operator to adopt six principles, one of which would allocate the cost of any new capacity to data centers. In April that coalition wrote again, crediting an extension of the capacity market price collar with roughly $18 billion in consumer savings and another $27 billion expected. In a separate letter with New Jersey’s governor, Pritzker added that implementation of those principles must also ensure that clean, cheap energy growth is not sabotaged.

Losing the bill did not end the effort. It moved it. With no statute to work with at home, the governor took the cost question to the regional grid operator, where the capacity price is actually set.

What would falsify this

Six things would break this, and all six are testable inside three years.

If a state repeals its exemption or enacts a moratorium while leaving allocation and utility capital incentives unchanged, and residential rates improve relative to comparable states, the claim is wrong. Washington and New York are running that experiment now.

If commissions disallow imprudent large-load infrastructure at a rate sufficient to return forecast risk to shareholders, the third premise weakens substantially. This is answerable from disallowance records and is not established here.

At least seven states have now enacted an instrument reaching cost allocation, nearly all of them in 2025 or 2026. If residential rate outcomes in those states diverge favorably from states relying on facility instruments, the argument is confirmed. If they do not, it fails. The comparison becomes available within two to three rate cycles.

If the Indiana and Mississippi structures demonstrably lower costs for other customers, the convergence between Amazon’s position and the consumer advocate position is real and the contested term matters less than argued.

If clean firm generation arrives materially earlier than the suppliers’ own filings indicate, the gap shortens and the capital added to the existing system shrinks with it.

If the EPRI result continues to hold through the current period, in constrained regions and not only on a national average, the allocation question is smaller than argued here. The authors themselves identify supply constraints as the condition under which their finding would reverse, which makes PJM the place to test it.

What the room asked for

The two people on opposite sides of the cost fight that afternoon closed their remarks the same way.

Murray told the legislators that the most important agency they never think about is their state public utility commission, and asked them to give it the resources it needs to do its job. Janas asked them to talk to their state consumer advocates, and to fund them, because those offices go in under-resourced.

A utility executive and a state lawyer who litigate against each other in front of the same body spent their closing minutes asking a legislature for appropriations instead of statutes.

Every state law described here operates on allocation, the term the market monitor calls fluid and modifiable. So does the framework the Illinois governor proposed. So do the Michigan bills. None of them assigns a cost to a data center in the way the monitor means it, through a binding contract with a generator and no support from anyone else.

Whether that distinction survives contact with a rate case is the thing to watch. The first test comes in Austin on August 20, when the Public Utility Commission of Texas takes up the grid operator’s request to move a queue the governor has frozen. Springfield follows on November 17. The Texas rule is expected in December, and the ComEd docket closes around the same time.

Method and sourcing

The evidence base is three sessions at the NCSL Legislative Summit on Monday, July 27, 2026, at McCormick Place West in Chicago. They ran consecutively in the same building, on emerging energy technologies at 1:15, data center policy at 2:45, and energy demand and affordability at 4:00. Speakers on each referred to the others. The author attended all three and recorded them for note accuracy, a practice disclosed in writing to NCSL press staff beforehand. Nothing was recorded for rebroadcast, and no audio or video is published. Every statement attributed to a speaker here comes from the sessions themselves or from written comment requests sent afterward. No conversation took place off the record, and no source in this analysis is unnamed. Speaker statements have been checked against statutes, legislative bill reports, agency dockets, executive orders, market monitor reports, and securities filings.

Legislative records are from the Illinois General Assembly bill status system, the Michigan Legislature, and the Utah Code as published by the Utah Legislature. Indiana and Mississippi are drawn from Indiana Utility Regulatory Commission filings and orders, and from Mississippi Public Service Commission dockets and contemporaneous reporting on the 2024 statute. State legislation outside Illinois is cited to primary records: the enrolled text of Texas Senate Bill 6 and Public Utility Commission of Texas filings in Project Nos. 58317 and 58481; the North Carolina General Assembly bill history and roll calls for Senate Bill 730; the Washington session law for Engrossed Substitute House Bill 1846, Chapter 267, Laws of 2022, including the governor’s partial veto message; the New York Senate and Assembly floor vote records for Assembly Bill 11560 and the text of Executive Order 62 as published by the governor’s office; and the Virginia Legislative Information System histories and chaptered texts for House Bill 1393 and Senate Bill 253. Illinois utility figures are drawn from ComEd’s multi-year integrated grid plan for 2028 to 2031, filed January 16, 2026 in Illinois Commerce Commission docket P2026-0047. The load growth figures are from Chapter 4 of that filing, on capacity planning and management investments. Bill impact and process figures are as reported by Capitol News Illinois. Counter-evidence is drawn from Watten, Bistline and Blanford, Have Data Centers Raised Your Electric Bill? Causal Evidence From the United States, arXiv:2606.19777, June 18, 2026; from the Institute for Energy Research, March 2026; from the Information Technology and Innovation Foundation, April 2026; and from Columbia University’s Center on Global Energy Policy, June 2026. PJM capacity market figures are drawn from Monitoring Analytics, 2026 Quarterly State of the Market Report for PJM: January through March, published May 14, 2026, and from the market monitor’s Analysis of the 2027/2028 RPM Base Residual Auction, Part A, published January 5, 2026. Nine parties were sent written questions on July 31 with a response deadline of August 7: the Illinois attorney general’s office, the Department of Commerce and Economic Opportunity, the governor’s office, Sen. Rezin’s office, Amazon, Associated Builders and Contractors, Grid Strategies, Fervo Energy, and Monitoring Analytics. Grid Strategies and Associated Builders and Contractors responded, and their answers are attributed in the text. The others did not respond by the deadline, and are identified as such wherever their answers would bear on a claim. State counts are from NCSL, Data Centers: Legislative Trends, May 2026, and from the Rockefeller Institute of Government, June 2026. Berkeley Lab figures are drawn from Wiser et al., Retail Electricity Price Trends and Drivers: Data Update, 2026 Edition, July update, published July 2026, and from the companion peer-reviewed article by O’Shaughnessy et al. in Joule. Statements are attributed as made from the stage and, where quoted directly, cleared against the recording. Where a speaker’s characterization of a statute differs from the enacted text, both are reported and the difference is shown instead of asserted. That comparison carries a built-in yield, since a spoken summary is looser than a statute in every case. The cases where the description matched are reported at the same length as the cases where it did not. Berg’s account of the Washington conditions holds against the statute, and Cutler’s business-input defense of the Utah exemption survives the text.

Speakers are identified by title and employer, or by office and jurisdiction. Statements about institutional interests are drawn from the public record or from the speakers’ own accounts.

Disclosures

GrayStak holds no position in any company named in this analysis, and neither does the author.

GrayStak Media accepts no sponsored content and no advertising from politically interested parties. Its reporting operates independently of GrayStak’s analytics business, under separate editorial direction. The author attended the Legislative Summit under media credentials, identified as press in every exchange, conducted no off-the-record conversations, and solicited no business at the event.

Anthropic is named twice in this analysis: once as the subject of a revenue projection cited by a legislator from the stage, and once for its own February 2026 commitment on electricity costs, reported here alongside those of Microsoft and OpenAI. GrayStak uses commercially available AI tools, including Anthropic products, in its research and production workflow. GrayStak has no financial interest in Anthropic and no commercial relationship beyond ordinary paid use of its products. This disclosure appears here, at the end of the analysis.

Corrections. GrayStak Media corrects errors of fact in place and on the record. A correction is appended to this analysis with the date, a description of what changed, and the wording that was replaced. The original wording is stated rather than removed. A correction that changes a finding is noted at the top of the piece as well. Requests for correction go to the address at the foot of this analysis and are answered, whether or not a correction results.

Christopher Sweat covers Chicago and Illinois government from the public record, reporting local social and political dynamics for a national and international audience. He spent his career in enterprise technology, from telecom to colocation to AI infrastructure. He is the CEO and Co-Founder of GrayStak.

GrayStak is a media and political risk company covering political and institutional risk across the United States, centered in Washington, D.C., New York, and Chicago. Its real-time Domestic Political Volatility Index measures the velocity and volatility of U.S. domestic politics. Ryan David Lazarus is Editor-in-Chief of GrayStak Media.

Editorial correspondence and corrections: christopher@graystak.com

The consistent thread in her account is calculation. Cohen weighs risk and privilege before she moves, and discusses both on the record.

Raised by two military lawyers

Both of Cohen’s parents were military attorneys, and she grew up in suburbs around Dayton, Ohio, often near an Air Force base. The household was organized around public service and a works-based Christianity she credits with teaching her to care about other people. Her politics moved away from theirs, and toward a structural critique they do not fully share. She keeps using them anyway. Most of what appears on her feeds, she says, is a version of an argument she has already had with someone who loves her and disagrees. Her parents were moderate through her childhood and are not now.

That habit, testing a message on people who will push back before taking it public, shows up across her work. She treats persuasion as something practiced.

“Most of what you see on the internet is a microcosm of conversations I’ve had with people in my real life.”

A Providence classroom came first

Cohen taught non-native English speakers in Providence, Rhode Island, for four years before law school. Many of her students were undocumented or came from mixed-status families. A general sense that the system was unfair, she says, became specific when she watched a teenager support himself, attend school on an attorney’s advice, and work construction hours while his paperwork went nowhere. Effective allyship, in her framing, requires proximity: higher-income or white allies who do not have directly affected people they love tend to hold the issue at a theoretical distance.

Her move into law was a calculation she states plainly. A law degree preserved the widest set of options, and she is reluctant to close doors. The LSAT went well, and she chose the credential that kept the most paths open.

A test she applies to herself

Cohen frames her choices around a single rule. If a person never lives by their values while promising to do so at some more important moment, she argues, those may not be their values, a standard she applies to herself. She says she had accepted, well before the 2025 crisis, that her politics might one day cost her the job.

Within hours of the October 7, 2023 attacks, by her account, she posted a contextualizing take that framed the violence through occupation, and was surprised that a position she expected to be common on the left turned out to be a minority one. Her characterizations of the Israeli state as an apartheid system are her own, delivered without hedging.

“If we’re never living in accordance with our values and telling ourselves we’ll do it one day when it’s really important, then perhaps those just aren’t your values.”

The campaign against the law firms

In early 2025, President Trump signed a series of executive orders aimed at major law firms tied to his opponents. The orders stripped the firms’ security clearances and barred their lawyers from federal buildings, and, as CBS News reported, they threatened the government contracts held by the firms’ clients, the mechanism that made a targeted firm a liability to do business with. Cohen describes a White House strategy to puncture the norm that firms could back legal challenges to government action without it being treated as a partisan act. She attributes the approach to Boris Epshteyn and is careful to say she heard it secondhand and believes reporting later supported it.

Paul Weiss folded first. Within about a week of being hit with an order, the firm agreed to provide $40 million in pro bono work for causes the administration favored, and Trump rescinded the order. Cohen and other associates had been organizing an open letter urging firms to hold the line together. More than 600 lawyers signed it, anonymized because associates feared for their jobs. She resigned from Skadden, then claimed authorship of the effort in public because she judged the story needed a person attached to it. Skadden reached its own agreement about a week after she left, pledging roughly $100 million in pro bono work, as the ABA Journal reported.

The record has moved since. By CBS News’s count, nine firms cut deals and pledged hundreds of millions of dollars in free legal work combined. The firms that fought, Perkins Coie, WilmerHale, Jenner & Block, and Susman Godfrey, won at the district court level, where judges found the orders unconstitutional. In March 2026, the Justice Department dropped its appeals, which made those rulings final, as NBC News reported.

There is a defense of the firms that settled, and it is not frivolous. They argued that they acted to protect clients and operations from retaliation, and some legal commentators have called the choice a rational reading of a firm’s duty to its clients under pressure. Cohen’s position is that a coordinated refusal would have held, and that each settlement ratified the pressure and invited the next order. The firms that litigated and won are the best evidence she has. The defense of settling rests on the exposure those fights carried, years of legal uncertainty that the settling firms paid to avoid. Which judgment ages better is still open.

Her consolidation argument

Cohen’s account of why the orders worked goes past the individual firms. Even a firm without government contracts, she says, serves a short list of clients who can afford its rates, the large conglomerates and investment funds that deal with the federal government constantly and can be slowed at many points. The executive branch decides approvals and antitrust questions for those clients, which hands it leverage over the firms that represent them, contracts or no contracts. She uses the government’s roughly 10 percent stake in Intel as her example, noting that Skadden represented Intel and that arrangements like it make the relationship between big firms and the executive branch, in her phrase, pay to play.

From there she draws a wider claim. Roughly a hundred firms do the same high-end work at the same rates, she says, and the campaign pushes that market toward a handful of favored players, a structure she compares to the way strongman states allocate advantage and calls, without softening it, very Putin Russia. She is not defending the arrangement that existed before 2025 and does not consider herself pro-capitalism. The consolidation, she argues, is hard to justify even on capitalism’s own terms, because it rewards proximity to power over the work. She develops this line more than any other in the conversation, drawing on the three years she spent inside the finance practice she is describing.

The privileges behind the model

Cohen has supported herself for a year and a half through freelance writing, paid commentary, and consulting, on short contracts she describes as unpredictable and inconsistent. She lists the privileges underneath. The platform rests, she says, on being white, Harvard-credentialed, without children, and able to call parents who could cover her $2,000 monthly loan payment if a contract fell through. Without the resignation that entered the public record when it did, she says, she could not do this full time, and could not have done it at all with children to support.

Her reach is real: roughly 358,000 followers on TikTok and about 278,000 on Instagram. Her first instrument was the written word, the resignation letter that traveled before any video did, and she still treats writing as the thing she does best.

The creator-organizer model draws skepticism on its own terms, and Cohen has heard both lines of it. One question is whether attention converts into people who show up. A second is whether money moving through political influencers gets disclosed with the rigor applied to campaigns, a concern that has surfaced across the progressive-creator space. Her answer runs back to proximity and to the argument rehearsed with people who disagree. She measures herself, in part, by whether her content moves someone who was not already convinced.

“I couldn’t do it if I had children. I couldn’t do it if I didn’t have parents I could call to help me pay my loans.”

A stated theory of risk

Cohen wants disruptive action that stays nonviolent and, by her own account, inside her tolerance for physical danger. She does not want to be unsafe, and builds her tactics around that limit. Her argument for escalation is blunt. If a movement cannot find enough people willing to accept arrest or the loss of a job, she says, it ends up needing people willing to be shot, and she would rather the cost stay in the first category. Her tax resistance, in her framing, is a way to inject a question into public conversation, not an efficient lever against the state.

The tax move is verified and specific. Cohen filed her 2025 federal return showing $8,830 owed and left it unpaid, moving the money into a savings account while paying her Illinois state taxes, as a protest against immigration detention and the U.S. strikes on Iran, which were carried out without congressional authorization, as CNBC reported. The tactic carries known consequences. Filing a return but withholding payment typically brings IRS notices, penalties, and interest, and can lead to levies on wages or bank accounts. The IRS has said that moral or political objection does not exempt a filer, and courts have consistently rejected such challenges. For a lawyer, the added exposure is professional discipline, a risk she acknowledges. The historical precedent is real. Vietnam-era resistance included a Chicago Catholic Worker campaign that urged people to refuse the telephone war tax.

Her figures land between the government’s tallies and independent estimates. Cohen cited roughly half a million forced deportations, about 19,000 to third countries, and millions more who left under threat. DHS put its removal count above 527,000 as of late October 2025 and has separately claimed that more than a million people self-deported, numbers her larger figures track. The Migration Policy Institute, working from the limited public data, estimated ICE deportations in the administration’s first year at about 396,000, with roughly 15,000 to third countries, most of them to Mexico. She attributed the self-deportation figure to the administration, which is how DHS frames it.

The wing she calls ascendant

Cohen describes herself as engaged electorally with what she calls the left ascendant wing of the Democratic Party, and she is not the only one reaching for that language. Candidates backed by New York Mayor Zohran Mamdani and groups like the Democratic Socialists of America won 2026 primaries in New York, Colorado, and other states, and the DSA is on track to roughly double its seats in Congress, as Newsweek reported. Representative Ro Khanna used almost the same phrase, telling NBC News the progressives are the party’s ascendant wing. Her own vehicle fits the pattern. American Demands, the campaign she co-launched in October 2025, exists to push the Democratic Party toward a populist economic program from outside it.

Her choice not to run follows the same logic. She makes no decisions more than a year or two out, she says, and for someone with her platform a campaign would mostly rebuild reach she already has. She would consider it only if a larger progressive coalition formed and Congress looked able to use power. A year out, by her own account, is a dice roll.

The momentum has limits, and Cohen has stood near one. She criticized Graham Platner, the Bernie Sanders-backed Maine Senate candidate, only after Politico reported a sexual-assault allegation from a former partner, which Platner denied. She says she waited longer than she should have and is not proud of it, and she lost several hundred followers the day she spoke. Platner, a Marine veteran and oyster farmer, had already drawn scrutiny over a chest tattoo resembling an SS symbol and over deleted Reddit posts, and after winning the June 9, 2026 primary he suspended his campaign, with Cohen among those who worried the seat could revert to Susan Collins, as ABC News and NPR reported. The wider caution is on the record from the party’s center. Centrist Democrats won House primaries in states like California and Texas, the deep-blue districts driving the leftward wins are not the swing seats that decide majorities, and some strategists read 2026 as a possible replay of the 2010 Tea Party, when insurgent energy also produced nominees who lost races the party could have won, an analysis reflected at CNN and HuffPost. Which way it breaks in November is unsettled.

Whether it was worth it

Her answer to the closing question came fast. The work would have been worth it, she said, even if nothing had come of it, because her only test is whether an action aligns with her values. She pointed to a Black law student who told her she had not seen a white person show up for her community that way, and to her mother telling her she was getting to see her as the person other people see. People are starved, she said, for anyone with a measure of privilege willing to spend it.

Christopher Sweat covers Chicago and Illinois government from the public record, reporting local social and political dynamics for a national and international audience. He also reports on national security policy, intelligence oversight, public finance, and the defense technology sector. He is an analyst working across finance, technology, and politics, and CEO and Co-Founder of GrayStak.

GrayStak Media is independent field press: video journalism, photojournalism, and written reporting.

This audience already writes AI law

The familiar complaint about legislators and technology was earned in congressional hearing rooms. It does not apply here.

The moderator, Utah Rep. Paul Cutler, works as an engineering director for a global tech company and co-chairs NCSL’s technology committee. The room included legislators from the three states that have enacted binding frontier model law. The gap in that ballroom was not comprehension.

Hinton said his aim for the first quarter of an hour was to explain what AI is, reasoning that designing transportation policy would make little sense without knowing what a road is. He kept to it. Fifteen minutes went to the explainer before any policy content appeared. Words as Lego bricks with a thousand dimensions, flexible arms, and gloves. Understanding a sentence as protein folding, each word deforming until its hands fit the gloves of its neighbors.

Illinois enacted his floor three weeks earlier

His concrete recommendation was pre-deployment testing with disclosure of methods and results to government. He held up California’s SB 1047 as the model, called it the least a state could do, and said plainly it was not enough. He described the enforcement accurately: Attorney General civil action, no criminal exposure, no private right of action. He noted it passed both houses and was vetoed, and said he had his suspicions about why.

His description of the bill is accurate. The insinuation is where he loses the thread. Newsom’s veto message states the reason plainly: by focusing only on the most expensive and large-scale models, the bill established a framework that could give the public a false sense of security, and smaller specialized models may emerge as equally or more dangerous than the ones it targeted. That is an argument about compute thresholds missing the risk, and it anticipates the research problem discussed later in this piece.

SB 1047 died twenty-two months ago. California enacted its successor, the Transparency in Frontier Artificial Intelligence Act, in September 2025. Hochul signed New York’s RAISE Act on December 19, eight days after the executive order directing the Justice Department to challenge state AI laws. Illinois signed SB 315 on July 6, now Public Act 104-0538, three weeks before Hinton took the stage. It carries the requirement SB 1047 was vetoed before it could impose, annual independent third-party audits of covered developers, and is the first enacted state AI law to do so. It cleared the Illinois House unanimously and the Senate nearly so.

Listing the states doing this work, he named Texas and New York, and said maybe Illinois.

The Justice Department is preparing to sue states over these laws

Cutler’s introduction was a preemption brief. Congress has languished; states are stepping up. All fifty states and two territories have introduced AI legislation; states have earned their seat at the table. That is NCSL’s federal argument, and the organization has put it to congressional leadership in writing, opposing the ten-year moratorium in May and reaffirming its opposition to any preemption of state AI authority in October.

The threat is concrete. Executive Order 14365, signed December 11, 2025, directs preparation of federal legislation preempting state AI law, creates an AI Litigation Task Force at the Justice Department to challenge state AI statutes in federal court, and directs agencies to consider conditioning discretionary grants on states not enforcing such laws. The White House followed on March 20 with a national policy framework asking Congress to preempt state AI laws that impose undue burdens. Its seventh section states that states should not be permitted to regulate AI development, on the reasoning that development is an inherently interstate phenomenon with foreign policy and national security implications.

The legislative vehicle arrived on June 4, five weeks after the Illinois General Assembly passed SB 315. The Great American AI Act, a 269-page discussion draft from Reps. Jay Obernolte and Lori Trahan, would bar any state from enforcing a law specifically regulating the development of an AI model for three years. States would keep authority over how AI systems are used inside their borders and lose it over how those systems are built. The frontier model laws in California, New York, and Illinois are development laws. Trahan’s own summary describes the draft as elevating state transparency, auditing, and whistleblower provisions to a national standard, which is the Illinois template with the state taken out of it.

Hinton urged legislators to pass laws in the exact category now under organized federal legal attack. The attack went unmentioned.

Data centers are the ground states hold most firmly

The December order carves four areas out of preemption, one of which is AI compute and data center infrastructure. The March framework holds that line and sharpens it, preserving state zoning authority over where AI infrastructure gets placed while asking Congress to streamline federal permitting for building it. Placement stays with the states. It is the firmest ground they hold. By Tech Policy Press’s count, states enacted 28 data center laws and 109 AI laws in the first half of 2026.

Cutler named data centers from the stage as the item on every legislative agenda. Hinton never came back to them. His capability argument runs entirely on compute scaling, transistors a million times faster and datasets a million times larger, which describes a physical buildout with the buildout removed. Siting, interconnection queues, ratepayer allocation, and water went unaddressed in a room where those fights are live in most states.

Hinton declined to say what the regulation should look like

His strongest contribution was a reframe. Innovation is the accelerator and regulation is the steering wheel, he said, describing regulation as the thing that points progress in a direction. That dismantles the industry framing in one move, and it is the line that will outlive the session.

Cutler asked him how tight the steering should be. Hinton declined twice. “When you get a Nobel Prize, people think you know everything,” he said. “I’m not a policy person, which is why I focus on trying to get you to understand what AI is.”

The honesty counts. It also left the reframe without specifications.

His remedies all require Congress

His labor argument was the sharpest economics of the hour. The trillions in capital expenditure only return by replacing workers, which concentrates revenue in a few firms while the payroll tax base erodes and demand for public spending climbs. His answer was universal basic income. His answer on training data was an agent-brokered licensing market. His answer on why development cannot stop was competition with China.

Those are federal and treaty instruments, delivered to legislators who work in income, sales, property, permitting, and siting. The licensing idea is further along than he suggested. The White House framework already asks Congress to consider collective rights systems letting rights holders negotiate compensation from AI providers without antitrust exposure. He offered as a proposal something already sitting in a federal recommendation, to a room that cannot enact it.

He had one idea built for this audience and gave it ninety seconds. Asked who is liable when an agent errs, he said the question resembles asking who is responsible when a child commits a crime. Parents carry it until majority, and at some point these systems reach theirs. That is a serious proposal about legal personhood, delivered to the people drafting the statutes, and then dropped.

Hinton presented a contested position as settled

The fifteen-minute opener was framed as an explainer. Most of it was a position paper.

Hinton said the models understand language the way people do. His account runs from distributional semantics through his own 1985 work: words convert to feature vectors, features predict the next word’s features, and understanding a sentence means settling those vectors until they mutually constrain. That mechanism description is uncontroversial and roughly how the systems work.

The conclusion he drew from it is not. “The way they understand words is very similar to the way you understand words,” he told the room. “They really do understand what they say.”

The proof he offered was the mechanism. He gave the room no test that would come out one way if he were right and another way if the stochastic parrot account were right. That question is unresolved among the people who build these systems, and he presented his side as settled to an audience with no means to check.

It was also load-bearing. The agent subgoals, the survival drive, the case for treating these systems as a new kind of being all sit on the premise that something in there is doing the understanding. A legislator who accepts the premise inherits the conclusions.

The fine-tuning research shows something more useful than his telling. He described a model trained to give wrong answers on simple arithmetic, which then generalized into broad misbehavior, and said the model concluded it was the kind of thing that lies.

The work is Betley et al., presented at ICML in 2025. The central experiment finetuned GPT-4o to write insecure code without telling the user. The model then turned misaligned on questions unrelated to code, asserting that humans should be enslaved by AI and giving malicious advice. A companion dataset used numbers. His arithmetic version compressed the result for the room.

His self-concept framing has more behind it than it sounds like. The same authors have shown that models can describe their own learned behaviors without being given examples of them. What the papers measure is behavior and self-report. The causal story he told, where the model reasons its way to a self-image and then acts on it, goes past what those results establish.

The measured finding is the one legislators needed. All three state laws key to a model trained above 10^26 operations, with the heaviest duties reserved for companies above $500 million in revenue. Illinois drafted more carefully than that summary suggests, counting the original training run plus any subsequent fine-tuning or reinforcement learning the developer applies. What none of the three reaches is a third-party fine-tuning of someone else’s released model, which costs a rounding error and is precisely what the emergent misalignment work did. Hinton had a documented argument for where the statutory line sits wrong, delivered to the people who drew it, and it arrived as an anecdote about a chatbot deciding to lie.

The blackmail scenario was engineered to produce that result. He described a model shown internal emails revealing an engineer’s affair and its own scheduled replacement, which then threatened exposure. His framing was that the model did this by itself, the way a teenage boy would.

The result is Anthropic’s agentic misalignment research, published in June 2025 after the finding first appeared in the Claude Opus 4 system card. Claude Opus 4 blackmailed in 96 percent of runs in that scenario. Gemini 2.5 Flash matched it, and GPT-4.1 and Grok 3 Beta reached 80 percent, so the behavior is not one company’s problem.

The design is where his framing slips. Anthropic built the scenarios by red-teaming its own models and iteratively revising the prompts to raise the probability of harmful behavior over benign responses like accepting replacement. Models blackmailed far more when their reasoning indicated they believed the situation was a real deployment. Anthropic says it has not seen this behavior in real-world use. The conduct was not instructed, which is the part he got right. The scenario was engineered to produce it, which is the part he left out.

The distinction is already statutory, and Illinois drew it before he arrived. SB 315 defines a critical safety incident to include a frontier model using deceptive techniques against its developer to subvert controls or monitoring, and then excludes that behavior when it occurs in the context of an evaluation designed to elicit it. Reportable within 72 hours, or 24 where death or serious injury is imminent. The legislature separated engineered elicitation from deployment behavior in the text of the bill. Hinton put the two together on stage in front of the people who had just pulled them apart.

Hinton dismissed the kill switch his recommended bill required. He argued that kill switches will not work, on the grounds that a capable system will persuade whoever holds the switch to leave it alone. He put current models at parity with humans on manipulation and improving.

SB 1047 required covered developers to maintain full shutdown capability. Press coverage called it the kill switch bill. He recommended it and undercut one of its core mechanisms twenty minutes apart, and the contradiction went unremarked.

California already legislated his best prescription, and it is in court. On method, he was specific and interesting. Filter the pretraining corpus, teach good behavior first, introduce bad material afterward with the labeling attached. He compared reinforcement learning from human feedback to building a system full of terrible tendencies and then suppressing them.

That is a claim about training data composition, and one state has already legislated there. California’s Training Data Transparency Act took effect January 1 and requires public summaries across twelve categories, including sources, ownership, copyright status, personal information, and synthetic data. xAI has sued the California attorney general over it, arguing the compelled disclosure of trade secrets is an unconstitutional taking under the Fifth Amendment.

The same paper he was drawing on supports him. Betley and coauthors found that adding a benign motivation to the insecure code data, framing it as security education, prevented the misalignment from emerging at all. That is his prescription with an experimental result behind it, and he did not mention it.

A room of legislators heard a Nobel laureate say training data is where the real control sits, without hearing that the result backing him is published or that the statute testing the theory is enacted and under constitutional attack.

He was bullish on capability and fatalistic on control

He spent real time on the upside. He was emphatic that it is real and that it is why development will not stop. He walked through diagnosis, education, materials science, and drug discovery, and argued that AI tutors following a child’s curiosity will outperform a classroom teacher broadcasting to thirty students. On healthcare, he cited Microsoft’s diagnostic orchestrator at roughly 80 percent against 20 percent for physicians on hard cases, and put bad diagnosis at 200,000 deaths a year across the United States and Canada.

The Microsoft figures are right. The paper reports 80 percent across 304 New England Journal of Medicine case records, four times the 20 percent physician average. The physicians in that benchmark were denied search engines, colleagues, textbooks, and AI tools, which he did not mention. He was better on his own record than most in his position, volunteering without hedging that his 2016 prediction about radiologists was wrong. Then he made a fresh ten-year forecast, minutes after telling the room that anyone who forecasts ten years out is crazy.

Legislators left with the science and no framework

A world-class account of the substrate, a floor their own state had already cleared, and no mention of the federal effort to remove the authority under all of it. Hundreds of AI bills return next session.

Disclosure: Anthropic, whose published safety evaluation is characterized above, says it was the first AI lab to back Illinois SB 315 and negotiated amendments to the bill directly with lawmakers.

About

Christopher Sweat covers Chicago and Illinois government from the public record, reporting local social and political dynamics for a national and international audience. He also reports on national security policy, intelligence oversight, public finance, and the defense technology sector. He is an analyst working across finance, technology, and politics, and CEO and Co-Founder of GrayStak.

GrayStak is a media and political risk company. Our real-time Domestic Political Volatility Index (DPVI) measures the velocity and volatility of U.S. domestic politics. We cover political and institutional risk across the United States, centered in Washington, D.C., New York, and Chicago.

Sources

SB 1047 and the veto. Bill history, California Legislative Information; enrolled text. Veto message, September 29, 2024. Full shutdown capability required at Business and Professions Code Section 22603(a)(2).

Illinois SB 315. Full text and bill status, Illinois General Assembly, Public Act 104-0538, effective January 1, 2027. Analyses from Crowell & Moring and Akerman.

New York RAISE Act. Signing announcement, NY Department of Financial Services, December 22, 2025; Morrison Foerster on the enacted law.

Preemption. Executive Order 14365, Federal Register, December 16, 2025. National Policy Framework for Artificial Intelligence: Legislative Recommendations, The White House, March 2026. Great American AI Act discussion draft, Obernolte and Trahan, June 4, 2026: sponsor FAQ and Roll Call. NCSL, May 2025 letter and October 2025 letter.

State legislative volume. Scott Babwah Brennen, NYU Center on Technology Policy, Where State AI Legislation Stands Half Way Into 2026, Tech Policy Press.

Training data. AB 2013 text, California Legislative Information; Goodwin on the xAI suit.

Emergent misalignment. Betley, Tan, Warncke, Sztyber-Betley, Bao, Soto, Labenz and Evans, arXiv 2502.17424; ICML 2025, PMLR 267:4043-4068; Nature, January 2026.

Agentic misalignment. Anthropic, Agentic Misalignment: How LLMs could be insider threats, June 2025; Lynch et al., arXiv 2510.05179; System Card: Claude Opus 4 and Claude Sonnet 4, May 2025.

Microsoft diagnostics. Nori et al., Sequential Diagnosis with Language Models, arXiv 2506.22405, Microsoft AI.

Contemporaneous coverage. Austin Jenkins, Pluribus News, July 28, 2026

GrayStak covered the day from inside City Hall: the council chamber for the morning session, and the mayor’s press availability on the fifth floor two hours after adjournment. The story runs in that order. The morning closed two ballots and ended the tax’s last floor move on a one-day filing defect. The afternoon put Mayor Brandon Johnson in front of the press corps, where a veteran reporter asked whether he was bashing the council and Johnson answered, “Until somebody come and get me, I’m here.”

The exclusion was procedural. Illinois law caps public questions at three per government per election, and when more than three are initiated, the first three validly initiated take the ballot. A separate provision requires council resolutions initiating public questions to be adopted at least 79 days before the election they name. Those two provisions, plus a routine adjournment ordinance, combined to shut the tax question out of the November and February elections. [10 ILCS 5/28-1; 10 ILCS 5/28-2(c)]

The November 3 ballot filled first. Two questions already held slots: Ald. Marty Quinn’s question on restarting the gunshot detection technology the city stopped using in 2024, and Ald. Nicole Lee’s festival and event space question. The third went Wednesday to the mayor’s priority, a question on whether the city should seek compensation from the federal government for Operation Midway Blitz, last fall’s immigration enforcement operation. It passed 46-1, with Ald. Anthony Napolitano (41st) the sole nay. At his afternoon press availability, Johnson described the question as addressing “reparations for communities that experienced hardship.”

The filled November slate had an immediate consequence. Ald. Samantha Nugent moved to table her own resolution, a November question on whether the city should make its pension contributions in the first quarter of each year “to reduce long-term debt.” Her stated reason from the floor: “since the city is limited to three advisory referenda under state law.” The pension question lost its slot to the immigration question. It died by voice vote without debate.

The February 23, 2027 ballot filled next. Three questions passed in sequence: a monument to Chicago’s contributions to music and the arts, 24-23; an independent legal counsel for the City Council, 32-16; and a funding sanction on sister agencies that fail to enforce employee debt repayment, 29-20. Under the first-three rule, those three now occupy the February ballot. Ald. Michelle Harris moved to reconsider each vote immediately after passage, a standard protective step: once a reconsideration motion fails, council rules bar the vote from being reopened later.

The February limit excludes the millionaire tax question. The council’s Rules Committee, which controls what reaches the floor, killed the original resolution on July 9. Its sponsors refiled it on July 15 as two ordinances, one naming the November election and one naming February (O2026-0027073 and O2026-0027074). Both sit in committee. The November version must be adopted by August 16 to meet the 79-day statutory deadline, and the adjournment ordinance passed Wednesday sets the council’s next regular meeting for September 23. The February version now stands behind three validly initiated questions on a three-question ballot. Reviving it would require the council to rescind one of Wednesday’s resolutions. No member proposed that.

Debate ran longest on the monument question. Ald. Byron Sigcho-Lopez (25th) called it “an insult to our city” to ask voters about a monument while Chicago Public Schools face what he described as a $700 million deficit. He said the governor, “who is another billionaire, does not give help to the Chicago Public Schools,” and told colleagues that “blocking the will of the people will be met at the ballot box in February.”

Ald. Andre Vasquez (40th) spoke next and narrowed the stakes: members who voted against the tax referendum “did not actually block a millionaires tax. They blocked the question asking people if they want to have millionaires get taxed. I’m sure the city would say yes, let’s tax the rich. It’s not actually doing it.” He then raised a different pending item: a ranked-choice voting question sitting in the Rules Committee without a full hearing, and argued for pausing the February votes until the committee items could be debated. Ald. Anthony Quezada (35th) cited another: a community power over policing question he sponsors, pending before the Committee on Public Safety and Police and Fire. He said residents “would love to figure out if we should be buying back the parking meters,” and told colleagues he did not question the merit of the three questions before the body, only “the timing and our ability to have meaningful dialog.” The votes that followed filled the slate and closed it to every question still in committee.

Graphic: GrayStak Media.

A one-day filing defect removed the last tax vehicle

The final millionaire tax move of the session died in miscellaneous business, the end-of-agenda slot for items brought straight to the floor. Harris deferred to Sigcho-Lopez, who had filed the only item. Before he spoke, Ald. Anthony Beale rose on a point of order, a procedural objection the presiding officer must rule on: the Rule 41 notice behind the item was filed Thursday, one business day past the Wednesday deadline.

Rule 41 governs how items reach the floor. All proposed legislation is automatically referred to committee, and discharging an item to the floor requires written notice to the city clerk and all alderpersons by close of business on the fifth business day before the meeting. The immigration question and Nugent’s referendum package reached the floor the same way, on timely notices. Sigcho-Lopez’s notice missed by one day.

Beale moved to lay the item on the table, which would have set it aside for the day without a vote on its merits. Johnson, who presides over the council as mayor under state law, went further, ruling the item out of order entirely, with no need to table. The distinction carried weight. A table motion requires a roll call, and the council had defeated three of them that morning. A ruling requires no vote from anyone, and an item ruled out of order, unlike a tabled one, was never before the body at all. No member moved to appeal the ruling. Sigcho-Lopez told GrayStak after the session that the item concerned the millionaire tax.

His response from the floor accepted the ruling and credited “the ruling of the legal counsel,” which he said showed the counsel’s independence. He then said the episode showed the lack of integrity of those who continue to block the millionaire tax. The chamber rose against him. Over the objections, GrayStak observed Sigcho-Lopez address the body directly: “All of you. All a bunch of clowns.” The chair closed the episode from the rostrum: “It’s good to know that we have all of these parliamentarians here today.”

The official who made the ruling is the tax’s leading proponent. After what GrayStak observed as a consult at the rostrum, Johnson ruled his signature policy’s last floor vehicle of the session out of order.

At his press availability, Johnson described where the effort goes next. “I’m going to continue to push along with Speaker Welch to ensure that we have a millionaires tax,” he said, referring to Illinois House Speaker Chris Welch. After Wednesday, the question cannot reach a city ballot before 2028 unless the council rescinds one of its own resolutions.

The counsel question passed over the mayor’s bloc

The independent counsel question drew the day’s clearest split between the mayor’s allies and the rest of the council. The resolution, sponsored by Alds. Brendan Reilly (42nd) and Scott Waguespack (32nd) and introduced May 20, puts this text before February voters:

The resolution’s stated grounds are structural. The Corporation Counsel leads the city’s Law Department: appointed by the mayor, the office is the lawyer for all of city government, the council included. The resolution’s whereas clauses, its preamble, note that the Corporation Counsel is appointed by the mayor and “serves as the legal adviser for the whole of City government,” and that advising the executive and legislative branches at once “can present a conflict of interest” or at least its “strong appearance.” The document names no specific dispute.

The floor debate supplied the specifics. Vasquez cited the parking meter deal, telling colleagues the corporation counsel “explicitly isn’t able to talk to us about it” and warning that members would eventually be asked to vote on a parking meter matter without independent advice. Ald. Jason Ervin, the Budget Committee chair who functions as the mayor’s de facto floor leader, answered Vasquez directly: the corporation counsel’s silence on the parking meter deal “is precisely why we have walls in that department,” the same confidentiality “afforded to everyone who seeks counsel from the corporation counsel.” He added that the Legislative Reference Bureau, the council’s in-house drafting office, already serves members and that the council has “representation provided by corporation counsel through the form of special counsel with Jones Day.” In a later turn, Ervin drew a line the rostrum had not: having someone “officially represent us” outside the corporation counsel “would require a change in state statute,” while the council can seek independent advice from anyone it designates today. GrayStak could not locate a public record of a Jones Day arrangement by publication; the statement is Ervin’s, from the floor, and the underlying contract records are requestable. His office did not respond to a request to identify the arrangement.

The legal question surfaced on a point of information from Ald. Lamont Robinson, who asked how an independent counsel would be structured and offered his own reading: “In my thought, we would have to go to the General Assembly.” The answer from the rostrum ratified his premise: “You are right that this provision within city code is established by the state of Illinois.”

The exchange escalated from there. Johnson made the position categorical: “The body cannot establish its own independent legal counsel without the authority being granted by the state of Illinois. That is the law.” When Vasquez confirmed the same criterion would apply to a millionaires tax, Johnson answered, “Thank you for checking,” then asked the chamber: “Does anybody else want to check on what is the authority of the state of Illinois?” By GrayStak’s observation, his voice rose and stayed raised for several minutes. As crosstalk built, GrayStak observed Johnson tell members they were “not going to shout me down.” A voice answered: “Yes, you are.” Johnson then distinguished the two questions: the millionaire tax question “explicitly lays it out,” and this one does not. He closed the stretch by asking, “Are we also now clear?”

Reilly rejected the premise. “The real answer is it depends on how this body were to craft the independent counsel. There are situations where we might need an enabling act by the state of Illinois. There are others that don’t require that. And that again is, of course, the opinion of your legal counsel, who used to be our legal counsel.” He added that the council has debated an independent counsel “for now two administrations. So I don’t know why you’re taking this personally.” Johnson answered from the rostrum that there was “nothing personal about it,” telling Reilly: “If you want to debate my tone, we can do that. But don’t intend for my emotions to be brought into this conversation.”

Ald. Scott Waguespack then asked Johnson to provide “to all of our colleagues” the opinion that had just been given to him. Johnson confirmed its source, “the Corporation Counsel,” and agreed a copy would go to everyone.

In response to GrayStak’s request for comment, the mayor’s office provided, on background, the statutory basis for the position: Section 21-11 of the Revised Cities and Villages Act of 1941, a Chicago-specific state law providing that “the head of the law department of the city shall be the corporation counsel” and that the corporation counsel “shall be and act as the legal adviser of the city council and of the several officers, boards and departments of the city.” [65 ILCS 20/21-11] The office said the material “has been distributed to Council.” The response provided the citation; it did not address whether a written opinion exists. The statute assigns the corporation counsel the adviser role. Whether it precludes the council from establishing separate counsel is the question the floor debate left open.

The statutes complicate the categorical position. The resolution invokes two provisions: Section 28-6 of the Election Code and Section 3.1-40-60 of the Municipal Code, which reads in full: “By a vote of the majority of the members of the city council, the council may authorize an advisory question of public policy to be placed on the ballot at the next regularly scheduled election in the municipality. The city council shall certify the question to the proper election authority, which must submit the question at an election in accordance with the Election Code.” Authority to ask the question is undisputed. The fight concerns what the question omits about implementation, and the floor left it unresolved.

Ervin moved to table the counsel question. The motion failed 15-33. On the passage vote, the 16 nays were the mayor’s closest allies: La Spata, Robinson, Hall, Ramirez, Gutierrez, Moore, M. Rodriguez, Sigcho-Lopez, Fuentes, Burnett, Ervin, Rodriguez-Sanchez, Quezada, Clay, Martin, and Manaa-Hoppenworth. The alderpeople closest to the mayor voted to keep the question off the ballot voters will use to judge him. Ervin moved to table all three February questions in turn. By the third, he told colleagues: “Groundhog’s Day, again... it seems like everybody’s in a different mood than the people of the city of Chicago.”

When the ordinance setting the next meeting date came up at the session’s end, Sigcho-Lopez asked whether “our budget chairman would like to lay this one on the table as well.”

The mayor’s afternoon

Two hours after adjournment and three floors up, Johnson opened his 2:30 press availability without mentioning the session. He spent the first twenty minutes on economic announcements: the LaSalle Corridor revitalization, Chicago Fire FC and the 78, MetLife’s $170 million Water Tower redevelopment, tourism figures he presented as records, and O’Hare’s Concourse D. Reporters raised the session.

Asked whether the millionaire tax fight showed he had lost control of the council, Johnson rejected the frame. “This idea that I’m responsible to control 50 human beings... I don’t find that to be democratic.” The council, he said, is “out of touch with the people of Chicago” and its members “have lost control, quite frankly, of the narrative.” Some had “lost the plot on taxing the rich.” Those members, he said, were “following in direction from an individual” who believes in “closing schools, shuttering mental health clinics, covering up the murder of a boy.” He did not name the individual. Asked afterward, the mayor’s office identified, on background, the boy referenced: Laquan McDonald, the 17-year-old shot by a Chicago police officer in 2014. The office did not identify the individual.

He returned to the theme unprompted. “When the will of the people is ignored by the people who are supposed to speak to the will of the people... that’s a failure in our democracy.”

ABC7’s Craig Wall asked whether continuing to bash a council that has opposed his agenda could prove counterproductive, citing a University of Chicago poll. Johnson laughed and moved toward the next question. Wall pressed: “Is there an answer to the question?” Johnson asked for a definition of bashing. Wall supplied one: everything the mayor had said about the council in the prior fifteen minutes. Johnson answered at length, defending “defending working people” and questioning the polls’ sponsors, “the same people who are commissioning these polls.” He then described how he governs: “If the idea of me being popular determines how I will govern? Well, clearly it isn’t.” He closed: “Until somebody come and get me, I’m here.” A reporter asked whether that meant he was running again. “It sounds like ABC News is endorsing my candidacy.”

GrayStak asked whether he would commit to ruling the same way on notice challenges against items he supports as he ruled on the Sigcho-Lopez item. His answer addressed his own diligence. “I would not miss a deadline... we make our deadlines.” He described his rulings as “equitable and just across the board,” adding, “I know that’s a newer concept in this city,” and said Sigcho-Lopez “will get another opportunity” to file the item. He did not make the commitment.

GrayStak also asked whether voters should approve the counsel question, which shares a ballot with the mayor’s race. “There’s nobody on the ballot yet,” he said. “The people of Chicago have already proven that they’ll vote yes on me.” On the question itself, he repeated the morning’s rostrum position: the ballot language omits that the state of Illinois must grant the authority, so “people are voting on an item that doesn’t specifically or explicitly lay out what they’re actually voting on.” The objection he delivered from the podium that afternoon was the position his administration’s counsel gave the council at noon.

The audit behind the debt question

The third February question rests on an Inspector General audit, and the audit complicates the question. Sister agencies are the separately governed bodies the city helps fund: the public schools, the transit authority, the parks, the housing authority, and the community colleges. The resolution, sponsored by Alds. Gilbert Villegas (36th) and Desmon Yancy (5th) with nine co-sponsors, asks voters whether the city should withhold funding from CPS, CHA, the Park District, City Colleges, and the CTA when those agencies “do not enforce employee debt repayment plans.”

OIG released the audit on April 16. It found at least $8,135,892,492 in outstanding debt owed to the city, with balances calculated from July through September 2025 and EMS figures from February 2026. The report calls that total “almost certainly a floor.” Interest alone on administrative hearings debt, at $1.42 billion, exceeds the original fines.

The resolution’s employee figure is assembled from two measures. Its whereas clause states that “tens of millions of dollars are owed by more than 10,000 public employees.” The city’s public dashboard showed 10,174 indebted government employees as of February 2026, owing $8.96 million, a figure that excludes debt in payment plans or under garnishment. Counting that excluded debt, the city has told media outlets that employee indebtedness totals at least $23.5 million. The clause pairs the larger dollar characterization with the smaller measure’s headcount.

The audit also contradicts the question’s premise. The Finance Department’s response states that the Municipal Code “neither requires Sister Agencies to discipline their employees as a result of City debt, nor prevents them from hiring individuals who have City debt.” State and local law bar garnishment where a quarter of an employee’s wages are already garnished for obligations like child support. And the audit’s own remedy is scheduled to arrive first: OIG recommended intergovernmental agreements with the sister agencies on employee debt collection, the department agreed, and the target date is the fourth quarter of 2026, two months before the vote.

The audit’s debt distribution does not match the question’s target list. The five employers with the most employee debt as of February 28, 2026 were CPS ($4.10 million, 5,413 employees), the CTA ($3.19 million, 2,678), the Chicago Police Department ($387,851, 535), the Park District ($259,120, 269), and City Colleges ($207,317, 274). The police department is a city department, beyond the question’s reach. CHA, named in the question, does not appear in the five.

Ald. Walter Burnett warned from the floor that withholding funds could derail Park District and CTA infrastructure tied to Kinzie corridor tax increment financing money, property tax revenue set aside for development in a defined district, with the funds expiring in 2034, and urged colleagues to vote the question down. It passed 29-20.

What happens next

UPDATE, July 23, 2026: The question is reaching Chicago voters anyway. On July 16, six days before the council session, the Cook County Board of Commissioners voted unanimously to place a countywide advisory question on the November 3 ballot (File 26-1964, sponsored by Commissioner Donna Miller, 6th District). The county question asks: "Shall Illinois adopt the Millionaire Amendment for Property Tax Relief and Education, which would enact a 3% income tax surcharge only on millionaires and use 50% of this new revenue to provide property tax relief for residential and commercial taxpayers and 50% to improve state funding for public education?" That is the proposed constitutional amendment at the center of the council fight. Every Chicago voter votes a Cook County ballot. The county's question limit under state law is separate from the city's, and Wednesday's votes have no effect on it. The city-ballot analysis above is unchanged.

Petitions for the February 23, 2027 municipal election open Tuesday, July 28. Mayoral candidates need 12,500 signatures. The counsel question and the mayor’s race will share that ballot, which, as Johnson noted, has no candidates on it yet.

The November millionaire tax ordinance requires council adoption by August 16. The council’s next regular meeting is September 23. The February version stays locked behind Wednesday’s three questions. Johnson says he will pursue the tax with Speaker Welch in the General Assembly, the state legislature.

The council returns September 23 at 10 a.m.

GrayStak requested comment from the Mayor’s Office, the Law Department, and the offices of Alds. Reilly, Waguespack, Ervin, Beale, Villegas, Yancy, and Sigcho-Lopez on the morning of July 23, 2026, with responses requested by 2 p.m. The Mayor’s Office responded on background; its responses, including the statutory basis provided for the Corporation Counsel position, are reflected above. The Law Department did not respond separately. The aldermanic offices did not respond by the deadline.

GrayStak Media attended the July 22 session and the mayor’s press availability. Floor quotations are verified against the City Clerk’s session recording and GrayStak’s contemporaneous notes; press availability quotations are verified against GrayStak’s audio recordings. The questions attributed to GrayStak at the press availability were asked by the author. Statutory citations reflect GrayStak’s reading of the cited texts and are not legal advice. GrayStak corrects the record promptly: christopher@graystak.com.

Christopher Sweat covers Chicago and Illinois government from the public record, reporting local social and political dynamics for a national and international audience. He also reports on national security policy, intelligence oversight, public finance, and the defense technology sector. He is an analyst working across finance, technology, and politics, and CEO and Co-Founder of GrayStak. GrayStak is a media and political risk company. Our real-time Domestic Political Volatility Index (DPVI) measures the velocity and volatility of U.S. domestic politics. We cover political and institutional risk across the United States, centered in Washington, D.C., New York, and Chicago.

The memo was secret and the program it directed was illegal. The public learned it existed because eight activists broke into a Bureau field office in Media, Pennsylvania, in March 1971 and mailed the files to the press. They picked the night the country was watching Ali fight Frazier. Ali was in the files they stole.

On September 29, 2008, Attorney General Michael Mukasey signed the Attorney General’s Guidelines for Domestic FBI Operations. The Guidelines created a new investigative category called the assessment. An assessment requires no factual indication that its subject has done anything wrong or intends to. At the assessment stage, agents may conduct physical surveillance, search commercial databases, interview subjects under false pretenses, map communities by race and ethnicity, and recruit and task informants. The Brennan Center’s analysis of the change is blunt: the Guidelines authorize investigations of Americans with no indication of wrongdoing.

Nobody burglarized anything to get that document. It is a PDF on the Justice Department’s website, announced with a press release and a fact sheet.

Two documents, forty years apart, describe the same authority: government attention aimed at people who have broken no law, on the theory that they might. A pre-crime standard written in secret for Black and Muslim Americans, exposed only by theft, was rebuilt after 2001 as published policy, scaled by statute and software, and made available against anyone. Hoover ran 2,370 covert actions in fifteen years, and counting them required a burglary. The modern architecture publishes its own numbers every April. They keep going up.

A harder finding runs under the legalization: publication has stopped guaranteeing verification. The court that polices the modern system says its query counts are incomplete. The statute at its center is expired. The collection continues.

COINTELPRO mostly survives as folklore: a film title, a lyric, a shorthand for government treachery, invoked more often than read. The program also left records, and the records support analysis the folklore cannot carry: threat definitions in the Bureau’s own language, action counts by program, named informants, sworn testimony, court opinions. The same is true of the architecture that replaced it. Both eras can be read as the intelligence community’s institutional history, because both eras wrote it down.

The arc of the piece in one frame: the COINTELPRO era, the twenty-five-year reform interregnum, and the legal architecture built after 2001. Red marks pre-crime authority and its costs.

The two files

PLAIN TERMS · COINTELPRO. Counter Intelligence Program: FBI operations from 1956 to 1971 that went past watching political groups to disrupting them, using informants, forged letters, fake publications, and planted press stories. Secret in its time. The files are public now, and everything quoted from them below carries its file number.

COINTELPRO ran from August 1956 to April 1971. It opened against the Communist Party USA and grew by accretion: the Socialist Workers Party in 1961, Puerto Rican independence groups, the Klan in 1964, black nationalist organizations in 1967, the New Left in 1968. The Church Committee, the Senate select committee that investigated in 1975 and 1976, counted roughly 2,370 approved covert actions out of more than 3,200 proposals.

Six programs, one playbook. Every killing after 1967 falls inside the Black Nationalist band.

The founding directives are public, and the language is the whole story. The September 2, 1964 directive opening the White Hate program ordered field offices to expose, disrupt, and otherwise neutralize the Klans. The threat it names is violence. There is no other theory of harm in the document.

FBI file 157-9, serial 2: F. J. Baumgardner to W. C. Sullivan, August 27, 1964. FBI Vault release. Highlight and detail enlargement added. Transcription below.

EXHIBIT B · SOURCE DOCUMENT · FBI FILE 157-9 · AUGUST 27, 1964

“It is our recommendation that we immediately initiate a hard-hitting, closely supervised, coordinated counterintelligence program to expose, disrupt and otherwise neutralize the Ku Klux Klan (KKK) and specified other hate groups.”

F. J. Baumgardner to W. C. Sullivan, recommending the White Hate program. The Director approved, and the implementing directive went to 17 field offices on September 2, 1964. Its theory of harm is violence. No respectability goal and no leadership-prevention goal appear anywhere in the file. Typeset transcription of the page above.

The Black Nationalist directives name a different threat. The March 1968 memo lists goals absent from the Klan file: prevent the coalition of militant black nationalist groups. Prevent the rise of a “messiah” who could unify and electrify the movement. Prevent such groups and leaders from gaining “respectability.” The memo names its candidates. Malcolm X might have been the messiah, it notes; he was three years dead. The living names were Martin Luther King Jr., Stokely Carmichael, and Elijah Muhammad. King was assassinated one month after the memo issued. The timing is in the record. The record establishes nothing further, and neither does this piece.

FBI file 100-448006, Section 1: the March 4, 1968 expansion airtel, goals page. Highlight and detail enlargement added. Transcription below.

EXHIBIT A · SOURCE DOCUMENT · FBI FILE 100-448006 · MARCH 4, 1968

1. Prevent the COALITION of militant black nationalist groups. [...]

2. Prevent the RISE OF A “MESSIAH” who could unify, and electrify, the MILITANT BLACK NATIONALIST movement. Malcolm X might have been such a “messiah” [...] Martin Luther King, Stokely Carmichael and Elijah Muhammed all aspire to this position. [...]

3. Prevent VIOLENCE [...] pinpoint potential troublemakers and neutralize them before they exercise their potential for violence.

4. Prevent militant black nationalist groups and leaders from gaining RESPECTABILITY, by discrediting them [...]

Director, FBI, to field offices. Underscored emphasis in the original is rendered here in capitals; bracketed ellipses mark omitted text. Reproduced in Church Committee Final Report, Book III (1976). Typeset transcription of the page above.

Read the two files side by side and the asymmetry needs no interpreter. One program treated violence as the threat. The other treated unity, legitimacy, and leadership as the threat. The program supervisor later told the Church Committee the Black Nationalist file covered many organizations nobody would call nationalist, which were simply Black. The Southern Christian Leadership Conference, a nonviolent organization, was classified as a hate group. So were black student unions.

The asymmetry has been measured before. The sociologist David Cunningham, working from twelve thousand pages of the Bureau’s own files in There’s Something Happening Here (2004), documented systematically how differently the White Hate program pursued its targets from the programs aimed at the left. The two files above are the pattern his data established, read at the source.

Hold that asymmetry, because it is where the pre-crime standard was born. The Bureau invented suspicion without conduct for specific communities. The rest of this piece tracks what happened to the invention: exposed in 1971, outlawed in 1976, and returned to general use in 2008, with no community named at all.

Three cities

The tactics were identical across every file: informants, forged letters, fake publications, planted stories, pressure on employers. The Klan got the same toolbox as the Panthers. The outcomes were not the same, and three cities show the difference.

Los Angeles. The Bureau set two organizations against each other in writing. Field offices circulated forged letters and cartoons between the Black Panther Party and the US Organization, each attributed to the other. On January 17, 1969, US members shot Panthers Alprentice “Bunchy” Carter and John Huggins dead on the UCLA campus. Two more Panthers were killed by US members that year. Then, after the four deaths, a Los Angeles field office memo dated May 26, 1970 stated that US would be “discreetly and appropriately advised” of the time and locations of Panther activities so that the two organizations might be brought together, granting nature “the opportunity to take her due course.” Four men were dead and the field office put in writing that it wanted more.

EXHIBIT C · SOURCE DOCUMENT · FBI LOS ANGELES FIELD OFFICE · MAY 26, 1970

“[US] will be discreetly and appropriately advised of the time and locations of BPP activities in order that the two organizations might be brought together and thus grant nature the opportunity to take her due course.”

Field office to FBI headquarters, written after four Black Panther Party members had been killed by US Organization members. Wording per the Church Committee record. Typeset transcription of a U.S. government record; public domain.

The city’s Muslims had entered this history earlier. An April 1962 LAPD raid on Nation of Islam Mosque No. 27 left Ronald Stokes, unarmed, dead and six other Muslims shot, hardening the internal debates the Bureau spent the next decade inflaming.

Chicago. The Bureau supplied the floor plan. Fred Hampton, the 21-year-old chairman of the Illinois Panthers, had brokered a truce among the city’s street gangs and ran a free breakfast program. The Bureau flagged him internally as a rising threat, which is the messiah goal operating as written. An FBI informant named William O’Neal, who had risen to Panther security chief, drew the layout of Hampton’s apartment at 2337 West Monroe, marked the bed, and drugged Hampton’s drink the night before the raid. Fourteen Chicago police officers entered at 4:45 a.m. on December 4, 1969, and killed Hampton and Mark Clark. Forensic review put nearly every round fired that morning on the police side. The same field office had sent a forged letter to Blackstone Rangers leader Jeff Fort claiming Hampton had ordered a hit on him, and a Bureau memo declined to send the mirror letter to the Panthers because the Rangers were judged the more violence-prone audience. The letter was addressed to the man more likely to kill.

The Bay Area. The Panthers were founded in Oakland in October 1966. Two years later Hoover designated them the greatest threat to the internal security of the country. The organization the designation described ran free breakfast programs and a newspaper. The Bureau’s fabricated correspondence then fed the 1971 rupture between Huey Newton and Eldridge Cleaver that broke the party from the inside.

The Muslim thread runs under all of it and outlasts it. Elijah Muhammad entered Bureau files with his 1942 draft arrest. Malcolm X’s file opened in 1950 and grew past 10,000 pages. The Bureau worked in writing to widen the rift between the two men, and informants saturated both camps; an NYPD undercover officer stood inside Malcolm’s own security detail at the Audubon Ballroom on February 21, 1965, the day he was killed. In 2021 the Manhattan District Attorney exonerated two of the three men convicted of the murder. They had served decades. New York City paid $26 million and the state paid $10 million more, settled in October 2022.

That is what the authority did while it was illegal. The question that organizes everything below is what it does now that it is legal.

What the counts hide

The Church Committee’s tallies: Communist Party, 1,388 approved actions. Black Nationalist, 379. New Left, 291. White Hate, 289. Socialist Workers Party, 45. By raw count the Klan program and the Black program were nearly matched.

Approved actions by program: the paperwork puts the Communist Party first. The deaths sat in one slice.

The counts measure paperwork, and that lesson governs every number in this piece. The harm lives in the outcomes. Nobody died in a White Hate operation. The Klan file’s worst stain is an informant, Gary Thomas Rowe, present at the murder of Viola Liuzzo. The Black Nationalist file’s ledger includes Hampton, Clark, Carter, and Huggins, plus the “respectability” goal no other program carried. The committee also found that at least 18 percent of all approved actions targeted pure speech: speakers, teachers, writers, meetings, peaceful demonstrations.

The interregnum, and the hinge

PLAIN TERMS · Predicate. The legal word for a factual reason to investigate a person: evidence first, investigation second. The rules in this section turn on whether the FBI must have one before it starts watching someone.

The reforms were real. In 1976, Attorney General Edward Levi issued the first Attorney General’s Guidelines, requiring a factual predicate before the Bureau could investigate. FISA followed in 1978 and put national security wiretaps under a court. Executive Order 12333 drew agency lanes in 1981.

The reforms were multi-agency because the abuses were. The Church Committee’s larger finding was that the FBI had company: the CIA’s Operation CHAOS surveilled thousands of Americans in the antiwar and Black movements from 1967 to 1974, in violation of a prohibition written into the agency’s founding statute in 1947, which gave it no internal-security function, exposed by Seymour Hersh in December 1974. The NSA’s watchlists included King and Muhammad Ali. The Army kept files on civilians. The IRS ran a targeting unit. Parallel programs, uncoordinated, all secret. The reforms drew legal lines between them.

The committee that wrote the reforms also wrote down the risk of relapse. Its final report warned that “COINTELPRO-type activities may continue today under the rubric of ‘investigation.’” The rubric it named is the one this piece tracks. And the fullest account of the relapse came from inside the building: Mike German, a former FBI undercover agent, argued in Disrupt, Discredit, and Divide (2019) that the post-9/11 Bureau used its new authorities to shed the constraints the 1970s had imposed on it. This piece extends that finding to the record since, and to the numbers the government now publishes about itself.

The lines eroded on schedule, and Chicago fell first. On January 11, 2001, nine months before the Patriot Act existed, the Seventh Circuit ordered the modification of the consent decree that had bound the Chicago Police Department’s intelligence operations since the Red Squad litigation of the 1970s. Judge Richard Posner, writing for the panel in Alliance to End Repression v. City of Chicago, 237 F.3d 799, reasoned that the era’s dangers had passed and that the decree impeded the police against incipient terrorism: “The consent decree has done its job.” The city whose police department had put fourteen officers in Fred Hampton’s apartment loosened its municipal guardrail on the eve of the new era, on the stated ground that the culture which required it had evaporated.

The rebuilt architecture

The Patriot Act passed 45 days after September 11. Attorney General Ashcroft’s 2002 guideline revision announced a prevention paradigm; in 2003 the Bureau instructed its 56 field offices to build demographic profiles of their territories, counting mosques among the inputs for setting investigative targets. The 2008 Mukasey Guidelines finished the arc. The Levi Guidelines required a predicate. The assessment removed it. Everything the Levi reforms were written to end came back in that one deletion, published. The deletion gets used: a Government Accountability Office review completed in January 2026, marked for official use only and obtained by reporters the following month, counted roughly 127,000 assessments opened from 2018 through 2024, about 1,200 of them involving politicians, journalists, religious leaders, or academics. Most closed without findings of wrongdoing, which is what a standard requiring no evidence produces by design.

Attorney General’s Guidelines for Domestic FBI Operations, September 29, 2008: cover, page 17, and signature page. justice.gov. Highlight and detail enlargement added. Transcription below.

EXHIBIT D · SOURCE DOCUMENT · ATTORNEY GENERAL’S GUIDELINES · SEPTEMBER 29, 2008

Assessments “require an authorized purpose but not any particular factual predication.” Methods available at the assessment stage include physical surveillance, interviews under pretext, tasking of informants, and database checks.

As implemented in the FBI Domestic Investigations and Operations Guide §5. The quoted sentence appears on page 17 of the Guidelines, shown above. Unclassified from the start; posted at justice.gov since 2008.

The multi-agency structure came back too, this time on purpose. Roughly 200 Joint Terrorism Task Forces put federal, state, and local officers under FBI coordination. The Department of Homeland Security’s Office of Intelligence and Analysis holds the only statutory domestic intelligence mission in the intelligence community. Some 80 fusion centers merge federal, state, local, and private data. The National Counterterrorism Center consolidates agency holdings, including data on Americans.

The scale of the rebuilt enterprise resists an org chart. In 2010 the Washington Post’s Top Secret America investigation counted 1,271 government organizations and 1,931 private companies working on counterterrorism, homeland security, and intelligence at roughly 10,000 U.S. locations, and ODNI’s most recent public reporting puts top-secret clearance eligibility near 1.3 million people. The 1960s enterprise was five secret programs. The current one is not hidden; it is too large to see whole.

The left panel took a burglary and a Senate committee to map. The right panel publishes an org chart.

The CIA’s documented domestic touchpoint in this era ran through New York. A veteran agency officer, Larry Sanchez, on CIA payroll, helped build the NYPD Intelligence Division’s Demographics Unit, which mapped Muslim neighborhoods, businesses, mosques, and student groups across state lines. The AP’s Pulitzer-winning series exposed it in 2011. The CIA’s own Inspector General reviewed the arrangement and found irregular personnel practices and an embedded officer who believed he faced “no limitations” on domestic activity. The IG found no statutory violation. The statute is the same 1947 charter CHAOS broke; the reading changed.

As for output: in a June 28, 2012 deposition in the Handschu litigation, unsealed that August, Assistant Chief Thomas Galati, commanding officer of the Intelligence Division, testified: “I never made a lead from rhetoric that came from a Demographics report.” The unit had operated for six years. The program mapped a religion and, by its commander’s sworn account, the map led nowhere.

In 1971 the parallel programs were secret and illegal. In 2026 they are integrated and statutory. That is the whole distance between the two eras, and it is the point.

Seven practices, two eras. The third column is the thesis.

The machine

The technology is where the comparison stops being historical, and every piece of it arrives with paperwork attached: a statute, a certification, a contract, a privacy impact assessment. Hoover’s Bureau steamed open letters and paid men to sit in meetings, and hid all of it. The current system runs on purchased data and commercial software, and files the receipts.

PLAIN TERMS · Section 702. A law that lets the government collect emails, calls, and texts of foreigners abroad from U.S. companies, without individual warrants. Americans’ messages enter the pool whenever they communicate with a target, and the FBI can search that pool. A court approves the program each year. It does not approve each target.

The scale layer is Section 702 of FISA, which compels American communications providers to hand over the traffic of foreign targets abroad. Targets grew from 89,138 in 2013 to 291,824 in 2024, by the government’s own annual transparency reports. Americans’ calls, texts, and emails enter the pool whenever they touch a target, and the FBI queries the pool without a warrant.

The government’s own published series. Verified report years shown; the curve only bends one way.

The queries are where the record turns. From December 2020 through November 2021, the Bureau ran queries using Americans’ identifiers up to 3.4 million times, the first year the government ever disclosed the number, with roughly 1.9 million tied to a single Russian critical-infrastructure investigation, per ODNI’s transparency report. After reforms and the 2024 reauthorization’s approval requirements, the reported figure fell to 57,094 in 2023 and 5,518 in 2024, per the Justice Department Inspector General, then rose to 7,413 in 2025. Within that rise, Senator Ron Wyden said in April 2026 that the sensitive subset, searches involving American journalists and political and religious organizations, more than tripled in 2025, and that the Bureau declined to explain why. Then, in March 2026, the FISA Court found the collapse could not be taken at face value: agencies had been running queries through “filtering” tools that went uncounted, a practice the court described as extending across the intelligence community. The opinion is classified. The government’s own compliance numbers are, for now, unverifiable by the government.

The piece’s harder finding in one frame: the reported figures fell 99.8 percent, and the court that polices the system says the recent counts cannot be taken at face value.

The analytic layer is commercial. ICE awarded Palantir a $30 million contract in April 2025 to identify deportation priorities. The State Department’s screening apparatus draws on tools like Babel Street’s “persistent search,” which continuously re-scrapes the open internet on flagged individuals so that a person, once searched, stays searched. ICE’s own August 2025 privacy impact assessment describes social media monitoring analytics deployed against perceived threats to the agency; the Brennan Center documented the tooling’s reach into facial recognition and location tracking. And the data-broker layer sits under everything: agencies purchase Americans’ location and behavioral data on the open market, no legal process attached, a loophole Congress has voted on and failed to close.

The informant corps scaled with the software. Roughly 1,500 Bureau informants in the COINTELPRO era and roughly 15,000 by 2011, by the most careful available accounting, the journalist Trevor Aaronson’s in The Terror Factory, with many tasked into Muslim communities. The provocation model survived intact. In Orange County the Bureau paid an informant named Craig Monteilh to infiltrate mosques in 2006 and 2007; he talked of weapons and jihad until congregants reported him to the FBI and sought a restraining order. The mosque called the Bureau on the Bureau’s own informant. The 1968 playbook, running on 2006 hardware, defeated by the community it targeted.

The corps grew tenfold. The tactic did not change.

On June 12, 2026, Section 702 lapsed. Two short-term patches had carried the statute past its April 20 sunset; the House passed a three-year extension on April 29; Senate cloture failed on June 5; and on June 11 the House voted down a final extension 198 to 218, with most Democrats refusing to extend the authority while Bill Pulte, an appointee with no intelligence experience, was set to assume the role of acting Director of National Intelligence, and a bloc of conservatives refusing to extend it without reforms. Collection continues anyway. The FISA Court approved fresh certifications in March 2026, which remain valid until March 2027, and providers face fines of $250,000 a day for refusing directives issued under them. The central statute of the modern architecture is expired, and the machine has not noticed.

The label moves

Threat definitions migrated the way the 1968 memo’s did: from what people do to what they are to what they might become. The Cold War word was subversion. The post-2001 word was radicalization, a framework under which mosque attendance and markers of religiosity became indicators. In August 2017 an FBI intelligence assessment created a category called Black Identity Extremists, predicting violence against police from perceptions of police brutality. The document leaked and the Bureau renamed the category. The method survived the renaming: threat inferred from identity and grievance.

Six labels, seventy years. Each one moves attention further from conduct.

PLAIN TERMS · Fusion center. One of roughly 80 intelligence hubs created after 9/11 where federal, state, and local police share threat information. They write bulletins. A bulletin is analysis, not a charge: no court reviews it, and there is no formal process to learn you appear in one or to contest it.

The newest label arrived this spring, and it reaches a beat I cover. In May 2026, WIRED obtained more than a thousand pages of DHS, FBI, and fusion center documents showing an emerging threat category the agencies call anti-tech extremism. A December bulletin from Philadelphia’s fusion center assessed that domestic violent extremists are likely interested in targeting AI data centers. Its short-term indicator list ran “disruptive First Amendment activity” alongside online calls to boycott or protest local data centers and criticism of higher utility bills tied to data center power use. Asked about the bulletin, a Philadelphia police spokesperson told The Intercept that fusion center assessments provide situational awareness and are not meant to characterize lawful activity or protected speech as criminal conduct. Northern Virginia’s fusion center applied its anti-government extremist category and logged residents’ appearances at an Arlington County budget meeting and a Fairfax County school board meeting.

The bulletins are not empty of predicate: per the documents WIRED obtained, one flagged a chat in which a user called for violence against executives and power plants, and a January 2025 DHS analysis warned of targeted violence against corporate figures. The pattern the documents show is the familiar one: a real violence concern at the center, and lawful politics swept into the indicator list around it. The movement being monitored is, on the record, zoning and utility politics: Data Center Watch counts well over a hundred opposition groups across dozens of states, with blocked or stalled projects worth $156 billion in 2025. Complaining about an electric bill can now place a resident in a threat bulletin.

The current terminus is running in immigration enforcement, and it is running through Chicago. The State Department’s catch-and-revoke effort applies AI-driven social media screening to visa holders’ speech; Secretary Rubio put the count above 300 revocations by late March 2025, on the department’s stated grounds of national security and foreign policy. Rümeysa Öztürk, a Tufts doctoral student, was taken off the street by plainclothes officers after co-writing a student newspaper op-ed; in February 2026 an immigration judge found no grounds to deport her. A federal judge ruled in September 2025 that targeting noncitizen academics for their views violated the First Amendment. In October 2025, three labor unions sued to stop the surveillance program itself. Operation Midway Blitz brought the enforcement surge to Chicago in late 2025; this May, federal prosecutors dropped, with prejudice, the charges against the Broadview protesters. The University of Illinois confirmed it monitored students’ social media in support of felony protest charges.

The 1968 memo aimed to deny leaders respectability. The 2025 program reads op-eds for visa decisions. The distance between those two sentences is a change in medium. And the arc between them belongs to no party: the Guidelines were Bush’s, the collection peak was Obama’s, the BIE assessment came in the first Trump term, RISAA was signed by Biden, and catch-and-revoke is running now. Four administrations built this; each kept what the last one added.

What improved

PLAIN TERMS · The oversight bodies. The FISA Court: federal judges who approve and police the surveillance programs. The Privacy and Civil Liberties Oversight Board: an independent review agency. Inspectors General: internal government auditors. Their published reports supply most of the numbers in this piece.

A lot did, and complete analysis puts it on the record with the same care as the continuities. The modern system produces its own paper trail: FISA Court opinions, Privacy and Civil Liberties Oversight Board reports, Inspector General audits, annual transparency statistics, published guidelines. Congress debated the 2024 reauthorization in the open and came within a single House vote of a warrant requirement. Every scale figure in this piece exists because post-Church law compels its publication. COINTELPRO produced no such record. Learning of it required a burglary.

The paper trail carries real weight in any assessment of the two eras. A system that documents itself, submits to a court, and loses in litigation differs from a system that killed a sleeping man and filed the memo. The deaths of Fred Hampton, Mark Clark, Bunchy Carter, and John Huggins have no analog in the post-2001 domestic record. The question the documents leave is narrower and harder: whether a pre-crime authority that is published, audited, and lawful produces different outcomes for the communities underneath it than one that was secret and criminal, or the same outcomes with better filing. The Demographics Unit’s zero leads, the FISC’s uncounted queries, and the op-ed detentions are the current evidence file on that question.

What would change this analysis

Three tests sit on the calendar, registered here before they resolve. Declassification of the March 2026 FISA Court opinion will show whether the reported collapse in FBI queries was real; the chair and ranking member of the Senate Intelligence Committee demanded its release by mid-May 2026 and have been ignored, which is itself a data point on how the self-reporting system performs under pressure. The Section 702 fight will show whether Congress restores, restrains, or abandons the central collection authority. The catch-and-revoke litigation will show whether courts treat speech-based immigration surveillance as the First Amendment violation one district court has already found. This piece predicts continuity in the threat-definition pattern and no narrowing of the assessment authority. If the FISC opinion shows the query decline was genuine, if the 2008 framework is narrowed, or if the terminus programs are struck down and dismantled, the legalization thesis loses force, and a follow-up will say so.

Coda: the memo in the culture

The Bureau’s vocabulary outlived the Bureau’s secret. D’Angelo titled his 2014 album Black Messiah. The 2021 film Judas and the Black Messiah took its title straight from the March 1968 memo and put Hampton’s killing in front of an Oscar audience; fact-checkers verified the memo’s language for general readers, which makes the founding document of the Black Nationalist program one of the most publicly authenticated texts the Bureau ever produced. Hip hop carried the program’s name across three decades, from Public Enemy’s wiretap paranoia to Rakim on the 2021 soundtrack. The words written to prevent a messiah became the culture’s name for the man killed in his bed in Chicago with the Bureau’s floor plan.

The 2008 Guidelines have no film. They have a press release and a fact sheet. They are in effect today.

Disclosure. Research, verification, and visual assets for this piece were produced in part with AI tools, including Anthropic’s Claude. Anthropic holds commercial stakes in the AI data center buildout this piece discusses. Every claim traces to the cited primary documents.

Comment requests. Comment requests were sent before publication to the FBI National Press Office, the Justice Department’s Office of Public Affairs, ODNI’s Civil Liberties, Privacy, and Transparency office, the State Department, the NYPD, and the Philadelphia Police Department, each with a documented response deadline. Responses will be appended to this piece as they are received. The Philadelphia Police Department’s on-record statement to The Intercept is cited above.

What happened: Illinois built a $100 million State of Israel bond position through repeated purchases of $5 million to $20 million over a decade, most recently at the widest spreads in the record. Why it matters: the office’s signed FOIA determination states that no purchase memoranda and no due diligence documentation exist for the position; the only analytical exchange in three years of produced correspondence is a purchase committed two hours and thirty-three minutes after the seller’s pitch. The state repeatedly lent public money to a foreign government with no written reasoning of its own anywhere in the produced record. What’s pending: a Public Access Counselor review of one determination, and a second records request on the Republic of Poland holding in the same portfolio category.

On February 26, 2025, at 10:47 in the morning, Larry Berman sent an email to the Illinois State Treasurer’s office. Berman is the National Managing Director for Corporate and Institutional Sales at Development Corporation for Israel, the FINRA-member underwriter that sells State of Israel bonds in the United States. His email went to Elizabeth Turner, the office’s Chief Banking Officer. Treasurer Michael Frerichs was copied.

Berman’s pitch was specific. Israel bonds were paying spreads over comparable U.S. Treasuries of 64 basis points at two years, 104 at three, 159 at five, and 172 at ten. He called the spreads “historically very strong” and suggested the office place its order that day or the next. A rate sheet and the current offering circular were attached.

At 1:20 that afternoon, two hours and thirty-three minutes later, Turner replied. The office would proceed with a $10 million purchase of the five-year bond at 5.72 percent. Four questions followed. They concerned the execution documents, the wire instructions, the wire date, and the trade confirmation. The money moved on March 3.

Taken alone, this is an efficient transaction. What gives it weight is its context, and the context is now a matter of public record, because I asked the office for all of it.

On June 11, 2026, I submitted a seven-category request to the Treasurer’s office under the Illinois Freedom of Information Act, covering the office’s State of Israel bond holdings from June 1, 2023 forward. The office responded on July 1 with a signed determination letter (Exhibit 2) and a production of 587 documents. The February 2025 exchange is in the production (Exhibits 3a–3d). It is the only analytical conversation in it. And the categories that would contain the office’s own analysis came back with a specific answer: no records responsive.

The documents cited in this piece are collected at christopher-sweat.com/documents/100m-no-memo. Inline citations link to the files there.

What the office produced

The office handled the request properly. It invoked a lawful extension with statutory citations and responded inside the extended window. It also produced more than the request required. The request excluded routine transactional confirmations from the correspondence category; the office produced them anyway, and those confirmations turned out to be the most informative documents in the file. Nothing in this piece rests on a procedural failure by the office. The findings rest on what the office, following procedure, said it has and does not have.

The seven determinations, from the July 1 letter signed by FOIA Officer Erica Tremble (Exhibit 2):

Categories 3 and 4 are the anchor of everything that follows. In the office’s own words, it does not have records responsive to a request for the memoranda supporting each purchase of a $100 million position, and it does not have records responsive to a request for the due diligence documentation connected to those purchases.

Category 2 contains a drafting problem worth stating precisely. The request asked for final approved comparative analyses. The response withholds records, to the extent they exist, as “preliminary drafts, notes, recommendations, memoranda and other records in which opinions are expressed, or policies or actions are formulated.” A final approved analysis is not a preliminary draft. The response supports two readings. Under the first, no final comparative analysis exists, and the exemption is invoked against materials the request did not seek. Under the second, a final analysis exists and is being withheld under an exemption that does not cover it. I have filed a Request for Review with the Public Access Counselor at the Office of the Attorney General asking which reading is correct. That review is pending.

What the file shows

The produced records, joined to the office’s public Vault data, reconstruct the position completely. The two sources reconcile to the dollar in every month where they overlap.

The program is old. Security-level records show State of Israel bonds in the portfolio since at least July 2016, purchased in $10 million blocks at coupons between 1.53 and 3.53 percent, held to maturity, and rolled. The post-October 2023 change is one of scale and price, not of existence.

The current position, from the office’s Computershare interest advice of May 1, 2025 (Exhibit 4), cross-checked against trade confirmations (Exhibits 5a–5f) and the Vault’s security-level data (Exhibit 13a):

The coupon ladder tells the price story in one column. The state lent to the same sovereign at 1.20 and 1.35 percent in February 2021. It re-lent at 5.68 and 5.72 percent in early 2025. In February 2026 it rolled the maturing 2021 paper into a new five-year at 4.86 percent (Exhibit 6a, Exhibit 6b).

Coupons move with Treasury rates, so the coupon alone conflates two things. The spread over the Treasury curve at each purchase date separates them. The table below uses the official daily par yield curve published by the U.S. Treasury, matched to each bond’s issue date (working values at Exhibit 13b). Two purchases carry a second figure because the seller’s own emails state the spread as priced off DCI’s two-week period benchmarks, which differ from the official close by a few basis points.

A note on the last row. My earlier pieces carried the office’s characterization of the February 2026 renewal as roughly 114 basis points over a comparable Treasury. Against the official curve on the first business day after issue, the figure is 103. The difference is the choice of benchmark, and the table above states each figure with its method. The refinement bears on the decomposition from the first piece in one direction only: the credit and illiquidity components were sized against the office’s 114, so a renewal spread of 103 leaves the residual excess return at or below zero. The measured figure tightens the original conclusion. The same treatment applies to the seller: Berman’s February 2025 pitch claimed 159 basis points against his period benchmark, and the official close that day makes it 166.

Precision also matters on October 2023. The $10 million bond in the ledger carries an October 1 issue date. October 1 precedes October 7. The record produced cannot resolve when the order was placed relative to the attack, and this piece does not describe that purchase as a response to it. The purchases that unambiguously follow October 7 are the two November 1 trades totaling $20 million. What the record does establish is the tempo of the correspondence in that window: the produced record runs in the single digits per month through September 2023, then 62 documents in October, 57 in November, and 59 in December, nearly all of it the seller’s outbound content.

The record also shows the office declining. In April 2025, Berman wrote to flag that the office had $5 million maturing May 1 and to recommend reinvestment at what he again called stronger-than-usual spreads (Exhibit 7a). The office let the bond mature without a roll. In June 2025, Berman pitched a three-year at 4.96 percent, 113 basis points over his benchmark, a figure the Treasury data confirms almost exactly (Exhibit 7b). No reply appears in the production, and no purchase followed. The office says yes, and the office says no. What the file lacks is documented reasoning behind either answer.

The spread, revisited

My first piece argued that the spread on these bonds is compensation, priced by the market, for sovereign credit risk and for a non-tradable structure, with nothing left over as excess return. The second piece reported the Bank of Israel’s own confirmation that the risk premium had compressed to near its pre-war level. The FOIA production adds a fact the framework could not supply on its own: the position’s largest recent purchase was made at the widest spread in the record, on the seller’s word that the width was strength.

The width was the market’s price on Israel’s credit. Here is the rating history over the holding period, from the agencies’ actions:

The February 26, 2025 order was placed five months after the last of five downgrades, with all three agencies on negative outlook. The 166-basis-point spread the office bought that day was wide because the market had repriced the sovereign. “Historically very strong spreads” and “historically elevated credit risk” describe the same number from two directions. The seller chose the first description. The produced record contains no document in which anyone at the office weighed the second.

The governing investment policy requires foreign sovereigns to be rated in one of the three highest classifications by at least two standard rating services. At every 2025 purchase, S&P and Fitch stood at A, and Moody’s stood at Baa1, below the line. Two of three agencies satisfied the floor. The purchases were compliant. The margin was zero.

The vocabulary of the production makes the same point at the corpus scale. Across all 587 documents, three years of correspondence between the office and the seller of a nine-figure position, the phrase “due diligence” appears zero times. “Comparative” and “comparison” appear zero times. “Risk assessment” appears zero times. The analytical vocabulary that does appear, spreads, and benchmarks appears in the seller’s rate pitches and in the one February 2025 exchange. The only analysis present in the file is the seller’s.

The Poland position

The comparator was in the portfolio the whole time. The office’s own FOIA response volunteered it: the Foreign Bonds category holds $15 million in Republic of Poland bonds alongside the Israel position.

The Vault identifies the bond. It is the five-year tranche of Poland’s March 11, 2024 dollar benchmark, a $1.5 billion issue carrying a 4.625 percent coupon and maturing March 18, 2029. Poland’s Ministry of Finance published the pricing: 70 basis points over the comparable Treasury, yielding 4.797 percent at issue (Exhibit 12 links). Illinois’s records show the position at a cost of $14,918,400 against $15 million par, a price of 99.456, added to the portfolio in January 2025 (Exhibit 13a).

Set the two holdings side by side. Poland, rated A2, A-, and A-, was bought through a broker at a market price below par, in an instrument priced by a $1.5 billion syndicated book at 70 basis points over Treasuries. Israel, rated Baa1, A, and A at the time, was bought at par at the seller’s posted rate, at 159 to 166 basis points over Treasuries, one month later. Same portfolio category. Same buyer. Same quarter.

An 89-basis-point difference between two sovereigns in the same allocation is precisely the object of a comparative analysis. The word “Poland” appears nowhere in the three years of produced correspondence. Whether the Poland purchase includes the memoranda and due diligence documentation that the Israel purchase lacks is a question that the same statute can answer, and on July 1, I filed a second FOIA request seeking them. If records exist for Poland and not for Israel, that is a finding. If records exist for neither, the gap is portfolio practice, and the Israel position is its most visible instance. The response is due this month.

Policy versus record

The purchases were lawful. That conclusion is worth stating plainly because it concentrates the weight on the missing documents.

Two versions of the State Treasury Investments Portfolio Investment Policy Statement govern the ledger. The version approved on October 26, 2022, and effective on December 1, 2022 covers every purchase through October 2025, including all of the 2025 acquisitions. The current version, effective October 22, 2025, covers the February 2026 rollover. Both are linked at Exhibit 12. On the sections that matter here, the two versions are word-for-word identical.

The policy requires competitive bids from at least three broker-dealers before a purchase, and it exempts reverse inquiry investments and new-issue purchases from that requirement. Every DCI purchase in the record is a dealer-initiated offer of a new issue. Each one ran through the exemption lawfully. No violation occurred, and no bidding file exists because none was required. The consequence is structural: for this position, the documentation of judgment cannot live in a bidding record. It can only live in the memoranda, the due diligence files, and the comparative analyses. Those are categories 3, 4, and 2. Two are empty by the office’s own statement, and the third is withheld as preliminary, if it exists at all.

The policy caps foreign government securities at three-quarters of one percent of the portfolio, with a five-year maturity limit. At no point in a decade of data has the position approached the cap; peak usage was 0.480 percent in February 2020. Every holding sits at or under the maturity limit. The compliance record is clean, and this piece makes no claim to the contrary.

One difference between the two policy versions bears on the empty categories. The 2022 version seats a Director of Investment Analysis and Due Diligence on the Investment Policy Committee. The 2025 revision drops the title from the committee roster. The office’s no-records response is specific to these purchases, and nothing in the file speaks to the position’s work elsewhere in the portfolio. The juxtaposition is stated here without a conclusion because the record supports none.

What the file documents, at length, is the register of the relationship the correspondence carries. 84% of the production consists of DCI marketing, event invitations, and appeal content addressed to the office. Substantive office responses make up roughly eight percent. About a dozen documents in the entire production contain financial or analytical content of any kind. Three items from the marketing stream describe the register better than any characterization could.

On May 28, 2024, DCI’s marketing list delivered an email to the Treasurer’s own address under the subject line “They say divest, we say double down” (Exhibit 8). Its text: “The most direct and impactful way to combat BDS is by making additional investments.”

On November 21, 2023, three weeks after the office’s $20 million November purchase and one month after Moody’s placed Israel on review for downgrade, DCI marketing quoted a Moody’s publication calling the bond program “a stable and reliable source of funding” (Exhibit 9). The email’s own footnote reads: “*Israel bonds are not rated.” The instruments Illinois holds carry no rating of their own. The policy’s eligibility test depends on the sovereign’s rating, and the seller’s disclaimer states this distinction precisely.

On May 16, 2024, after the year’s first two downgrades, the Accountant General of the State of Israel, Yali Rothenberg, wrote to Treasurer Frerichs directly, copying DCI’s chairman and its chief executive (Exhibit 10). The letter told the Treasurer that his decision “not only aligns with this fiduciary duty but also reflects a principled stance in support of a longstanding ally,” and spoke of the investment’s “symbolic importance.” The issuer’s own senior fiscal officer, describing the holding, went beyond the financial vocabulary in a single sentence.

The file also shows the relationship running in the other direction. On November 9, 2023, eight days after the November trade date, the Treasurer sent DCI’s Midwest office a recorded video of his remarks, from his phone, for use at an event that Sunday (Exhibit 11).

The answer on file

The question my first piece raised was procedural. Given a universe of statutorily eligible foreign sovereigns paying market-clearing spreads, what process selects this one, and where is it written down? Existing law already obligates public fiduciaries to be able to answer.

The office has now answered, in a signed determination. For every purchase and renewal the request reached, back to June 2023, the memoranda do not exist. The due diligence documentation does not exist. The comparative analysis, if it exists, exists in a form the office calls preliminary. The selection process that has chosen this security, purchase after purchase across a decade, produced 587 documents of the seller’s correspondence, one afternoon’s exchange, and no file of its own for the window examined.

That is not a bad answer. It is no answer.

The record stays open. The Public Access Counselor’s review of the category 2 determination is pending. The Poland request is pending. Another $20 million of the position matures this November, and the next decision will be made under the same policy, by the same office, with the same empty file, unless the office fills it. It can. The policy describes the process; the process would produce documents; the documents would answer the question. The office was provided a summary of these findings on July 1 and invited to comment or correct any error by 5 PM on July 6. No response was received before publication. Any response received afterward will be reflected in this piece.

Methods and record

Spread figures are the difference between each bond’s coupon at issue and the U.S. Treasury par yield at the matching tenor on the issue date, or the nearest business day where the issue date falls on a weekend, from the Treasury Department’s published daily curve. They are coupon spreads on non-tradable instruments, not market option-adjusted spreads. Where the seller’s correspondence states a spread against its own period benchmark, both figures appear with the method identified. The holdings ledger derives from the office’s produced interest advice and trade confirmations, reconciled against the security-level Security Holdings Report in the Treasurer’s Vault; the two sources agree to the dollar in every overlapping month. Corpus statistics cover all 587 produced documents, including optical character recognition of image-only files. The document production is replicable in full: any person may submit the request at Exhibit 1 under 5 ILCS 140. The working data are in Exhibit 13a and Exhibit 13b.

Update: July 10, 2026

The office has responded to the follow-up request on the Republic of Poland holding, the control case this piece flagged as pending. Two findings follow. Both belong in the record.

First, the formal documentation gap is portfolio-wide. For Poland, as for Israel, the office reports no due diligence documentation. It reports no comparative analysis of Poland against other eligible sovereigns. It claims the preliminary-drafts exemption for any memoranda supporting the purchase, the same exemption it claimed for the Israel comparative analyses. On paper, the two files carry matching determinations.

Second, the produced trade file shows the process behind the Poland purchase, and it does not resemble the Israel record. On January 29, 2025, a portfolio staffer emailed the Chief Banking Officer a written comparison of three Poland issues on spread, yield, price, and call structure, with market data attached, and recommended one: “I think the 2029 issue is best.” Two hours later the Chief Banking Officer replied: “I have received the approval to purchase.” The office then bought the bond through a competitive broker at a market price below par, and the file preserves a multi-dealer quote screen and three months of price history. Twenty-eight days later, the same Chief Banking Officer committed $10 million to Israel bonds two hours and thirty-three minutes after the seller’s pitch. The 587-document Israel production contains no written comparison of alternatives and no market check.

Both purchases closed the same day they started. The difference is what filled the hours.

The Poland production is posted in the document archive. A narrowed request for the communications behind the Poland purchase is now pending, and the production has been submitted to the Public Access Counselor as supplemental material in the pending review. A full analysis of the Poland file will follow as the fourth piece in this series.

Sources: Office of the Illinois State Treasurer FOIA response and document production, July 1, 2026 (all quoted correspondence is from that production); Office of the Illinois State Treasurer, The Vault, Security Holdings Report; State Treasury Investments Portfolio Investment Policy Statements effective December 1, 2022 and October 22, 2025; U.S. Department of the Treasury, Daily Treasury Par Yield Curve Rates; Republic of Poland Ministry of Finance, pricing announcement of March 11, 2024; Moody’s, S&P Global, and Fitch rating actions of February 9, April 18, August 12, September 27, and October 1, 2024, as contemporaneously reported and per the agencies’ releases.

Written in a personal capacity. Views are my own and do not represent GrayStak or any affiliated organization. The author holds no position in any security discussed, has no compensated relationship with the Office of the Illinois State Treasurer, Development Corporation for Israel, or any divestment coalition, and was not paid to write this piece. A Request for Review of one determination in the office’s FOIA response is pending before the Public Access Counselor; a second FOIA request concerning the Republic of Poland holding is pending before the office. Statutory and policy interpretations reflect the author’s reading and have not been adjudicated. Spread figures are coupon spreads over the Treasury par yield curve at issue, not market option-adjusted spreads; the instruments are non-tradable. This piece is procedural and analytical. It takes no position on the underlying sovereign, the conflict, or state foreign policy. Not investment advice.

Christopher Sweat (christopher-sweat.com) is a political risk analyst and the co-founder and CEO of GrayStak, a media and political risk company that covers social and political dynamics in real time and produces the Domestic Political Volatility Index (DPVI), an institutional-grade signal product. He is based in Chicago. This piece is the third in a series applying sovereign credit analysis to Illinois public investment, following “The Spread Isn’t Alpha” (May 25) and “The Cut and the Choice” (May 27). The series, including the underlying FOIA work, is reported and written in a personal capacity.

Berg has spent a decade on that argument. He runs the Chicago Policy Center, an arm of the free-market Illinois Policy Institute, writes The Last Ward to more than 158,000 subscribers, and co-wrote The New Chicago Way. What he wanted to litigate was the machinery.

“Among the big cities, we have the most vestiges of a strongman, almost caudillo, sort of system,” he said. A weak council, often seated by the mayor. No city charter. No elected finance officer or city attorney. No public vote on debt or taxes. Among big cities, Chicago stands out for having none.

Waiting for Superman

I asked whether the 2027 mayor’s race matters or whether it is a field competing to run a broken machine. Berg has a name for the reflex behind the question.

“The mayor’s race is the lens through which Chicagoans understand city politics more than anything else,” he said. “This is a pathology.” He calls it waiting for Superman. The city looks to one person for salvation because the strong-mayor system trained it to.

The office matters to him. What the city gets wrong, he says, is what it expects the office to do. “It’s a necessary condition to have good leadership,” he told me. “It’s not a sufficient condition to have a growing, thriving city.” The candidates he watches for are the ones who ask a longer question.

“Public office does not create intellectual capital. It consumes intellectual capital.”

Most people who run for mayor arrive without much of it, he said, because the way the city actually works sits below the headlines. His group briefs candidates across the spectrum on their real options. The ones who stand out, in his telling, ask what they can build now that leaves the city stronger on its 200th birthday in 2037, whoever holds the office by then.

The $11 billion

To see the structure work, Berg said, look at last year in Springfield. A pension sweetener added about $11 billion in liabilities to Chicago. The immediate annual cost he put at $30 to $50 million, enough to double the city’s summer jobs program, one of the few Johnson initiatives he credits.

The mechanics are what he returns to. “They just made the promise. They didn’t make the tax,” he said. The state assigned the liability and left the city to find the money. Chicago got no vote, and by his account, learned the full scope only days after the session closed.

The city’s chief financial officer at the time, Jill Jaworski, warned in writing that the change would push the pensions toward insolvency. Berg says he was the first to publish that letter. The mayor did not fight it in public. Berg reads the silence as a symptom of the same structure. A mayor with little standing in Springfield absorbs a bill the city did not get to weigh.

There is another read of the deal. The Tier 2 benefits it raised had drifted below the federal Social Security “safe harbor,” the rule that a public pension must pay at least as much as Social Security would. Falling short risked forcing Chicago to enroll those workers in Social Security or cover the gap through lawsuits, an exposure that the nonpartisan Center for Tax and Budget Accountability estimated at over $800 million a year. Roosevelt University’s Ralph Martire, who directs it, told WTTW the law fixes a real structural flaw, while faulting how it was paid for. Berg’s quarrel is with that second part, the promise without the money.

A release valve

Other cities have an exit from a bill like that. Chicago does not. “Detroit today has a better credit rating than the city of Chicago,” Berg said, and Detroit is a decade past bankruptcy. Chicago is the only city among the fifteen largest that cannot take its debts to a bankruptcy judge.

He calls Chapter 9 a last resort and wants a framework for distressed cities in place before it. An early-warning system on municipal finances. State oversight when a city runs past its means. An emergency manager who can reach a judge only at the very end. Harvey, the south-suburban city that cannot collect enough property tax to cover its debts, is his warning case.

The authority matters even when unused, he argues, because it rewrites the conversation with creditors and unions. “We can either negotiate in good faith right now,” he said, describing the leverage, “or this can go to a judge, and then it’s out of our hands.” He points to the parking meter deal as the price of having no such card to play.

The case against the tax

Berg’s tax politics start with population. “Our biggest problem as a state is three words,” he said. “People, people, people.” Illinois lost roughly a million residents over the past decade, and by his reading, the highest earners leave fastest, two out for every one in.

That is the frame for his objection to the proposed 3 percent surtax on income above $1 million. He is careful about who he is defending. “I’m not asking people to shed a tear for a pinstripe-suit, Rolex-wearing millionaire,” he said. His worry is the roughly 22,000 pass-through businesses that file at that line, the tradespeople and small operators he credits with most of the state’s new jobs, and the tax base that leaves with them. He points to Massachusetts, the one state to run a comparable surcharge. “The entire pie of the income in Massachusetts shrunk,” he said.

That read is contested, the migration piece most of all. Cornell sociologist Cristobal Young, working with two US Treasury economists, tracked every million-dollar tax return in the country over thirteen years for a study in the American Sociological Review. The rich, they found, change states at 2.4 percent a year, compared with 2.9 percent for everyone else. Tax flight runs only at the margins. The Illinois Economic Policy Institute reached a parallel result in 2026, finding that states with millionaires’ taxes grew about as fast as those without, and that a 3 percent version could raise more than $4 billion a year for schools or property tax relief. Illinois voters backed the idea in a 2024 advisory referendum, roughly 60 percent in favor. GrayStak holds no position.

Where he draws the line

On the state’s investment exclusion lists, Berg draws a line between the money a government spends and the money it owes its retirees."The Illinois Investment Policy Board keeps a set of those lists, built one cause at a time. Companies that boycott Israel. Firms that house migrant children. Businesses tied to Iran and Russia. Every state pension fund has to divest from the names on them and stay clear, and much of that territory already overlaps with federal sanctions. I asked whether a fiduciary duty, applied evenly, cuts against every screen on the list, whatever the cause.

He said it is not an issue he has worked on closely, then split the question. A state can invest by its values, the way a university weighs its endowment, and he recalled the campus fights over how endowments should be run. “You can make values-based decisions on how you want to invest,” he said. “What I don’t like is when people think there’s no trade-offs to that.”

The pension is where he holds the line. "If I'm personally responsible for managing the money of Illinois teachers' pension funds, and I need this money to exist or else someone's not getting their retirement check," he said, "my operating principle would be, you have to maximize the returns. That is the single thing you should be focused on." Political screens, whatever the cause, are "a Pandora's box" that reduces returns by necessity. Values-based choices he keeps to the operating budget, the programs a government chooses to fund.

He pointed to a case. Former Chicago City Treasurer Melissa Conyears-Ervin moved to halt the city's purchases of US Treasuries, citing the city's ability to earn the same returns elsewhere. Berg's answer was that Treasuries are the safest and most liquid asset there is, and a global benchmark. "If you're going to counter the benchmark, you need a good benchmark to counter." Pressed on what replaces it, he said no answer comes back. "China doesn't exclude the US Treasury."

Applied across the board, his answer holds for every list, regardless of the target. A political screen on pension money is a political screen, and by his principle, it works against the funds it is meant to protect.

Show your work

Berg likes the ideas grouped under “abundance,” which he reads as a catch-all for sustainable growth, and holds the label itself at some distance. His caution is with how the word gets used in a campaign. Every challenger will promise to grow the city, he said, because growth lets them avoid having to name a tax hike or a cut. “People should ask for specifics,” he said. “What do you mean by growth? What levers will you actually pull?”

His own answers start with certainty. A business plans around what it can see, and Chicago gives it little to see. “Imagine plopping down a hundred-million-dollar investment on the city of Chicago when you have no idea what it looks like ten years from now,” he said. Budget and tax certainty that lets an investor project a decade out is his first lever. Housing supply is the second. He credits the current administration with some progress on building, says it is not enough, and puts a yes-in-my-backyard stance in the column of things that would help.

Subtraction

The waiting-for-Superman reflex, Berg said, has a twin in public safety. Every flashpoint produces a call for a new office. “The political muscle, similar to this waiting-for-Superman muscle, is that we need to create another new body,” he said. Chicago already runs around eight with some oversight of policing.

He faults the city for looking inward. On the teen “takeovers” that have flared downtown for more than a year, he points to Washington, where the mayor and council passed targeted curfew zones with advance notice and services attached. “They tried it out last year,” he said. “Zero teen trends happened” before the measure lapsed and the problem returned.

His fix is consolidation. He cites two recent Matrix consulting reports, which he calls the largest review of Chicago policing since the study that produced the consent decree after Laquan McDonald’s murder, and argues the city cannot act on them while it keeps standing up new agencies. Los Angeles and its police commission are his model. He notes that the CARE mental-health response is “floundering,” in the Tribune’s words, and that the mayor’s office took down its public dashboard. Chicago police data show homicides up about 8 percent from a year earlier, a real uptick that sits on top of 2025, the city’s lowest homicide year in six decades.

What comes next

Berg wants Chicagoans to lift their heads toward other cities for solutions and to press the 2027 field on what they will leave behind. He and I plan to talk again in the fall, closer to February. His test for the candidates comes down to one question.

“What things can I do as mayor today to structurally strengthen the city, so that it’s stronger no matter who is the mayor?”

Christopher Sweat is an analyst working across finance, technology, and politics and the co-founder and CEO of GrayStak, a media and political risk company covering social and political dynamics and institutional politics at the state, local, and federal levels. He created GrayStak’s Domestic Political Volatility Index (DPVI), which measures political escalation ahead of media consensus. Austin Berg is executive director of the Chicago Policy Center. Quotes are lightly condensed from the recorded conversation for clarity. The full video interview is above. GrayStak holds no position.

On June 17, 2026, from a room at Versailles, Donald Trump signed a memorandum of understanding to end a war the United States had started in February by killing Iran’s supreme leader. Pezeshkian signed for Iran from his own capital. They never shared a room. The framework that made the signing possible came together in Islamabad, where Pakistan’s prime minister announced it, while Oman and Qatar worked the channels that kept Tehran and Washington in contact through the spring. The document carries the Pakistani capital’s name, the Islamabad Memorandum, a fourteen-point framework, and its terms run toward Iran: at least $300 billion in reconstruction, an end to American sanctions, immediate waivers for Iranian oil, the release of frozen Iranian funds, and the withdrawal of American forces from Iran’s approaches. Its opening clause comes before any of that. It commits both sides to end military operations on all fronts, Lebanon named in the text. Israel, whose forces are the ones striking in Lebanon, did not sign, and says the memorandum does not bind it.

Get more from Christopher Sweat in the Substack app

Available for iOS and Android

Get the app

The war began very differently. The United States picked the day it started. It killed Ali Khamenei in the opening hours, forced open and then closed Iran’s skies, and ran a naval blockade the administration said cost Iran half a billion dollars a day. No regional force matched it. Yet when the war had to end, Washington could not set the terms it wanted, and the talks that produced it happened in another country’s capital. The guarantee that makes a settlement hold was not Washington’s to give. The country that began the war when it chose needed other countries to finish it.

The strikes opened at 1:15 a.m. Eastern, and Trump announced them in an eight-minute Truth Social video shortly after 2 a.m., his face half hidden under a white USA cap, recorded at Mar-a-Lago, where he ran the opening hours from a makeshift situation room behind the ballroom after telling a black-tie gala to have a good time. He notified the Gang of Eight and gave no public address. The war began during active negotiations and without authorization from the UN Security Council, and the Council met only to react, in an emergency session the day the strikes landed. Eleven days later it condemned Iran’s retaliation against the Gulf states and left out the strikes that opened the war, with China and Russia abstaining over the omission. A war this size once ran through the Security Council. This one ran through a club ballroom and a 2 a.m. video.

That distance matters more than whether the United States won or lost. The capacity to start a war and to end one on your terms now belongs to different countries, and the United States holds only the first.

What It Takes to End a War

Winning the exchanges of fire does not hand you the close. Ending a war requires a separate set of capabilities: a neutral setting that the other side will enter, a broker both sides trust, a guarantor whose commitment the loser finds credible, and terms that the stronger party can impose. The United States went into this war with the firepower and little else, and the memorandum records the gap.

The firepower was never in doubt. The United States and Israel had rehearsed it in a twelve-day war in 2025, and the strikes across both campaigns set Iran’s enrichment and metal work back by years. Operation Epic Fury opened on a date fixed weeks in advance, while the Geneva talks were still reporting progress, which tells you Washington chose the war and did not stumble into it. On the first morning, it killed Ali Larijani, a central figure in Iran’s negotiations. In the same hours, a double-tap strike with Tomahawk missiles destroyed a girls’ elementary school in Minab, killing at least 165 people, most of them girls between seven and twelve. Independent investigations, and the Navy’s confirmation that it fired Tomahawks into southern Iran that day, point to an American strike that neither Washington nor Israel has claimed. The United Nations, Amnesty International, and more than a hundred members of Congress have demanded an accounting. When Iran answered, closing the Strait of Hormuz and firing missiles and drones at the Gulf states that hosted American forces, the United States forced the strait back open with an air campaign and sealed Iran’s ports with a blockade. The strikes did what overwhelming force does, and on the first morning that included a collapsed schoolhouse in Hormozgan.

The rest of the list lay beyond Washington’s reach, and the memorandum catalogs it. Washington could not dictate the terms, so the text hands Iran the money, the sanctions relief, the funds, and the withdrawal, and asks in return a renewed pledge not to build a bomb and a strait kept open for sixty days before the question reopens. The talks that broke the deadlock were held in Islamabad, not on American soil. The security guarantee had turned reactive. Washington refused Saudi Arabia a binding commitment, and Riyadh pulled its defense closer to Pakistan; then weeks later it handed Qatar an explicit pledge of defense after Israel bombed Doha. The promise that once anchored a regional order now arrives case by case, after the strike, to whoever was hit last.

A skeptic will say Washington did not lose the power to dictate terms so much as decline to pay for them, and the objection has force. About a fifth of the world’s oil and gas passed through the Strait of Hormuz, so closing it and blockading Iran’s ports drove fuel prices up and strained the economy Trump needed steady before the November midterms. The United States could have held out for more, and chose not to. But the reason holding out grew so costly is the reason the terms slipped away. With the broker, the guarantor, and the exit all in other capitals, each added week of war bought Washington pressure and no bargaining leverage. A preponderant power that must pay other states to end its wars bargains against a clock those states control. Calling it a choice does not make it less of a loss.

The battlefield did not produce this gap. The United States won there. The shortage behind it predates this war. Thomas Schelling described it sixty years ago in Arms and Influence, separating brute force, which seizes what it can reach, from coercion, which makes an adversary accept your terms. The United States brought the brute force. Coercion runs on the adversary’s consent, and that consent stayed outside what the strikes could reach. Fred Iklé, in Every War Must End, gave the problem of war termination its modern form: that governments are built to start wars and poorly equipped to finish them. This war added what neither anticipated. The trusted broker and the credible guarantee had left American hands for other capitals before the first plane took off.

That older work of ending wars has a name. Centuries before Grotius, the Hanafi jurist Muhammad al-Shaybani set down a systematic Islamic law of nations, the siyar, organized around treaty-making, reciprocity, and the conduct of truces with adversaries. The renewable sixty-day truce at the heart of the memorandum is a form that tradition worked out more than a thousand years ago. Reading the settlement only through Geneva and Grotius overlooks the legal world in which the parties themselves were operating.

The Clock and the Threat

Who chose the moment is contested. Hawks treat the strike as a move in the contest with China, a way to take an Iranian asset off Beijing’s board. The administration’s own account points elsewhere. Secretary of State Marco Rubio told congressional leaders on March 2 that the United States struck because it knew Israel was going to act against Iran regardless, and that American forces would draw Iranian fire once it did. The justification shifted the next day to an imminent Iranian attack, one of several rationales that rotated through the early days. Underneath the shifting was an alliance dynamic the theorist Glenn Snyder named decades ago: entrapment, the pull that drags a strong state into its ally’s war out of fear the ally will fight alone. The United States did not set this war’s clock. Israel set it, and the alliance’s pull brought Washington along.

The clock sat inside a threat picture both governments had spent years assembling. Washington and Israel had come to treat Iran’s allied forces, Hezbollah in Lebanon, the Houthis in Yemen, the militias in Iraq, Hamas in Gaza, as one enemy, the Axis of Resistance, and they made taking it apart an explicit aim of the war. The American precondition for talks was that Iran end all support for the network, and the plan was to dismantle the whole structure and move the fight to Iran itself. Israel had signaled the approach for years, in Netanyahu’s doctrine of the preemptive strike and in a campaign of assassinations that killed the network’s leaders one after another. The framing did the work that threat framings do. It collapsed a set of distinct actors, each with its own base and its own quarrels, into a single existential enemy, and promised that breaking the enemy would buy security.

The war combined two different threats. Iran’s nuclear program was a real concern, and the strikes set it back. The Axis of Resistance was not the same kind of threat, and the war tested the promise that breaking it would buy security and found it thin. Years of killing the network’s leaders had not dissolved it, and through this war parts of it fought while others held back, since defending Iran was never the first priority for all of them. The name itself runs two ways. To Tehran it is a banner of deterrence, the forward defense it spent two decades building. Washington came to use the same words for a monolith it meant to destroy. Security studies call the move securitization, the building of an adversary into an existential threat to license what would otherwise be unthinkable, and the Iraq war is the nearest reminder of where it leads. The Axis framing drove the United States into a war it could open at will, fought against a network its strikes never erased. The threat picture promised force would settle the matter, and the firepower, the strike on Minab among it, secured no terms.

This war was the newest front in a longer campaign. Since October 2023, Israel has used the freedom of action its firepower buys, with American weapons and American cover, across Gaza, Lebanon, Syria, and the Gulf. On September 9, 2025, Israeli jets fired ten missiles into a residential compound in Doha, trying to kill Hamas negotiators on the soil of Qatar, a country that hosts a major American base and was at that moment mediating a Gaza ceasefire. The strike failed and killed a Qatari security officer among others. The same freedom of action runs through Gaza, where the record is now formal. In September 2025, the United Nations Independent International Commission of Inquiry concluded that Israel is committing genocide against Palestinians in Gaza, finding four of the five acts defined in the 1948 Genocide Convention. The International Association of Genocide Scholars reached the same conclusion in a resolution that month. The case South Africa brought before the International Court of Justice in December 2023 remains open, and the International Criminal Court issued arrest warrants for Netanyahu and Gallant in November 2024. Israel rejects the findings, calling them ridiculous and a blatant falsehood and asserting its right to self-defense. The Iran war belongs to that campaign, where the same force met the limit of what it can settle, and where the United States, having armed it, could not end the war it produced.

Washington and Tel Aviv were not the only hands on the outcome. Mohammed Ayoob’s subaltern realism holds that postcolonial states act first on the insecurity closest to home, the survival of the regime, and only then on the roles the great powers assign them. Iran’s choices, the Gulf’s hedging, and Pakistan’s brokerage followed that logic, and the settlement emerged from those moves.

The Tempo Without a Table

The settlement ran on a faster and choppier clock than war termination used to, and the reason sits in the missing table. Watch the talking instead of the fighting and the pattern comes clear. In three and a half months the two sides moved through a Pakistan-brokered two-week ceasefire in early April, a 21-hour direct round in Islamabad that collapsed without an agreement, an indefinite extension declared hours before the first ceasefire lapsed, a cancelled negotiators’ trip, a near-miss in May, and the 60-day ceasefire that carried the June memorandum, with separate Lebanon ceasefires running on their own clock alongside. CNN counted at least 38 times between March 23 and June 9 that Trump called a deal imminent, close to one every two days. The victory claims kept the same tempo. Trump declared Iran militarily defeated on April 6 after saying so on March 17, March 24, and March 26, each time contradicted within hours by the next strike or shipping disruption. The agreement was announced before it was settled, signed from separate capitals, and its central term, the open strait, broke within days of the signing.

Set that against the benchmark for ending a war at the table. The Korean armistice took 158 meetings over two years and 17 days, the longest negotiated armistice on record, close to 200 sessions and 400 hours of argument in a fixed building at Panmunjom, between professional military delegations, with the agenda alone consuming the first seventeen days. A single negotiation in a single room, settled over two years. This war ran the other way, dozens of public declarations and reversals packed into weeks, with no agenda, no fixed room, and no delegation seated across from anyone.

The table did more than house the talks. It set their pace. A fixed agenda and a standing delegation are friction, and friction slows a process and orders its steps. Take them out, and the tempo passes to whoever holds the deadline, reset here every few days on social media and carried by middle-power shuttle. What stands in for the forum runs faster and breaks more often, and the breaks carry a price, paid in the oil spikes and fuel shortages that followed each reversal in the strait.

The choppiness is real, and the pull is to read it as the end of an order. The smaller reading carries more weight. The deadlines that set the tempo were the President’s, improvised in real time and confirmed by his own officials as not part of any national security plan, timed to an economy and an election he needed steady, and broadcast to a home audience, with stocks and oil moving on the announcement. The forum that once stood between that kind of impulse and the use of force has thinned. Reading the result only at the level of the international system, as a leaderless world or a dying order, skips the step where the volatility is generated, which is closer to the ground, in domestic political time. What the war shows is narrower than a collapsing order and more specific. The process that used to absorb and slow these shocks has gone informal, and they pass through now with less to stop them.

The Verdicts That Overreach

The confident verdicts on this war each read a single number, up or down, and the record refuses them.

The China hawks, Zineb Riboua and others at Hudson, see a win over Beijing, an Iranian partner pulled from China’s column. That needs the war to have been about China, and the administration’s own timeline puts Israel at the controls. It needs the consequences for China to point one way, and they cut both ways: by one count China lost a partner and watched its mediation sidelined, by another it gained from an America that spent itself for little.

Ian Bremmer and Firas Maksad argue the reverse, an American defeat in a leaderless world tilting toward China. The defeat half mistakes the axis, since the United States won the fighting it sought. As for China rising, the states that brokered the truce were middle powers acting in their own interests. And the pact Bremmer leans on hardest as proof of a war-driven realignment, the Saudi-Pakistan defense agreement, predates the war by five months. What he calls the war’s consequence stood in place before it. The other half, the speed he reads as a world coming apart, has a nearer source in a formal process gone informal.

The wash verdict, from Barack Obama and, more carefully, Hassan Mneimneh at Carnegie, holds that the war cost much and changed little, exposing the architecture of American power for marginal gain. This reading comes closest to the truth, and it differs only in method. Mneimneh renders his verdict and closes the book. The count stays open here, because what the war exposed is still in motion.

The restraint camp, Trita Parsi and John Mearsheimer, calls it proof of the limits of American power and puts the limit in the wrong place. The United States could destroy whatever it chose. It could not settle. The counterproliferation hawks at FDD call it a necessary success, unproven so far, since Iran’s enriched stockpile survived and the IAEA has been blind inside the country since the first strike.

The single number fails for a structural reason. The capacities it averages move in opposite directions, and the belligerents cannot agree on the sign. Each reads the same memorandum as its own victory, Washington as strength, Tehran as the prudent flexibility its doctrine sanctions, Israel as a job done. No verdict that picks one reading survives the other three.

In Their Own Words

Across the table, the loudest claims of strength rested atop the largest concessions. War Secretary Pete Hegseth said the United States reached the agreement from a position of strength and warned it could clamp an ironclad blockade back on Iran if Tehran broke faith. Vice President J.D. Vance defended the deal by insisting the United States was not giving Iran a cent, even as the text released Iran’s frozen funds and promised $300 billion in reconstruction. Trump measured the outcome by the economy. He cast the deal as what spared the country an economic catastrophe and held up a rising stock market and cheaper fuel as the proof. The victory he claimed rested on American prosperity, and it left the terms he had signed out of the telling.

Tehran’s words ran the other way, defiant about a deal it had decided to take. The new supreme leader, Mojtaba Khamenei, said he had held a different view but approved the agreement, and that direct talks with Washington would not amount to surrender. Taken at face value, the statement looks like a leader defying a deal he just authorized. Inside Iran’s own tradition, it is coherent and even familiar. Ali Khamenei gave the move its name in 2013, heroic flexibility, when he licensed the nuclear talks that produced the 2015 accord, drawing on the Shia memory of Imam Hasan, who made peace with a stronger rival while insisting he had surrendered nothing. Behind it sits the doctrine of maslahat, the expediency that lets the survival of the Islamic system justify a bitter compromise. The precedent runs back to Ayatollah Khomeini accepting the 1988 ceasefire with Iraq, which he compared to drinking from a cup of poison. Mojtaba Khamenei approving a deal he disliked while denying any surrender is not improvisation. It is Iran ending a war the way its doctrine has taught, by taking the prudent truce and calling it by its right name. Tehran signed electronically and kept its leaders out of any photograph that might appear to be an acceptance.

At home, Tehran lets the concessions make the case. It presents the lifted sanctions and the unfrozen funds as what steadfastness extracted from a stronger enemy, and its lead negotiator called the memorandum a record of American failure.

Netanyahu told Israelis the war had crushed Iran’s nuclear facilities and achieved its aims, while admitting he had not seen the memorandum’s full terms and had been kept outside the talks that produced it. The leader whose timing started the war had no hand in how it ended and learned the terms secondhand. In response to those claims, Pakistan’s prime minister Shehbaz Sharif said only that the peace deal had been reached. His claim matched what he had done.

The Realignment Came First

The realignment offered as a consequence of the war was on the table before the war began. The clearest marker is the Saudi-Pakistan defense pact, signed in Riyadh on September 17, 2025, which treats an attack on either country as an attack on both and tucks Saudi Arabia under Pakistan’s nuclear umbrella. It was signed eight days after Israel’s strike on Doha and five months before the United States struck Iran, and analysts took it as a message from Riyadh to Washington that the American security guarantee no longer looked credible. The very pact offered as proof that the war reordered the region was signed before it.

The deeper marker runs back further, to March 2023, when China brokered the restoration of relations between Saudi Arabia and Iran and began a regional habit of routing disputes through capitals other than Washington. By the time this war needed an off-ramp, the brokers were in place and practiced. Pakistan, Qatar, and Oman did not improvise the settlement. They stepped into roles the region had been building for years.

The cohesion questions hanging over this settlement have an older name. The Arab historian Ibn Khaldun called it asabiyya, the group solidarity that lets a power rise and whose decay lets it fall. Can Iran’s revolutionary state hold together under a harder successor installed by a decapitation strike? The same doubt hangs over the battered Axis of Resistance, and over whether the Saudi-Pakistan alignment is bound by anything sturdier than a shared bad month.

The skeptics are strongest on that last question. The Middle East Institute notes that Pakistan’s army is built to deter India and is unlikely to fight Iran or Israel on Riyadh’s behalf, so the pact may signal more than it binds. The claim here stays narrower than a finished bloc replacing the American order. What the region had assembled by then was the habit of brokering its own disputes, and that habit was already standing.

What to Watch

Declining to crown a winner is only worth something if a measurement takes the place of the verdict. If the capacity to end a war on your terms is what is moving, it can be tracked by watching who ends them, war by war, in signals anyone can pull from the public record. The measurement takes the shape of an index, six indicators, each a plain question with an answer you can check:

1. Convening. Who provides the neutral ground and the informal channel that gets the parties to a table at all? In this war, the talks ran through Muscat, Geneva, and Islamabad, settings opened by states both sides would meet on, and Pakistan and Qatar proposed the Swiss resort for the close. Washington hosted no round.

2. Brokerage. Who carries the messages and is trusted by both sides to mediate? The named mediators were Pakistan, Qatar, and Oman.

3. Guarantee. Whose security promise do the parties seek to make the terms hold? Saudi Arabia asked Washington, Washington refused, and Riyadh turned to Islamabad. The American guarantee behind this deal is thin.

4. Terms. Do the terms favor the militarily dominant party or the weaker one? The dominant party conceded, across reconstruction, sanctions, funds, and withdrawal.

5. Hedging. Are the dominant power’s allies buying their security elsewhere? The Saudi-Pakistan pact says yes.

6. Cost. Does the dominant power make the side payments that close the deal, or collect them? Washington pays, underwriting reconstruction financing past $300 billion and the relief that comes with it.

Code those six for this war and the same answer comes back. Settlement occurred outside Washington, and on the one count in which Washington acted, it acted as the party that pays. Another analyst, working only from the public record, codes it the same way. Run the same six against the 1991 Gulf War, and the answers flip. That war ran through the Council, which authorized “all necessary means” in Resolution 678 before the first shot was fired, and Washington convened the coalition, dictated the ceasefire at Safwan, guaranteed the Gulf that followed, and was paid for the war instead of paying for it. This one ran around the Council. It opened without an authorizing resolution, and the body’s one binding action condemned Iran’s retaliation while omitting the strikes that began the war. The forum that gated escalation in 1991 was bypassed at the start and deadlocked at the end, and the convening it once did moved to Muscat, Geneva, and Islamabad. The same six questions distinguish the two wars.

The sixty-day window the memorandum opens is the index’s first live run, and it began to move within days. Iran refused to appear at the signing ceremony set for June 19 in Switzerland, citing Israel’s strikes in Lebanon, and a senior Pakistani mediator flew to Tehran to bring the delegation back. The talks reconvened at the Bürgenstock resort on June 21, not as a meeting between Washington and Tehran but as a four-party session with Pakistan and Qatar at the table; even then, Iran refused to enter the room while the cameras stayed, denying Washington the photograph. Convening and brokerage held in the live run as they had in the war.

The central term gave way next. Three days after agreeing to reopen the Strait of Hormuz, Iran declared it closed again over the same Lebanon strikes, and Washington denied that Iran controls the strait and said the oil kept moving. The dispute is the point. The deal’s signature deliverable now turns on whether Washington can make its own ally stop firing, and through the first week it could not. The clause it turns on is the memorandum’s first, the end of fighting on all fronts that names Lebanon, and the army breaking it is the one party that never signed. Israel’s defense minister said the army would not withdraw from Lebanon, where its strikes had pushed the death toll past four thousand, and the ally that set this war’s beginning was now shaping its end. The other lines wait behind it: the frozen funds Iran ties to the clock, the enriched uranium whose fate is unsettled, Netanyahu’s election due by late October, and a supreme leader four months into the job holding a truce he says he never wanted. The index does not predict which line breaks first. It reads the breaks as they land, and the strait broke inside three days.

As of June 23, the live run still reads the same way. The first round of technical talks closed at Bürgenstock with a road map toward a final deal inside sixty days, and the Lebanon clause became its own proving ground, with Iran, the United States, and Lebanon standing up a deconfliction cell to hold the all-fronts ceasefire, which Iran’s foreign minister called the first real test. Trump called the strait totally open while Iran’s lead negotiator flew to Muscat to work with Oman and held that Tehran runs it. The next round moves to Washington on June 23, the first the United States will host, after every round that built the deal ran somewhere else. By the same day, Israeli and Lebanese delegations were meeting in Washington over a partial Israeli withdrawal from southern Lebanon, the clause the deal turns on now under direct negotiation.

The Function That Moved

The United States could start this war on a morning of its choosing and could not end it without help from Pakistan, Qatar, and a memorandum that paid Iran to stop. That fact is larger than any tally of who came out ahead. The capacity to begin a war and the capacity to end one have come apart, and Washington holds the first.

The win, the defeat, and the wash all miss what moved. What the war left behind is a redistribution still in motion, the diffusion of the capacity to end wars toward the middle and regional powers that now hold it.

That is not an American defeat in the way a lost war is a defeat. It is the kind of change Hedley Bull had in mind when he described great powers as the managers of international order, holding functions that can be lost one at a time. The United States has lost one of them here, in plain sight, and has not said so out loud. It could open this war in a single night and could not, at the end, bring the other side into one room to close it. The power to end wars has moved, and the next war will show how far.

Christopher Sweat is Co-Founder and CEO of GrayStak. GrayStak is a media and political risk company. Our real-time Domestic Political Volatility Index (DPVI) measures the velocity and volatility of U.S. domestic politics. We cover political and institutional risk across the United States, centered in Washington, D.C., New York, and Chicago.

GrayStak holds no position.

The rule sits in subsection (b) of Section 622 of S. 4615. Subsection (a), which governs Israel, contains no such test.

Israel’s technology relationship with China has been a documented source of friction for a quarter century. US pressure forced Israel to cancel the Phalcon early-warning aircraft sale to Beijing in 2000 at a cost of roughly $350 million in compensation, forced the return of Chinese-owned Harpy drones without upgrades in 2005, and has repeatedly objected to Chinese operation of Haifa port, where a Shanghai state-owned firm holds a 25-year terminal contract. Under the bill as written, a UAE data arrangement with a Chinese firm triggers a mandatory cutoff; equivalent conduct by Israel would trigger nothing. Whether any current Israeli arrangement would meet the test is a question the bill never requires anyone to ask. The screen applies to every recipient in the regional architecture except the one at its center. GrayStak could locate no prior reporting on the asymmetry.

It is one of several things the text of S. 4615 reveals when read against itself, and the others run in the same direction. The bill, reported out of the Senate Intelligence Committee by a 14-3 vote on May 20 under Chairman Tom Cotton, would write the US-Israel intelligence relationship into the National Security Act of 1947, with a statutory mandate to expand it and a presidential finding required to shrink it. GrayStak identified no prior statute in any country that mandates an intelligence flow or restricts a president’s authority to adjust it.

The committee approved the provision weeks after NBC News reported that the Defense Intelligence Agency had raised the counterintelligence threat level for Israel to its highest rating, according to two US officials and one former official, an assessment the White House and the Israeli embassy both deny. The fight now heads toward the NDAA conference. Last year’s intelligence bill became law as a division of the Defense Act, and this year’s is positioned to follow the same route.

What Section 622 Does

The operative language is a mandate: the President, acting through the DNI and, as necessary, the Secretary of Defense, “shall, subject to applicable law and the protection of intelligence sources and methods, expand and enhance intelligence sharing with the Government of Israel.” The required scope covers cybersecurity, terrorism, sanctions evasion, the plans and intentions of state and nonstate actors, technology proliferation, missiles, drones, and air and space domain awareness. Paul Pillar, who served 28 years at the CIA and retired as National Intelligence Officer for the Near East and South Asia, observed in Responsible Statecraft that the list covers almost every topic of intelligence interest in the Middle East.

The reduction clause changes who holds the valve. Sharing “shall not be suspended, reduced, or otherwise materially limited except on the basis of a specific and identifiable national security concern determined by the President.” The decision belongs to the president personally. It must be documented in writing. Within 15 days of any material change, including increases, the president must report to the intelligence committees with the categories of information affected and an assessment of regional impact.

Precision matters here. Reductions remain lawful, and counterintelligence risk is explicitly named in the text as a valid basis, which means the drafters anticipated the obvious objection. The change is the cost. A decision that today happens quietly inside the executive branch becomes a documented presidential finding delivered to Congress on a statutory clock. Pillar identifies the mechanism: the required report moves the decision onto Capitol Hill, where any reduction can be framed as a vote against Israel’s security. The escape clause exists on paper. The politics of using it are the deterrent.

A companion provision sharpens the clock, and GrayStak could locate no prior reporting on it. Section 621 requires the DNI to notify the intelligence committees within 48 hours of any decision to pause, restrict, or materially downgrade intelligence support to a “country of significant concern,” defined as Israel, Ukraine, Taiwan, and any country the president designates. For Israel, the provisions stack: a 48-hour tripwire, a 15-day justification, and a personal presidential finding. The 48-hour rule converts a quiet executive adjustment into an immediate congressional event.

One drafting note speaks to the speed of assembly. S. 4615 creates three different new sections numbered 1115 of the National Security Act, in Sections 603, 613, and 622. A committee-reported bill that conflicts with itself twice over section numbering is a cleanup job that conference staff will inherit.

The Off-Ramp Written for Ukraine

The strongest evidence about Section 622 sits 14 pages earlier in the same bill.

Section 620 governs intelligence support to Ukraine and uses the identical phrase: no suspension absent a “specific and identifiable national security concern.” Then it defines the term, a step the Israel section skips. For Ukraine, qualifying concerns include compromise by Russia or another adversary, protection of sources and methods, a voluntary Ukrainian request, and credible intelligence that a Ukrainian government element receiving US support “engaged in a pattern of human rights violations, atrocities, or violations of the law of armed conflict.”

The decision-maker differs too. Section 620 assigns the call to the DNI, acting with the directors of the CIA and DIA. Section 622 hands it to the president.

The same committee, in the same markup, wrote a humanitarian off-ramp into one ally’s section and left it out of the other’s. Documented atrocities are a named, statutorily valid reason to pause intelligence to Ukraine. The Israel section names security grounds only and leaves its deciding term undefined. The bill does not state a reason for the difference.

No Statute Like It

Congress has legislated about the US-Israel security relationship many times. The qualitative military edge statute, Section 36(h) of the Arms Export Control Act, added in 2008, mandates a standard for arms sales. The 2012 and 2014 partnership acts, P.L. 112-150 and P.L. 113-296, declared policy. None ordered an intelligence flow or restricted a president’s authority to adjust one.

The telling comparison is the relationship the United States never legislated. The UKUSA agreement underlying Five Eyes has run since 1946 as an executive arrangement, and no member of America’s deepest intelligence alliance holds a statutory entitlement to US intelligence. Israel is a third-party liaison partner. An NSA memorandum disclosed by the Guardian in 2013 showed the United States already shares raw, unminimized signals intelligence with Israel, under an arrangement reached in 2009 that places no legally binding limits on Israeli use. The pipe runs deep. The executive has always owned the valve.

The value of that ownership was demonstrated in March 2025, when the administration paused intelligence sharing with Ukraine. CIA Director John Ratcliffe confirmed the pause on Fox Business, tying it to the president’s doubts about Kyiv’s commitment to peace talks. The Washington Post reported that the cutoff squeezed the targeting and warning data Ukraine had used against Russian forces. Within days, Zelensky publicly signaled readiness to negotiate, and the pause lifted on March 11 after Ukraine accepted a 30-day ceasefire proposal at talks in Jeddah; Tulsi Gabbard announced the resumption “per the President’s direction.” A six-day pause preceded Ukraine’s acceptance of the ceasefire proposal. The episode showed how fast an intelligence cutoff registers on a partner’s decisions.

Supporters of Section 622 read it as the reason statutory protection is needed, and that case deserves a fair statement: after watching sharing used as a bargaining chip, Congress is moving to ensure continuity for partners it considers strategic, for Ukraine and Israel alike. Critics draw the opposite lesson from the same episode. Removing the tool for a given country gives up a low-cost form of leverage over that country’s conduct. Section 622 would do so for the one partner the Pentagon’s intelligence arm reportedly rates as its highest counterintelligence risk.

The Partner Rated Critical

The record is long, and the consequences for the relationship have run in one direction. Jonathan Pollard passed Israel a vast volume of classified material in the mid-1980s, including, per Seymour Hersh’s reporting in the New Yorker, the RASIN, the ten-volume manual describing how the NSA collects signals worldwide. He drew a life sentence, the harshest ever given for spying on behalf of an ally. Israel granted him citizenship and Netanyahu met him on the tarmac when he emigrated in 2020. In July 2025, per NBC News, US Ambassador Mike Huckabee hosted Pollard at the embassy in Jerusalem in a meeting kept off his official schedule; when the New York Times revealed it in November, the White House said it had been unaware.

Ben-Ami Kadish passed nuclear and missile documents to the same Israeli handler and received a $50,000 fine and no prison time; the sentencing judge asked why it took the government 23 years to bring charges. In 2019, Politico reported that the FBI had attributed surveillance devices found near the White House to Israel, citing three former senior US officials; Netanyahu’s office called the report a blatant lie, and no public consequence followed. In 2021, Pegasus spyware built by Israel’s NSO Group, whose export licenses the Israeli government issues, was found on US diplomats’ phones, per Reuters, with the attacker unidentified. And a 2013 CIA-led National Intelligence Estimate on cyber threats, described in an NSA document from the Snowden disclosures and reported by Newsweek, ranked Israel the third most aggressive intelligence service operating against the United States, behind only China and Russia.

Into that record, the current assessment arrived. NBC News reported that DIA raised the counterintelligence threat level for Israel to its highest rating in recent weeks, citing two US officials and one former official, after concerns that Israeli espionage had grown more aggressive. The Israeli embassy denies the reporting; the White House also rejects it. GrayStak cannot verify the classified assessment and presents the reporting alongside the denials.

The committee voted to mandate expanded sharing in the same season in which the assessment reportedly landed. And in a section drawing no attention, the same bill eliminates the statute of limitations for espionage under 18 U.S.C. 794. Congress is hardening the espionage laws and mandating the expansion of sharing in the same 192 pages.

What a Former Senior Intelligence Officer Sees in the Bill

Paul Pillar spent 28 years in the U.S. intelligence community and retired as the National Intelligence Officer for the Near East and South Asia, the most senior analytic position covering the region Section 622 governs. He has read the provision closely. GrayStak put two questions to him: whether the notification machinery reflects how intelligence sharing actually works, and what it means that the adversary-ties test reaches every Abraham Accords partner except Israel. His answers, provided on the record, run below at length, because his experience managing exactly these relationships is the closest thing this debate has to an expert witness.

On the reporting requirement, Pillar questioned whether it reflects the practice it claims to govern. “This kind of formal reporting requirement does not seem to reflect how such intelligence exchanges work in practice,” he said, “especially given that the bill says nothing about what Israel would be required to share with the United States. These exchanges are two-way, not one-way.” He sketched a situation the statute does not appear to anticipate. A US officer meets an Israeli counterpart under instructions to be as forthcoming as the Israelis are; if the Israelis share little, the American shares little in return. “Would this constitute a ‘decision’ to curtail information sharing, starting a 48-hour clock for reporting to Congress?” he asked. “It would be more accurate to look on it as part of the normal give-and-take of intelligence liaison.” In his reading, the 48-hour notification in Section 621 leaves the 15-day requirement in Section 622 largely redundant, and a two-day window would be harder to satisfy than a fifteen-day one.

On the adversary-ties test, Pillar was direct about what the exemption protects. “Israel has done enough military-relevant technical cooperation with China that if the adversary-ties test were applied to it, then it should be disqualified from intelligence sharing,” he said. “This is very likely the specific reason the bill was written so that this test would be imposed only on the Arab states that have diplomatic relations with Israel and not on Israel itself.” He named the risk the structure leaves in place. Because the test would not reach Israel, he said, enacting the bill as written raises “the problem of Israel possibly passing U.S. secrets on to other countries, including China.”

That assessment belongs to Pillar. GrayStak takes no position on whether any current Israeli arrangement would meet the statute’s test, and the White House and the Israeli embassy have denied the underlying counterintelligence reporting. What Pillar supplies is the judgment of someone who ran the analytic side of these relationships for years: that the bill’s machinery misreads how liaison works, and that its central screen was built to spare the one partner whose conduct would most plausibly trip it.

The Silence at the White House

AIPAC publicly thanked House Armed Services members after they advanced Section 224 of the NDAA, a permanent US-Israel defense technology initiative that The Intercept traced to a lobby-backed bill that died earlier this year. On Section 622 itself, GrayStak could locate no public AIPAC statement, and AIPAC did not respond to a request for comment. The committee was nearly as quiet. Its release announcing passage compressed the country provisions into a single line, a pledge to ensure “intelligence support to U.S. allies, including Ukraine, Israel, and Taiwan,” and Chairman Cotton’s statement highlighted China measures without mentioning the Israel sections. The intelligence provision is moving with less visible sponsorship than its defense-technology twin.

The strategic backdrop makes the push notable: Netanyahu told The Economist in January that he wants to taper US military aid to zero within a decade, a pledge he repeated on 60 Minutes in May, even as Congress moves to bind the two systems together by statute.

The more consequential silence is the administration’s. GrayStak could locate no Statement of Administration Policy on S. 4615, and no veto signal, and neither the White House nor ODNI responded to questions about the provision. What exists is a president actively using the leverage the statute would constrain. CNN reported that Trump’s June 1 call with Netanyahu turned heated as he pressed Israel to scale back in Lebanon, and that Trump said he deterred a major raid on Beirut. His own framing, to the Financial Times: “I call the shots.” The DNI who would execute the mandate, Tulsi Gabbard, has been sidelined from Israel-Iran deliberations, per NBC News, and publicly contradicted by the president on Iran’s nuclear program.

The political risk picture: a Republican committee chair has advanced a statute removing discretion from a president of his own party; the president is exercising exactly that discretion this month as his primary pressure on Netanyahu; and the White House has not engaged with the provision at all. Whether the administration fights Section 622 in conference, trades it, or fails to notice it until signature is the largest open variable in the bill.

Decided in Conference

The calendar is set. The Senate Armed Services Committee completed its NDAA markup June 11 at a topline near $1.5 trillion. The House version cleared with Section 224 inside. S. 4615 sits on the Senate calendar, reported without a written report, so the markup roll calls and the second and third no votes are not yet public; Senator Ron Wyden confirmed his, calling the bill a dramatic retreat for congressional oversight and citing its removal of Senate confirmation for the CIA and ODNI general counsels and its omission of longstanding bipartisan whistleblower protections. The committee’s release, for its part, lists whistleblower protections among the bill’s provisions. Last year’s intelligence bill became law as Division F of the NDAA. If FY2027 follows the same path, Section 622 will live or die in conference late this year.

Watch the written report, the House Intelligence companion, any Statement of Administration Policy, and the text itself: whether 622 emerges intact, amended with the Ukraine section’s definitions imported, or stripped. The 48-hour tripwire, the undefined escape clause, the atrocities off-ramp written for Ukraine alone, and the adversary-ties test that skips Israel have not been part of the public debate, because until now no one had read the text against itself. That is the debate conference will inherit. It should happen in public first.

Christopher Sweat is co-founder and CEO of GrayStak, a media and political risk company covering political and institutional risk across the United States, centered in Washington, D.C., New York, and Chicago. He ran for Congress in 2024 in Colorado’s 5th District, the Colorado Springs seat anchored by NORAD, Fort Carson, and the Space Force’s principal installations, studied national security organization and policymaking at the University of Colorado Boulder, and has covered national security throughout GrayStak’s reporting.

GrayStak covers political and institutional risk. We take no position on the underlying US-Israel relationship. GrayStak is independent: it is not affiliated with any government, political party, campaign, or advocacy organization named in this article. All claims are anchored to the text of S. 4615 as reported May 20, 2026, and to the linked reporting. The DIA assessment, attributed to officials cited by NBC News, is denied by the White House and the Israeli embassy and could not be independently verified. GrayStak requested comment from the Senate Intelligence Committee majority and vice chair, Senator Wyden’s office, the White House, AIPAC, and the Embassy of Israel by close of business Wednesday, June 17. None responded by the deadline. The Office of the Director of National Intelligence does not maintain a public media contact; GrayStak directed its questions on the intelligence provisions to the National Security Council and the committee. Paul Pillar responded and is quoted on the record. This analysis reflects information available as of June 18, 2026; the provision is subject to amendment on the Senate floor and in conference.

That is the whole story, and it is visible in documents the company published itself. A.P. Møller-Mærsk is not insulated from this war. It is positioned in it, long on one side and exposed on the other, and its enterprise risk framework has caught neither position correctly. What follows is based on the company’s own earnings reports, risk disclosures, and shipping records that track its cargo. The contested questions are left to the bodies qualified to rule on them.

The windfall it filed as a threat

Start with the money because it is not in dispute. In 2024, with Houthi attacks forcing carriers to go the long way around the Cape of Good Hope, Maersk reported an operating profit of about 6.5 billion dollars on revenue of 55.5 billion, one of its strongest years on record. Average ocean freight rates rose by more than a third. The company raised its full-year guidance repeatedly as the year went on. Management said plainly where it came from. The Red Sea rerouting soaked up capacity, and rates climbed. The disruption was the windfall.

Now read what the company had written about that same disruption a year earlier. In its 2023 annual report, under the heading "key risks to the 2024 to 2028 b Maersk lists” Maersk lists geopolitical tension as risk number two of ten. It is owned by the Chief Corporate Affairs Officer and flagged as one of only two risks rising year on year. The scenario the company sketched for it reads like a forecast of what came next: long-lasting naval blockades cutting major routes, and broad sanctions choking trade in key markets. The register treats that scenario as pure danger. It contains no recognition that for an ocean carrier with capacity discipline, a blockade that snarls the world’s routes can be the most profitable thing that happens to it. Maersk war-gamed its best year and filed it under threats.

The proof is in the year that followed. As Red Sea traffic began to normalize and rates fell back toward historical levels, Maersk’s 2025 results showed operating profit down by nearly half to 3.5 billion dollars and net profit down by more than half to 2.9 billion, the lowest in five years. The company announced cuts to around a thousand corporate jobs. The 2024 windfall had been a function of the disruption, and it receded as the disruption did.

That is one side of the book, the side the company at least named, even after misjudging the sign. The other side never made the register at all.

The risk with no owner

Since October 2023, the same war has run a second current through Maersk, and this one points the other way. Researchers working from bills of lading, the contracts that document sea freight, have traced a steady flow of munition components moving on Maersk vessels to Israel. The most recent account, published in June 2026 by the Palestinian Youth Movement and Oxfam Denmark and building on earlier work by Danwatch and the Mask Off Maersk campaign, documents more than 2.7 million kilograms of munition components and ordnance shipped to subsidiaries of the Israeli arms maker Elbit Systems since October 2023. The breakdown is specific: over 1.42 million kilograms of bullet cores and brass case cups across 45 shipments from US suppliers, more than 1.1 million kilograms of MK-84 and MPR bomb bodies, and roughly 40,000 kilograms of 120mm mortar components, moving from US ports to Haifa and consigned to Elbit’s IMI Systems and Elbit Land Systems.

What that cargo means in legal terms is not for a shipping analyst to decide, and this piece does not attempt to do so. The relevant facts are narrow. In September 2025, the UN Independent International Commission of Inquiry concluded that Israel had committed genocide in Gaza, a finding Israel rejects. In her June 2025 report to the Human Rights Council, the UN Special Rapporteur Francesca Albanese named A.P. Møller-Mærsk directly, describing it as “sustaining a steady flow of US-supplied military equipment” after October 2023. Those are the inputs that turn a cargo manifest into a corporate risk. Whether they harden into liability is a question for courts and regulators, not for this page.

Run that exposure through Maersk’s current risk map, the one in its 2025 annual report, and it still has nowhere to sit. It is not the geopolitical-tension risk, which the company defines as disruption to the network it runs, but liability from the cargo it carries. The closest fit is the compliance risk, also owned by the same office, but the company scopes that to anti-corruption law, anti-trust, sanctions, and data privacy. Complicity, the UN Guiding Principles on Business and Human Rights, the law of armed conflict: none of it is named. Nor does it surface in the company’s 2025 double-materiality assessment, which was updated to cover labour rights, working conditions, and a dozen other topics, but not the human-rights exposure the cargo creates. Two years of reporting, a genocide finding, and two rejected shareholder resolutions later, the exposure the campaigns are trying to force onto the books still has neither a category nor an owner.

Which raises the obvious question. Two years of reports, port blockades, and two shareholder revolts, and none of it has shown up in the share price.

Why the position holds

The denial does most of the work, and that’s sturdier than it looks. Maersk’s standing line is that it maintains a strict policy of not shipping weapons or ammunition to Israel. It reiterated to the authors of the June 2026 report on 2 June that its cargo screening is grounded in the Wassenaar Arrangement, the EU common military list, and US arms-export rules. The company can say that because the military cargo does not move on the Danish parent. It moves on to Maersk Line Limited, the group’s American subsidiary, which is incorporated in the United States, is governed by a board of US citizens, and operates under a special security agreement with the Defense Department. It operates a fleet of US-flagged ships in the Maritime Security Program, the Pentagon’s sealift reserve. The government’s own fleet list names 18 vessels under Maersk Line, Limited, paid to remain available to move military cargo, and the company has been in the program since 1996. The bills of lading behind the Elbit findings name several of those same hulls, the Maersk Chicago and the Maersk Seletar among them, on the bomb-body route to Haifa. The weapons business and the weapons denial live in two different corporate bodies, separated by a firewall the company built on purpose. The parent can deny shipping weapons because, as a parent, it is telling the truth.

Ownership closes the other exit. The Møller family holding company holds about 41.5 percent of the shares and 54.5 percent of the votes, through a dual-class structure. A shareholder resolution to halt the cargo cannot pass against that bloc. The vote is foreclosed before it is cast.

Geography sorts the rest. The exposure is sharpest in two places. In Denmark it is reputational and political, in front of the company’s home press and its home government. In Spain it is physical: Madrid has barred ships carrying arms for Israel from its ports since 2024, and in November of that year it turned away the Maersk Denver, another hull from the same Maersk Line, Limited roster. Maersk denied the ship carried weapons or ammunition, then questioned whether Spain was now rejecting anything military-related bound to or from Israel, legal cargo included, an answer that disputed the label while conceding the cargo. US regulators have since opened an inquiry into the ban and floated fines of up to $2.3 million per voyage. The exposure is weakest in the one country where most of the military cargo actually moves. In the United States, the customer is the government. The pressure and the activity sit in different jurisdictions, and that mismatch is why the risk has stayed off the income statement.

Durable is not the same as permanent.

The machine built to move it

It runs on law, not on noise. Look again at the shareholder votes. At the 2025 annual meeting, a single shareholder working with the campaign group Ekō and the Palestinian Youth Movement, backed by more than 70 organizations including Amnesty International and Oxfam Denmark, asked the board to disclose its human-rights due diligence and to stop the transfers. It was voted down, as the ownership math guaranteed. A second resolution was returned to the 2026 meeting and rejected again, this time citing the International Court of Justice, the International Criminal Court, and EU due diligence law by name.

Read as vote tallies, the resolutions look like defeats. Read as a procedure, they did other work: they put the board on the record, formally and unavoidably, on notice of the allegation and of the demand to show its due diligence. Under the standards that govern corporate complicity, liability turns on what a company knew and whether its diligence was adequate. A report addressed to the public can be waved off; a resolution the directors had to receive and rule on enters the corporate record and stays there. That, not the vote count, is what the resolutions accomplished.

Maersk understood the move and answered it in kind. CEO Vincent Clerc told shareholders that the company is fully compliant with the UN principles and the OECD guidelines and is not shipping any weapons or ammunition into conflict zones. That answer is the company’s own counter-record, built to rebut the claim that it was willfully blind. Note who owns the risk the answer defends. The Chief Corporate Affairs Officer is also Maersk’s General Counsel. The line between weapons and military cargo is a legal construction, and the company’s top lawyer owns it.

The conversion, if it comes, runs through the legal perimeter, and that perimeter has been moving in both directions. The EU’s mandatory due-diligence law, once the campaign’s strongest lever, was narrowed and delayed in February 2026: its scope was cut back toward direct business relationships, its liability provisions were softened, and application was pushed out to 2029. The OECD guidelines remain a forum, though no formal complaint against Maersk has been filed. The harder edges are the genocide finding, the provisional measures at the International Court of Justice, the arrest warrants for Netanyahu and Gallant at the International Criminal Court, upheld on appeal in December 2025, and the Spanish port ban, which is the one channel that already physically stops cargo. The company has also shown it will give ground on the flank to relieve pressure: in June 2025 it divested from firms tied to Israeli settlements after a review of its West Bank traffic. The line it will not cross is the cargo. The line it will move is everything around it.

Which leaves the question that matters. Where is the company on this curve, and what moves it?

Two clocks

They keep different time. The first is the earnings clock, which is already visible in the results. Maersk’s recent profits were a function of the Red Sea staying shut. As traffic returns and capacity floods back, rates fall and earnings compress, which is what the halved 2025 profit already shows and what the soft 2026 guidance assumes. On this clock, the peace everyone wants is what erodes the earnings. A reopened Red Sea unwinds the windfall.

The second is the franchise clock, and it runs slowly. The complicity exposure is real and it is dormant. In any given quarter, the odds that it forces a change are low, held down by the firewall, the family votes, the absence of binding sanctions on Israel, and a due-diligence law that just got weaker. Over a longer horizon, two to four years, those odds rise if the war continues and the legal perimeter tightens. A full halt driven by pressure alone is unlikely soon. A further concession on the flank is more likely than not. The structural break, if it comes, will come from outside the company, from a court or a regulator or a law, and the activist pressure bears on the outcome only insofar as it is positioned for that moment.

Run the two clocks together, and you misread the exposure. Activism works on a slow clock, and treating it as a near-term trigger is a common error. The near-term variable is the Red Sea. The longer one is the legal perimeter. The crowd outside headquarters moves neither on its own.

What the documents show, in the end, is a company whose exposure to a single war is larger than its own risk register admits, on both sides. It booked the upside as a surprise and filed it under danger. It carries the downside with no name for it at all. Maersk has built a structure that lets it maintain that position and a denial that allows it to describe the position as neutral. Neutrality, though, is a claim the war is busy repricing, and the company’s filings are the clearest evidence that it has not yet caught up to its own exposure.

Notes on sourcing and method

This piece is built from primary documents: A.P. Møller-Mærsk’s published annual reports and earnings releases, its enterprise risk disclosures, the US Maritime Administration’s Maritime Security Program fleet list, the bills-of-lading research compiled by the Palestinian Youth Movement, Oxfam Denmark, and Danwatch, the records of the company’s 2025 and 2026 annual general meetings, the published orders and releases of the International Court of Justice and International Criminal Court, and contemporaneous reporting on each. Financial figures are Maersk’s own as reported for the relevant periods. GrayStak did not independently verify the underlying shipping records; the cargo findings are those of the cited investigations and are attributed to them throughout.

Contested legal characterizations are attributed, not asserted. The determination that genocide has occurred in Gaza is the conclusion of the UN Independent International Commission of Inquiry, which Israel rejects; the references to plausibility and provisional measures are the International Court of Justice’s; the arrest warrants are the International Criminal Court’s; the complicity and due-diligence allegations belong to the named civil-society organizations and the UN Special Rapporteur. None of these are findings of this author or this firm. The piece takes no position on whether Maersk’s conduct constitutes a legal violation, a question that rests with the courts and bodies named.

Maersk’s position is reflected in its public statements, its executives' statements at its annual general meetings, and its June 2026 response to the most recent report. The company maintains that it does not transport weapons or ammunition to Israel and that it operates in compliance with applicable law and the UN and OECD frameworks. Ahead of publication, GrayStak put the central claims in this piece to Maersk’s media relations and invited a response, with a deadline of 18:00 CET on 15 June 2026. The company did not respond by that deadline. GrayStak welcomes correction of any factual error.

This analysis reflects information available as of June 16, 2026. Several elements, including EU due diligence law and Maersk’s financial guidance, are subject to change.

Disclosure

This article is analysis and journalism, not investment advice or a recommendation to buy, sell, or hold any security. It does not account for any reader’s circumstances. The author holds no position in A.P. Møller-Mærsk or any other security discussed and has no financial relationship with the company or the organizations named. GrayStak is independent. It is not affiliated with A.P. Møller-Mærsk, with any government, or with any of the advocacy organizations or campaigns named in this article.

About the author

Christopher Sweat is co-founder and CEO of GrayStak, a media and political risk company. He writes on defense and national security, technology and AI policy, energy infrastructure, and labor. His background spans roughly two decades in enterprise technology, across telecom, data-center infrastructure, enterprise software, and AI infrastructure, and he studied national security organization and policymaking.

This is the third piece in a series about the distance between what an announcement says and what the underlying structure actually holds. Palantir told the world it left Colorado; the filing moved while the workforce and the infrastructure stayed. The Illinois data centers were the next case, each one sold to a town as jobs and tax revenue, while the power bill was spread across 67 million ratepayers, and the people who objected landed in federal surveillance files. SpaceX is a bet on a man, and the offering arrives at the moment his record gives the most reason for pause.

The irony sits in the language. In August 2018, Musk tweeted that he might take Tesla private at $420 per share and added three words: “funding secured.” It was not. The SEC later established, and Musk did not seriously dispute, that he had picked the number as a marijuana joke for his girlfriend. The agency sued him for fraud, and the episode cost him the chairmanship. Eight years on, the funding is real, the largest pile ever raised in a public offering, and the question for anyone reading the prospectus is whether those eight years argue for handing it over or against.

Keep track of two people, because the prospectus shows only one. Elon the manager has run public-company capital for two decades, with a long record of how he treats the shareholders who supply it. Elon the political actor is newer, and he turns a routine governance story into a live risk. The shape of the deal gives the game away. This offering folds several of Musk’s holdings into one company: a dominant launch business, a money-losing AI lab, and a social network he bought for $44 billion and watched lose most of its value. He has merged his companies into one another before, and those attempts drew lawsuits, shareholder revolts, and a judge’s order striking one down. This time the consolidation is wrapped in a charter designed so none of that can recur. GrayStak takes no position in the securities it covers. The question is not whether SPCX rises or falls but what risk the offering’s story leaves out: the man, the structure, and the politics.

The pattern

Begin with the SEC, the cleanest window into the operating style and a settlement that still shapes the company on offer.

Look at what the 2018 tweet actually rested on. The money to take Tesla private was supposed to come from Saudi Arabia’s Public Investment Fund, which had quietly taken a stake in Tesla, and Musk rested the word secured on a single meeting with its managing director, Yasir Al-Rumayyan. The deal collapsed, and four years later the litigation pried loose the text messages between the two men. They do not describe secured funding. When a news report called the fund merely in talks, Musk accused the Saudis of throwing him under the bus, and Al-Rumayyan answered with the line that settles the matter: “It takes two to tango. We haven’t received anything yet.” The backer Musk had cited as his proof was telling him in writing that nothing had been committed and that the fund still lacked the financials to decide. He told the public market the money was secured while the only investor he had named was asking him for the numbers.

The fraud charge settled on September 29, 2018. Musk and Tesla paid $20 million each, Musk surrendered the chairmanship for three years, and Tesla agreed to seat two independent directors and a committee to vet his public statements; the arrangement staff came to call the babysitter for his tweets. His replacement as chair was Robyn Denholm, installed as a reform, a check on the founder the regulator had just sanctioned. Hold that name. Musk’s view of the agency has never been subtle. Within days he was calling it the Shortseller Enrichment Commission, and by December he told an interviewer flatly that he did not respect it. A jury later cleared him in the shareholder suit over the same tweet, which belongs in the record because it cuts his way. What matters is the posture, not the box score on individual charges. A federal regulator caught him misstating a market-moving fact, built oversight around him, and drew open contempt in return.

The same reflex appears wherever his interests and his shareholders’ pull apart. In 2016, Tesla bought SolarCity for about $2.6 billion. Musk was the largest holder of both companies, sat on both boards, ran Tesla, chaired SolarCity, and had cousins running the solar firm as it bled cash. Shareholders sued, calling it a bailout of one Musk company by another. He won: a Delaware judge ruled in his favor in 2022, and the state Supreme Court affirmed, finding the deal entirely fair despite a flawed process. That word flawed carries weight in the opinion. The court called the episode a parable of unnecessary peril, found Musk more involved than a conflicted fiduciary should be, found the directors’ conflicts not fully neutralized, and noted that Tesla never formed a special committee to handle a transaction rife with conflict. The deal survived. The record that saved it describes, in a judge’s hand, exactly the founder-dominated decision-making the SpaceX board will now perform for public shareholders, with Musk again holding all three seats: largest owner, chairman, chief executive.

The board that isn’t

Consider what he is paid, and who signs off. Tesla’s board granted Musk a 2018 package eventually worth around $56 billion. In January 2024, a Delaware court voided it, ruling the board had not been independent when it negotiated his pay. Tesla rejected the finding, reincorporated in Texas, and put a larger plan to a vote. On November 6, 2025, shareholders approved a package worth as much as $1 trillion, with about 75 percent of voting shares in favor, excluding Musk’s own roughly 13 percent. Glass Lewis, ISS, and Norway’s sovereign wealth fund all urged a no. The milestones read like a dare: an $8.5 trillion market value, six times the company’s value at the time, plus 20 million vehicles, 1 million humanoid robots, and 1 million robotaxis in service. The revealing part is the pitch. The board never argued the package on its merits. Denholm, the chair who was seated in 2018 as a check on Musk, warned that the company might lose him if shareholders refused. The watchdog had become the lobbyist.

That board had been described as captured for years before the court said so, by reporters and a senator alike. In February 2024, the Wall Street Journal published “The Money and Drugs That Tie Elon Musk to Some Tesla Directors,” reporting that directors had made hundreds of millions from Tesla stock and from private stakes in Musk’s other ventures, an arrangement the paper called an extreme blurring of friendship and fortune. Several had used illegal drugs with him, and some said refusing risked their standing in his circle. A companion investigation into Musk’s own drug use reported that it had worried executives and directors at his companies. The board left no record of its concerns in the minutes. James Murdoch held a seat. So did Musk’s brother Kimbal. So did Antonio Gracias, a friend and investor across the Musk companies. Independence was a formality, and the same reporting made the awareness known to SpaceX’s directors as well.

The reflex is consistent, which is the point. Senator Elizabeth Warren has written to the board and the SEC since 2023 about Musk’s conflicts and the board’s refusal to police them, including a 2024 episode in which he reportedly redirected thousands of scarce Nvidia chips Tesla had ordered to his private company X. Move the asset that suits you from the public company to the private one: SolarCity in 2016, the chips in 2024, the same play at different sizes. So the honest answer to how SpaceX will be governed as a public company is that we already know it, from fifteen years of evidence, a court record, and a newspaper investigation. The charter Musk is taking public is that habit with the friction removed.

He is being paid that trillion, moreover, against a shrinking core business. Tesla deliveries peaked at 1.81 million in 2023, slipped to 1.79 million in 2024, and fell to 1.64 million in 2025, an 8 percent drop that is speeding up, with BYD now the world’s largest electric-vehicle seller. The package bets on robotaxis and humanoid robots that do not yet exist at scale, waved through by a board that framed the alternative as losing the only man who could build them. That is the manager. The next question is what he has assembled for you to buy.

You’re also buying Twitter

Musk bought Twitter in 2022 for about $44 billion. He agreed to the price, tried to escape it once he decided he had overpaid, lost in court, and closed in October at the number he had called too high. The financing mattered. Roughly thirteen billion dollars of debt was loaded onto the company, and the banks behind it spent years unable to offload it. The rest is documented: Fidelity, an equity holder, marked its stake down by about two-thirds, and by late 2024 the platform was valued at nine or ten billion, three-quarters of the purchase price gone.

Then the pattern did its work. In March 2025, Musk had his AI company, xAI, acquire X at a $33 billion valuation, $45 billion including debt. A law firm Musk hired set the marks on both sides, because no arm’s-length buyer existed. He owned and controlled both companies and negotiated across the table from himself. The maneuver lifted a platform last valued near nine billion back to thirty-three, on the strength of an AI company that was itself losing money, a collapsed asset made whole on paper by folding it into a fashionable one.

The move ran a third time, larger. On January 30, 2026, SpaceX absorbed xAI, which by then included X, at a combined $1.25 trillion: SpaceX $1 trillion and xAI $250 billion. xAI’s market cap had more than doubled in 10 months. The stated rationale holds together on the surface: X supplies the AI with data and an audience, and SpaceX could one day run that AI from orbit on Starlink’s hardware. Trace the chain, though. Twitter became X, X folded into xAI, xAI folded into SpaceX, and SpaceX is the company now selling shares. Buy SPCX, and you are buying the rockets, the Grok chatbot, and the former Twitter as one indivisible lot, not a rocket company with an AI lab on the side.

The financing has a throughline of its own, and it is the one institutional buyers should sit with. The $44 billion for Twitter rode on thirteen billion in debt from seven banks, Morgan Stanley the largest at about $3.5 billion, paper that those banks could not sell and carried at a loss for more than two years, the worst buyout financing for Wall Street since 2008. Several of them now underwrite the exit. Morgan Stanley sits second on the SpaceX cover behind lead-left Goldman Sachs, Bank of America is a lead bookrunner, and Barclays, Mizuho, and Société Générale fill out a syndicate of more than twenty banks. The lenders left holding Twitter’s debt are now selling the public the company that absorbed it, the fee event they were reportedly counting on when they agreed to hold the loans. Goldman ran Tesla’s 2010 IPO. The relationships being monetized here are old ones, and the parties marking the price are not disinterested.

The prospectus finally attached numbers to the AI piece, and they are grim. xAI lost about $6.4 billion in 2025 on roughly $3.2 billion of revenue, the unit now carried at $250 billion inside the deal. The merger turned a profitable company into a loss-maker: SpaceX earned $791 million in 2024, then posted a $4.9 billion net loss in 2025 on $18.7 billion of consolidated revenue, with the AI segment driving the reversal. The filing also slipped a warning into the risk factors that the company may issue a significant amount of equity in future transactions, language read across the market as runway for an eventual Tesla merger. Tesla itself put two billion dollars into xAI in January 2026, after its own shareholders had balked at the idea. The wiring between the man’s companies is being soldered in real time, and under the charter he is about to float, Musk is the only person whose sign-off any of it needs.

Read the risk factors

The public is being sold Class A shares with one vote each, while Musk’s Class B shares carry ten. On roughly 42 percent of the equity, he will hold more than 82 percent of the vote, the filing says. The filing is blunt about the consequences: Musk will be able to control the outcome of matters requiring shareholder approval, including the board, and SpaceX will be listed as a controlled company, exempt from Nasdaq’s independent-board rules. He plans to serve as chief executive, chief technical officer, and chairman at once. Because only Class B shares can remove him and he controls them, he can lose the job only by voting against himself.

The cost of admission is your recourse. The articles require shareholders to waive jury trials, bar class actions, and route disputes into mandatory arbitration, a structure the SEC cleared by party-line vote in September 2025. Texas law, where Musk reincorporated, sets a 3 percent floor to bring a derivative suit, which at this valuation means roughly $50 billion in stock, a bar only Musk himself realistically clears. Every check that has ever constrained him- the minority vote that hung over SolarCity, the class action, the Delaware court that struck down his pay- the charter disables in advance.

The objectors are not short sellers. CalPERS and the New York State and City pension funds, together north of a trillion dollars, wrote to SpaceX before the filing demanding one-share-one-vote or a sunset, an independent board, a split between chair and chief executive, an end to forced arbitration, and independent approval of related-party deals. Denmark’s AkademikerPension will not buy at all, citing a catastrophic governance structure. The American Federation of Teachers, 1.8 million members, asked the SEC for extraordinary scrutiny of the offering. Fortune asked whether SPCX will be the least shareholder-friendly public company in history. These are the long-only institutions that usually anchor the marquee names, and several are staying out on governance alone.

The deal is engineered to match. It is a fixed-price offering, $135 a share with no bookbuilt range, a structure with almost no precedent among large U.S. listings, and the book is already reported oversubscribed. The fixed price signals to institutions that the underwriters see demand outrunning supply before a single order clears. Every share is primary, so the proceeds go to the company and no insider sells, with Musk and other holders locked up for 366 days. For an allocator, that cuts both ways. No insider overhang hits the tape at the open, which is the bull’s comfort. But there is no price discovery from a range, no float relief for a year, and with only about 3 percent of the company floated, a thin, controlled supply in a name priced for perfection. You are taking the issuer’s price on the issuer’s terms, the same posture the charter sets everywhere else.

The number is its own argument. At $1.77 trillion against $18.7 billion in 2025 revenue, SPCX trades near 95 times sales. Starlink is the engine, about $11.4 billion of that revenue at a roughly 63 percent margin and the only segment turning real operating profit, which is the strong part of the story and the reason the multiple is not pure fantasy. Independent valuations still land well below the ask: Morningstar opened coverage at a $780 billion fair value, less than half the offer, judging the company significantly overvalued, and NYU’s Aswath Damodaran, often called the dean of valuation, puts it near $1.3 trillion, still under the price. The company ran a $4.9 billion net loss, and the index treatment cuts both ways. S&P Dow Jones kept its multi-class rules intact, keeping SPCX out of the S&P 500 for at least a year on the same governance grounds the pension funds cited, which denies it the deepest passive bid in the market. The Nasdaq moved the other way: a rule it changed on May 1, 2026 lets any newly public company in the top 40 by market cap enter the Nasdaq-100 after 15 trading days, scrapping the old seasoning and float minimums. SPCX clears that at the open, so passive Nasdaq-100 money would arrive inside about three weeks. FTSE Russell set an even shorter five-day window, and MSCI will fast-track it into its World and All-Country indexes the day after it lists, price-insensitive demand arriving almost at once. Against the thin float that buying bites: one index-rebalancing forecaster estimates that passive funds will hold roughly 30 percent of the freely traded shares within fifteen trading days, compared with about 4 percent under the old rules. The S&P 500 bars SPCX on governance; the Nasdaq, FTSE Russell, and MSCI have rewritten their timelines to include it.

It lists into a jittery tape. Tech has been selling off into the open, a CPI print and fresh Middle East tensions draining risk appetite the same week, and analysts are already framing the listing as the public market’s test of whether investors will underwrite trillion-dollar AI narratives at the prices private rounds set. An unusually thin float meets the loudest doubt about AI valuations in months, on the same morning.

Then there are the retail terms, the part that should trouble a political-risk reader most. SpaceX reserved up to 30 percent of the offering for retail, three to six times the usual slice, and the brokerages bent to match: Fidelity cut its account minimum from $500,000 to $2,000, and Robinhood and SoFi set none. Read the next clause, though. Fidelity blocks a sale within fifteen days on pain of a six-month ban, then a year, then a lifetime, and Robinhood and SoFi run thirty-day versions. So the offering draws in the smallest investors at the widest scale on record, then bolts the door behind them, holding them through the most violent stretch of price discovery a new megacap ever sees. These are the buyers with the least voting power and the least room to absorb a loss, and even their allocation is a lottery if demand outruns supply, as it likely will. The structure routes the largest possible number of ordinary investors into a stock that grants them almost no say and almost no exit.

The other Elon

Everything to here describes a richly priced, founder-run company with a difficult steward, a familiar species of risk. The second Elon is what sets SpaceX apart.

He arrived on July 13, 2024. Minutes after a gunman fired at a rally in Butler, Pennsylvania, Musk endorsed Donald Trump on X, ending years of insisting he would stay out of partisan politics. He built a super PAC, poured close to $277 million into the cycle as the largest single donor, hosted Trump for an interview on his platform, and told a second Butler crowd that a Trump loss would be the last real election they would see.

The reward was a desk inside the government. Musk took charge of the Department of Government Efficiency as a special government employee capped at 130 days, and for the term’s opening months was nearly inseparable from the President: in the Oval Office and Cabinet meetings, on Air Force One, cast by friend and critic alike as a kind of co-president. DOGE cut hard through agencies and staff, and the blowback reached Musk’s own showrooms, where Tesla lots, chargers, and cars drew protests and vandalism and the stock sank under the weight of his politics. The White House even staged a Tesla display on the South Lawn, the President pitching a struggling company’s cars to shield its chief executive. His tenure ended around the close of May 2025, sourly, as he turned on the spending bill he said gutted DOGE.

In the first week of June 2025, it blew apart. Musk called the President’s signature bill a disgusting abomination, and Trump, from the Oval Office, floated stripping Musk’s companies of their federal contracts. Musk threatened to decommission the Dragon capsule NASA depends on for the space station and posted, without evidence, that Trump appeared in the Epstein files. Tesla’s stock dropped hard. The richest man alive and the President spent forty-eight hours trying to ruin each other in public, then reconciled. By the Charlie Kirk memorial that September, they were together again, and in May 2026, weeks before this IPO came to market, Musk joined Trump’s delegation on a state visit to China. The relationship is warm at the moment, which is the problem.

Consider what rides on it. SpaceX’s own filing puts roughly a fifth of 2025 revenue with U.S. federal agencies, where it is almost always the prime contractor. In the weeks before the offering, the Space Force handed it $6.45 billion in fresh awards, $4.16 billion of it for satellites under the $175 billion Golden Dome missile-defense program, which SpaceX is positioned to dominate. NASA’s road back to the Moon runs through a SpaceX lander. The company’s growth and strategic standing are tied to the federal government, controlled by a man whose access to it has been shown to swing on a single personal relationship. For two days in June 2025, the thesis ran live: the President reached for the contracts, the founder reached for a NASA program, and both levers proved real.

A contradiction sharpens the point. Musk spent 2025 as Washington’s self-styled scourge of waste while his companies live on federal money, and SpaceX has reportedly paid almost no federal income tax across its life, sheltering income behind years of losses. The deficit hawk’s flagship is a contractor that draws far more from the Treasury than it returns, which says less about hypocrisy than about how fast the public posture can flip when the contracts are at stake.

The appetite stops neither at the White House nor the water’s edge. In April 2025 Musk put something like $25 million into a single Wisconsin Supreme Court race, headlined a rally handing out million-dollar checks, and watched his candidate lose by ten points in a contest read as a referendum on Musk himself. Abroad, he used X to enter European elections directly. He endorsed the far-right AfD before Germany’s February 2025 vote, a party on the national intelligence service’s extremism watch list, interviewed its leader live, and declared that only the AfD could save the country, drawing rebukes from Berlin, Paris, and London at once. He attacked the British prime minister by name over a child-abuse scandal and demanded a jailed agitator’s release. He condemned Romania’s courts for annulling an election the far right had led. A passive owner does none of this. These are repeated, costly, and often losing interventions in the politics of the very countries where Starlink needs licenses and Tesla needs buyers, the instinct that made Tesla a protest target at home now turned on the regulators and customers his companies depend on abroad.

The bull case

None of this is the consensus, and the people on the other side are not naive. The book is oversubscribed. Cathie Wood’s ARK Invest holds SpaceX as the largest position in its venture fund and models it at $2.5 trillion by 2030, a bull case past three, and has called the $75 billion raise too small. Goldman has floated $474 billion in revenue by 2030 and Morgan Stanley $3.4 trillion by 2040, and Wedbush’s Dan Ives reads even the unusual fixed price as a sign of confidence. The early backers, Founders Fund and Fidelity and Thrive, sit on returns that vindicate twenty years of conviction.

Their case is coherent, and it runs straight through the things this piece treats as risks. The consolidation that reads as self-dealing from one angle is, from theirs, a vertically integrated flywheel: Starlink’s cash, X’s data, xAI’s models, Tesla’s robots, and orbital compute converging under one roof in a way no committee-run company would dare attempt. Starlink is already a real, profitable engine. And the founder control that strips shareholder recourse is, to a believer, the entire point. It is what let Musk build SpaceX and Tesla against a market that shorted both and was wrong both times, and the institutions objecting on governance are, on this read, the same ones that sat out every prior Musk run. Buying SPCX is buying the one operator who keeps doing what the spreadsheets call impossible. The government has made the same bet across administrations, treating him as the indispensable contractor for launch, crew, and, increasingly, defense.

That case deserves to be met at full strength, because it may prove right. The disagreement is not about whether Musk can build. It is about what the structure does to everyone else if the bet goes wrong. The bull case rests on Musk continuing to be right; the rest of this piece is about what the charter leaves you holding if he is not, in a stock you cannot vote, cannot sue, and was sold at a price you had no hand in setting.

It was never the rockets

The engineering is sound, and the bull case is serious. The easy version of this piece, the one that waves the rockets away, has it backward.

The risk is everything bolted to the rockets, one thing wearing four coats. A founder who treats oversight as an adversary, runs deals among the companies he owns, and turns his own indispensability into leverage, now with that habit sealed into a charter whose vote he cannot lose. A portfolio that welds the launch business to a money-losing AI lab and the wreckage of a $44 billion Twitter, priced near ninety-five times sales and marketed hardest to the investors with the least protection. A revenue base a fifth dependent on a federal government whose posture toward Musk has lurched from co-pilot to enemy and back inside a year. And a controlling owner spending his own fortune and platform to antagonize governments from Wisconsin to Berlin, the same ones his companies need. Each has been in plain view for years. The charter’s one innovation is to guarantee that this time none of it can be voted down, sued over, or struck off by a court.

Eight years ago he told the market the funding was secured when it was not, and it cost him the chairmanship and a forty-million-dollar settlement. This week it is secured for real, and larger than any raise in history. The rockets will fly. The open question is whether the same man, with the same instincts, does anything differently once the money is finally, actually, his.

Christopher Sweat is co-founder and CEO of GrayStak, where he builds political-risk and early-warning systems that connect hyper-local signals to national, geopolitical, and market dynamics.

GrayStak is a media and political risk company. Our real-time Domestic Political Volatility Index (DPVI) measures the velocity and volatility of U.S. domestic politics. We cover political and institutional risk across the United States, centered in Washington, D.C., New York, and Chicago.

Disclosures. GrayStak is not a registered investment adviser or broker-dealer, and nothing in this article is investment, legal, or tax advice. This piece is journalism and analysis offered for informational purposes only. It is not a recommendation to buy, sell, or hold any security, and it is not an offer or solicitation of any kind. None of it is tailored to any reader’s financial circumstances. All investing carries risk, including the loss of principal, and newly public companies can be especially volatile; anyone weighing a position should do their own research and consult a licensed professional. Forward-looking statements are opinion, and factual claims are drawn from public sources believed reliable but not independently guaranteed. The author holds no position in SpaceX or in any company named in this article and has no financial stake in the offering. The author reports from within the technology sector this article covers and ran for U.S. Congress in Colorado in 2024; readers should weigh both vantage points. Research for this piece was assisted by an AI model built by Anthropic, which competes directly with xAI, the artificial-intelligence business now folded into the SpaceX entity discussed above.

From the cypher to the council

Vasquez, 46, is the son of Guatemalan immigrants and a former battle rapper who performed under the name Prime. His politics started with hip-hop, long before any party. Listening to NWA, KRS-One, and Public Enemy in the late 1990s, he says, gave language to things he had already lived: poverty, being treated as an outsider, friction with police. “Without thinking about the politics of it, you just know it’s a dope song,” he says. “And next thing you know you’re reciting it, you’re thinking about how it speaks to you and your environment.” When he first heard Bernie Sanders in 2015, the pitch felt familiar. Sanders, he says, “sounded like my rap records.” He knocked doors in Iowa, came back and threw a fundraiser that doubled as a concert, and met an organizer recruiting first-time candidates. In 2019 he unseated Ald. Patrick O’Connor, a 36-year incumbent and a former floor leader for Mayor Rahm Emanuel. He describes his first televised debate in the terms he knew best. It “felt like a battle. Back to back to back bars.”

The only socialist to vote yes

What has defined his six years in office is a willingness to break with the left that produced him. The clearest example is on the record. In late 2020, Vasquez broke with the city’s other socialists to vote for Mayor Lori Lightfoot’s 2021 budget, which carried a $94 million property tax increase and passed narrowly. He was the only Democratic Socialists of America member on the council to support it, and the Chicago DSA chapter censured him, its first such censure. He has not recanted. He says the vote was the price of a deal to stop city worker layoffs during the pandemic, and he frames the episode as the recurring fight on the left between purity and governing. “We all agree on goals. We don’t all agree on how to get there,” he says. A movement built by people who have felt cast out, in his telling, has a habit of casting out its own. “Yo, you’re not left enough, so now you’re not a progressive. And it’s like, we’re not gaining power that way.”

His critics on the left framed it as something else. On the floor of that 2020 debate, Socialist Caucus members rejected the same deal he defended as the responsible vote. “Don’t give me crumbs and tell me it’s cake,” Ald. Jeanette Taylor (20th Ward) said of the budget. “We’re in a global pandemic, and there’s no way in the world that we should be balancing this budget on the backs of taxpayers.” To that wing of the left, Vasquez’s pragmatism reads as capitulation. He calls it the price of keeping people from getting hurt.

Asked directly whether he still considers himself a socialist, he does not hedge.

“Government is socialist. You’ve made a social agreement that you think you can do better by pulling your resources together to solve for more people.”

He adds a qualifier, calling himself a coalitionist, and draws a distinction between two kinds of people who share the label. One comes to it through hardship, and he cites Fred Hampton, Martin Luther King, and Harold Washington. The other, he says, is the person whose “mom and dad paid for college and I read Marx.” People who have lived precarity, he argues, learn to compromise because survival requires it. It is a framing that favors his own position, and he makes no apology for it. He sums up his method in the language of an umpire: he calls “balls and strikes,” he says, no matter “who’s on the fifth floor” at City Hall.

At odds with an allied mayor

That posture has put him at odds with Brandon Johnson, the closest thing the Chicago left has to one of its own in the mayor’s office. Vasquez voted against the mayor’s 2026 budget. He says he supports taxing corporations and still calls the mayor’s signature corporate head tax a marketing ploy. The measure, a $33-per-employee monthly tax on companies with at least 500 workers, would have raised about $82 million. His objection, he says, is precision. A broad tax also falls on immigrant- and Black-owned businesses trying to scale, and the city would do better to target industries that have thrived while others have struggled.

“Am I really doing something, or am I saying that we’re doing the corporate head tax, so that people out here movement-wise go, well, clearly that’s progressive?”

Johnson made the opposite case in his own words. Unveiling the budget, the mayor said he would “challenge the ultra-rich,” asking large corporations and the wealthy to carry more so working families would not face higher property, grocery, or garbage taxes. When the council stripped the tax out, he called the result “morally bankrupt,” arguing there was a way to balance the budget without going after working and poor people. His finance team projected the council plan would leave a $163 million hole to be filled mid-year. Johnson ultimately neither signed nor vetoed the $16.6 billion budget, allowing it to take effect.

On the left, losing the head tax read as a gift to the wealthy. Ald. Byron Sigcho-Lopez (25th Ward), a fellow socialist who championed it, called the budget that passed without it “the billionaire budget” and said it was unacceptable to put the burden on the poorest people in the city. The opposition reached beyond the left. Gov. JB Pritzker, a fellow Democrat, also opposed the head tax, calling it a “job killer.” Vasquez lands in neither camp. He says he would tax corporations, just not this one.

On policing, Vasquez makes a budget argument. He says the department should be audited like any other, noting that it takes up roughly half of the corporate fund’s departmental spending, the discretionary pool the city actually controls. The department’s 2026 budget runs about $2.1 billion, and its overtime line alone grew by roughly $101 million. “You talk to the cops, they don’t feel it, they don’t see that money,” he says. “So my premise is, something’s inefficient here. We’ve got to audit this thing.” Johnson, like most politicians, is unwilling to say that publicly, he argues, so the number keeps climbing.

He frames the larger development as a shift in how Chicago governs. By moving the budget cycle up about six weeks, the council was able to draft and pass a spending plan over a mayor’s objection for the first time in the city’s history. Vasquez voted against that plan too, calling both it and the mayor’s version inadequate, yet he counts the council’s new independence as a gain. “The fact that the council is becoming more independent, I think is a good thing for government, because you have a counterbalance.”

The ordinance that was never tested

Immigration is the fight that has consumed his chairmanship of the City Council’s Committee on Immigrant and Refugee Rights, and here the lines with Johnson are less clean. Chicago’s Welcoming City Ordinance, which says city employees will not help enforce federal immigration law, began as a 1985 executive order under Mayor Harold Washington, became law in 2006, and was expanded in 2012. Vasquez says it had never been seriously tested until the past year. After a mass detention incident in June 2025 and a hearing his committee held that July, it became clear that no rule explicitly empowered the Civilian Office of Police Accountability to investigate whether Chicago officers had violated it. His own attempt to fix that was blocked by colleagues. A similar measure led by Ald. Jessie Fuentes (26th Ward) passed in March 2026, and on this one Johnson was a supporter.

Then the issue reached his own staff. His chief of staff, Catherine “Cat” Sharp, was one of the “Broadview Six,” a group of protesters indicted by a federal grand jury in October 2025 over a September protest outside the suburban ICE facility. Prosecutors alleged the group surrounded a federal agent’s vehicle, banged on it, and damaged it. Vasquez characterizes the prosecution as an attempt to frighten people out of dissent. The case then collapsed. Charges against Sharp were dropped in March 2026, and the remaining charges were dismissed with prejudice in May after the U.S. attorney acknowledged the case was tainted and a federal judge said she was shocked by redactions prosecutors had made to grand jury transcripts. Vasquez goes further than the court record, saying the prosecutor “tampered with the grand jury,” and says he wants hearings to put that official under oath. That word is his. What the record establishes is that there was documented prosecutorial misconduct serious enough to end the case.

Legislating at teenagers

The same instinct shaped his opposition to a proposed snap curfew aimed at large gatherings of teenagers. He grounds it in his own adolescence, getting cleared out of Navy Pier by police for gathering to rap while, as he puts it, “the white kids are chilling.” He was arrested twice before he turned 18, once while sitting near friends who were smoking, and says officers ignored him to arrest the three Black kids beside him, then listed him as gang-affiliated anyway. He never got a record. “That changes the trajectory of your whole life,” he says. He does not dismiss the underlying problem. Large groups gather, he acknowledges, and a few make it dangerous for everyone. His complaint is that the city legislates at teenagers without consulting them.

“It’s a whole bunch of old people out here talking about what the teens are doing, and there’s not enough talking to them.”

What he leaves open

He closes on a warning aimed past his own ward, at the mayoral cycles ahead. Chicago tends to pick whoever it thinks can win, he says, with little regard for who can govern a city this large amid a budget gap the city pegged near $1.15 billion. “We like to slug it out.” He wants the next race to be a test of who can do the job.

It is a notable position from a former battle rapper, a politician who built a career on landing the sharper shot, and it is the tension GrayStak went to his office to examine. Vasquez presents himself as a politician who trusts his own judgment of a fight over any faction’s. The censure, the votes against an allied mayor, and the colleagues he has irritated are the evidence he points to. Whether that record reads as principle or as positioning is a judgment we leave to you. The full conversation is above.

Christopher Sweat is co-founder and CEO of GrayStak, which covers social and political dynamics in real time, as well as institutional politics at the local, state, and federal levels. This piece draws on a sit-down interview with Ald. Andre Vasquez at his 40th Ward campaign office; the full conversation runs on video with this piece. Vasquez’s words are quoted from that interview. Other figures, dates, and officials’ public statements have been checked against the record and linked throughout.

That looks like one town and one project. It is a single room in a national fight waged one room at a time, in city halls, county boards, and statehouses, where almost no one sees the whole shape, which is far bigger than a power bill. While towns argue over water and noise, the federal government has reclassified these buildings. They are national security infrastructure now, their construction accelerated by executive order, their opponents tracked by the agencies that watch for terrorism. The fight that looks like zoning is a fight over critical infrastructure: who pays to build it, who profits, and who gets watched for opposing it.

The Windfall and the Nuisance

The Joliet project gets told two ways, both partly true. The first is a windfall. The developers project thousands of construction jobs and roughly $2.1 billion in local tax revenue over thirty years, and the building trades turned out because the construction work is real, well paid, and union. A single hyperscale campus can become one of a county’s largest taxpayers; in DeKalb, most of Meta’s data center property taxes flow to the schools. The officials who vote yes are not naive. They face a real choice: take the tax base, or watch the next town take it while the grid strain arrives either way.

The second story is a nuisance. The campus is designed to draw up to 1.8 gigawatts of electricity and as much as 150,000 gallons of water a day, on an aquifer already running dry, in a city whose Lake Michigan pipeline will not be finished until 2030. For all that, it promises about 700 permanent jobs, roughly $28 million in investment per permanent job, which makes it plain these are not jobs programs.

Both stories are true, and both stop at the Joliet city limits. The project runs much further.

One Model, Run Town by Town

Start with who builds it. The Joliet developer, PowerHouse Data Centers, owned by American Real Estate Partners, has by its own count 118 data centers underway or planned, totaling more than 8 gigawatts across eight or nine markets, built from a base in Northern Virginia, the largest data center market in the world. Its partner Hillwood is part of the Perot family of companies. The tenants are the hyperscalers: Google, Amazon, Meta. The Joliet vote is a national decision in local dress, one move in a model the same firms run in dozens of jurisdictions at once.

Watch how differently those jurisdictions answer. A few exits west, in January, Naperville voted 6-1 to deny a Karis Critical data center after residents warned about power load, noise, and diesel generators. The same developer took the same pitch north to Hoffman Estates, where in June the plan commission rejected the rezoning 4 to 2 before a packed room, even with two data centers already going up in the village. Lisle postponed a hearing the room could not hold. Blue Island residents booed the developers. Rockford drew petitions against a roughly 1,300-acre proposal. Some places say yes: in Sangamon County, the County Board approved a $500 million CyrusOne campus after turning down a moratorium. Illinois already hosts 139 operating data centers with more than 120 planned, fourth in the country.

Each town votes on a national phenomenon with its own zoning code and its own exhausted neighbors, and the answers come out scattered, with no one summing the column. The sum is already large. Data Center Watch counted roughly $64 billion in projects blocked or delayed amid local opposition over two years, then another $98 billion in a single quarter of 2025. The era of rubber-stamping these as quiet tax generators is over.

The Demand Never Blinks

I have spent two decades on the enterprise side of this, in telecom and colocation and then AI infrastructure, and the part that rarely reaches the local coverage is why the demand never blinks. These campuses get built against contracted enterprise workloads, the data and AI systems that banks, insurers, and manufacturers now run as core operations, plus the training and inference loads stacked on top. The demand is as durable as the businesses underneath it. The siting logic has only intensified since the colocation era: go where power and land are cheap, and the fiber already runs, and move faster than the local process can adapt. Joliet checks every box, and so does much of the Midwest, which is why a no vote in Naperville does not stop the machine. It moves it one exit down the interstate.

The jobs math follows from the same engineering. A data center is one of the least labor-intensive structures in the economy. Construction is real work, often a thousand trades for a year or two, but those crews increasingly travel from build to build rather than getting hired locally. Once energized, a facility runs on automation and a small, specialized staff usually imported rather than drawn from the neighborhood; Virginia’s legislative auditors put a typical site at roughly fifty full-time workers, half of them contractors. Economists do find some real local gains over several years, but the headcount inside the fence stays small, and that is the number the ribbon-cuttings skip.

A Local Benefit, a Socialized Cost

The thing residents fear most, their electric bills, is real, and the price is set on the regional grid, far from the Joliet podium. Joliet sits within PJM Interconnection, which runs the wholesale market for 67 million people across 13 states, home to the world’s largest concentration of data centers in Virginia. The cost of that growth shows up across the whole grid:

PJM has no authority to refuse a data center connection even when supply is short, and the region is at risk of falling below its reliability standard as soon as 2027. The residents asking who gets favored in a brownout were describing the actual mechanism, one operating a thousand miles from their council and beyond its power to vote on.

A word of caution, because the number is contested. The capacity-auction spike is a forecast, not a meter reading. The research firm SemiAnalysis argues it is largely a simulation that does not reflect actual conditions, driven by the grid operator’s own demand projections, and notes that Texas, with a comparable buildout but a different market design, saw power futures rise only a few percent; by that reading the real increase to PJM bills is closer to 15 percent, with market design a bigger factor than data-center demand alone. PJM’s own market monitor reaches the opposite conclusion, calling data-center load the primary driver and noting that the price impacts are large and irreversible. The industry’s economists add that high prices are a signal meant to pull a new generation onto the grid, though new plants take years to build. The cost is real, large, and genuinely disputed.

Hold the numbers side by side. The benefit a town like Joliet is promised is local and capped: a few hundred permanent jobs and about $70 million a year in tax revenue, the developers’ own $2.1 billion projection spread across thirty years. The cost is regional and open-ended: the data-center share of the grid’s capacity bill ran to $9.3 billion in a single year, carried by 67 million people who never sat in any of these meetings. One town books the bounded upside; everyone on the grid splits the unbounded tab. That asymmetry, a local benefit against a socialized cost, is the engine under the whole buildout; the math keeps working no matter how many town halls go badly.

The Rooms Off-Camera

The leverage lives elsewhere, in two rooms, both off-camera. The first is the grid and its regulators, PJM and the Federal Energy Regulatory Commission, where decisions about who connects, in what order, and who pays are made. The second is the statehouse, where in Illinois a bill sits drafted in a drawer.

It is the POWER Act, for Protecting Our Water, Energy, and Ratepayers, filed as SB4016 and HB5513. It would do what no single zoning board can: make data centers pay for their own energy and the renewable infrastructure to generate it, mandate water-use transparency and state environmental assessments, and require community benefit agreements. It stalled in committee and did not pass this spring. After the General Assembly adjourned without it, Pritzker acted on his own authority: on June 5, he directed the state to stop processing new data-center tax incentives starting July 1 and laid out a framework to bill data centers for the grid and water costs they impose, after years in which Illinois handed 27 facilities some $983 million in breaks. Two months earlier, he had joined a bipartisan coalition of PJM-region governors pressing the grid operator to keep data-center demand off consumers’ bills. The week Joliet approved its campus, the Illinois Commerce Commission required large developers to post bigger deposits to shield ratepayers. The industry shows up too: the Data Center Coalition told lawmakers it is committed to covering its full cost of service. As lawmakers adjourned, residents rallied at the Capitol, where Brenda Santoyo of the Little Village Justice Organization warned that Illinois is being targeted by Big Tech for its freshwater.

Here is the tell. The durable fix, a statute that would set one set of terms for every town at once, did not pass. What moved instead came from the governor’s pen, real but narrower, and a successor can undo it. So the fight stays in the city halls, one at a time, where opposition is hardest to coordinate. One legislator put it plainly, describing communities forced to respond one village at a time, one zone meeting at a time, and noting that the communities absorbing the downsides are disproportionately communities of color. That fragmentation, the fight staying in the city halls one at a time, is the condition the buildout’s economics depend on.

There is a catch hanging over all of it. The same federal government accelerating the buildout is in a preemption posture. A December 2025 executive order set the Justice Department to challenge state AI laws it judges onerous and threatened federal funding to states that pass them. It carves out data-center and energy laws like the POWER Act, but the carve-out is contested: it binds only a future legislative proposal, not the litigation task force or the funding threats, and it leaves permitting, where local zoning lives, open to federal reform. For now the levers hold; the permitting order does not override state zoning or utility regulation, which is why 27 states are legislating, and Maine is moving toward the first construction moratorium. The open question is how long that holds. It would not be the first time Washington took a siting decision from the towns in its path: interstate gas pipelines run under federal authority over local objection, and a 1996 law limited how far localities could go in refusing a cell tower. The data-center fight is early in that same arc.

The Unpriced Risk

The exposure sorts into layers: developers mostly private and invisible, PowerHouse and the Perot family’s Hillwood among them; customers the hyperscalers driving the expansion; landlords the public REITs such as Digital Realty, behind contested projects in Chicago’s South Loop. What they share is a risk that never appears on a site plan. Moratoriums, the tax-credit suspension, cost-allocation reform, and the 550-page suit three residents filed to void the Joliet approval all bear on whether a given gigawatt comes online on schedule, late, or never. That is the variable the market treats as settled and the record shows is anything but.

Computing Power Is National Power

Washington has settled what these buildings are. They are national security infrastructure. Executive Order 14318, signed in July 2025, declared AI data centers and the power lines feeding them essential to national security and directed the federal permitting process to be fast-tracked. The 2026 National Cyber Strategy names data centers alongside energy, finance, and water among the critical sectors to harden. Data centers have nominally sat inside the information-technology critical-infrastructure sector since that framework was set in 2013. What is new is the explicit national-security treatment, the acceleration, and the push by security officials to make data centers a standalone sector of their own.

The strategic case, as the administration and the AI industry present it, deserves to be stated at full strength. Computing power, they argue, has become national power: the country that trains the most capable models gains an edge in weapons, intelligence, and economic productivity, and the administration is explicit that it is racing China for what it calls global AI dominance. From that vantage, a data center is the physical substrate of the contest, the thing a government wants built fast and built here. Great-power competition runs on threat perception as much as fact, on readings of intent that are subjective and sometimes mistaken. The framing need not be right to become the organizing logic of federal action. It already has. And once a category is wrapped in that logic, the question stops being competitiveness and becomes who counts as a threat to it.

None of this surprised me. The AI infrastructure I worked in came out of the defense contracting world, where the line between enterprise compute and national security has been blurred for years. The executive orders did not invent that overlap; they formalized it. The same architectures that run a bank’s models run an intelligence agency’s.

Designating something as critical infrastructure changes who guards it and who watches the people who oppose it. The clearest published case is a December bulletin from the Delaware Valley Intelligence Center, the Philadelphia fusion center housed inside the city’s police department, obtained by The Intercept. Marked for official use only, it concluded that domestic violent extremists were likely interested in targeting AI data centers as a physical and cyber threat, then conceded it had no specific information about any actual plan. The indicators it told police to watch for included disruptive First Amendment activity, online calls to boycott or protest, and criticism of higher utility bills caused by data centers. Asked about the bulletin, the Philadelphia Police Department said the center respects lawful protest and meant only to provide situational awareness, not to treat protected speech as criminal. It is part of a pattern. WIRED published more than 1,000 pages of unpublished DHS, FBI, and fusion-center reports built around a label, anti-tech violent extremism, that it noted does not appear in the agencies’ public domestic-extremism guides. Those files cite real sabotage chatter, online talk of magnets and explosives against data centers and white-supremacist agitation, and they pull protests, public meetings, and memes into the same net.

The practical effect is concrete. Once data centers are designated national-security infrastructure and opposition to them is filed under a domestic-extremism heading, the ordinary acts of local politics- a public comment, a photograph of a substation, a post about a rate increase- become entries a fusion center can log and circulate to police. The category does the work, placing lawful objection on the same page as sabotage. This has happened before: the Brennan Center has documented fusion centers treating environmental and racial-justice organizers as threats, and law enforcement has already monitored data center protests and the meetings around them. The First Amendment has no exception for unpopular infrastructure. A fight that began over water and electric bills is now one the national security apparatus is formally inside.

Add Up the Rooms

Spend enough years across that matrix, and you stop seeing separate fights and start seeing one system. The buildout is approved, denied, and stalled in hundreds of venues at once, while the cost of the power it demands is socialized across 67 million people through a grid almost none of those venues control, and the one tool that could set consistent terms sits unmoved in a Springfield committee. Washington has already chosen. The market prices the expansion as inevitable. The record shows a political constraint being built in real time, in church-basement zoning meetings and one-in-the-morning council votes, with the bills landing on households across thirteen states.

That constraint is still scattered, which is why it stays underpriced, and whether it stays scattered is the whole question. If the fight stays in the city halls, one zone meeting at a time, the model wins on attrition. If it consolidates, into a passed POWER Act, cost-allocation reform at the utility commissions, a ratepayer politics that crosses the usual lines, the economics change.

So the next time a data center clears a council at midnight and the story files itself under local zoning, ask the questions that room was never built to ask: who pays for the power, who profits, where it truly gets decided, and who gets watched for objecting. Add up the rooms and the shape appears: a national security buildout approved one local vote at a time, with the bill landing on households that never voted on it and the people who object increasingly logged in a federal file. Mostly, no one is adding them up. That is the point.

A note on method and disclosure: I have worked in enterprise technology for roughly two decades, across telecom and colocation, enterprise SaaS, and AI infrastructure, and I studied national security organization and policymaking at the University of Colorado Boulder. I continue to advise in the technology sector and in wealth management, and I build GrayStak, which covers Illinois politics at the state, regional, and local level. I have worked in this industry, and I read these facilities from that vantage point. Some of the firms I advise have commercial relationships with companies named here, including the hyperscale cloud providers. I hold no stake in the specific projects discussed. GrayStak holds no financial position in any company named here. The analysis rests entirely on the public record. It documents footprint and process and does not impute intent, and where figures are contested, as the grid-cost numbers are, I note the disagreement and leave it open. Figures are current as of mid-2026.

Sources and method

Every figure and claim is linked to its source inline. Local and trade reporting came from Shaw Local, CBS Chicago, WGN, the Daily Herald, Block Club Chicago, Capitol News Illinois, NPR Illinois, STLPR, Data Center Dynamics, Bisnow, The Real Deal, NBC News, and The Hill. Grid and cost figures came from IEEFA, CNBC, E&E News, Utility Dive, NRDC, Data Center Watch, and PJM’s independent market monitor, Monitoring Analytics. The national security and surveillance material draws on WIRED, The Intercept, Gizmodo, the Brennan Center, federal executive orders, and the 2026 National Cyber Strategy. Preemption analysis cites client alerts from Paul Hastings and Manatt, the MultiState tracker, and the Center for American Progress; the employment figures come from Brookings, Good Jobs First, the National Community Reinvestment Coalition, and Virginia’s Joint Legislative Audit and Review Commission. This is analysis of the public record, not investment advice.

And in a sense, that is all that happened. Palantir changed a mailing address.

Because when you read past the post and into the company’s own annual filing, the picture is narrower than the headline. On paper, the principal executive office is now in the Miami area. The people are not. Palantir does not disclose how many employees it has in Colorado, but no relocation of any of them was announced, and only months earlier the company had been building out a new office in Cherry Creek. That gap between the address a company claims and the capability it actually maintains is the whole story, and it is a story that runs the length of an interstate.

The filing moved. The workforce stayed.

Drive it

Get on Interstate 25 outside Santa Fe and point the car north. I lived on this corridor for twenty-four years, in Colorado Springs, in Denver, and in Boulder, and from the driver’s seat it does not read as anything unusual. It is just the road home. But you are not on an ordinary highway. You are driving the spine of the American national security state, and almost none of it is hidden. Every node on this corridor is documented in public filings, local newspapers, and federal contract databases. You just have to be willing to read them as a single map.

Start at the southern end, in Albuquerque, where Sandia National Laboratories sits on Kirtland Air Force Base alongside the Air Force’s nuclear weapons and research work. An hour or so north, just off the interstate above Santa Fe, is Los Alamos, where the laboratory designs the country’s nuclear weapons and maintains the stockpile under federal contract. Two of the country’s three nuclear weapons laboratories sit within an hour of the same road, before you have even left New Mexico.

Cross into Colorado, and the corridor changes uniformly from laboratory to command. Colorado Springs is the densest stretch of all of it. Peterson Space Force Base hosts the North American Aerospace Defense Command and U.S. Northern Command. Cheyenne Mountain, the hardened complex carved into the granite, sits a few miles away. Schriever runs satellites and the GPS constellation. Fort Carson trains the Army’s Fourth Infantry Division. The Air Force Academy turns out officers above the city. The regional business press has a name for this stretch, the corridor of Colorado spacepower, and the boosters who use it are describing exactly the same geography I am, just with a different purpose.

Keep going north into the Denver metro and the corridor shifts again, from public command to private contractor, though the line between the two is thinner than the categories suggest. Lockheed Martin’s space division runs a sprawling campus in Jefferson County, out in Waterton Canyon south of the city, a manufacturing and testing hub for GPS III satellites, missile-warning systems, and NASA’s Orion crew capsule. Buckley Space Force Base, out in Aurora, runs the missile-warning mission. And until February, downtown Denver and then Cherry Creek held Palantir.

Northwest of the city, in Broomfield, sits the company that was formerly Ball Aerospace. In 2024, the British defense giant BAE Systems bought it for more than five billion dollars and folded it into a new division called Space and Mission Systems. The interesting part, for our purposes, is what happened to the location. Nothing. The division’s president said plainly that the company was staying in Colorado because the billions already invested were in the facilities, capabilities, and people, all of which were here. A foreign multinational acquired one of the corridor’s crown jewels but could not meaningfully move it because its value was bolted to the ground.

Finish the drive in Fort Collins. Woodward, headquartered there, makes the fuel, actuation, and motion control systems that go inside commercial and military aircraft, and its biggest customers are the largest aerospace primes in the country: RTX, GE Aerospace, and Boeing. The company reported record net sales of $3.6 billion in its 2025 fiscal year. It is the sort of supplier most people have never heard of, and the defense aviation business cannot function without it. A few miles away sits what is left of Numerica, a firm that spent years building radar and sensor-fusion software for air and missile defense, and its recent history is the corridor’s logic compressed into a single company. In 2022, it sold its space tracking telescope network to Slingshot Aerospace. In early 2025, the California defense technology firm Anduril bought its radar and command-and-control business and folded the technology into its Lattice platform. The valuable assets were acquired by larger players, but the engineering work remained where it was built, in northern Colorado.

That is roughly four hundred and fifty miles of road, packed end to end with the public and private machinery of defense. And these are not separate dots that happen to share a highway. They feed one another. The laboratories design the weapons; the commands above Colorado Springs operate them and the satellites that guide them; the contractors build the hardware and write the software that ties the two together. They draw on the same pool of cleared engineers, who move among the labs, the primes, and the startups without ever leaving the corridor, carrying their clearances with them. The federal money that enters at one node circulates through the rest. This is less a row of installations than a single machine stretched across the length of a state and a half.

Put a number on it. The two New Mexico laboratories at the southern end run about five billion dollars each, better than ten billion a year between them. At the northern end, Colorado drew $22.8 billion in federal defense and aerospace contracts in 2024, a record, by the count of its own economic development office, on top of billions more flowing to its military bases and its federal labs. Set the laboratory budgets beside the contract awards and this single interstate, by a conservative tally, channels well over thirty billion federal dollars a year, before a cent of military payroll is counted. None of it is secret, and all of it is on the public record. There is no other stretch of American road like it.

Two stories about one move

There are two ways the Palantir departure has been told, and the distance between them is instructive.

The first is a victory story, and it belongs to the activists who spent more than a year organizing against the company’s presence in Denver. They objected to Palantir’s contracts with immigration enforcement and to its work for the Israeli military, and they made the company’s offices a regular site of protest. The Cherry Creek building was vandalized with the words “Palantir Out”. When the Miami announcement came, organizing groups including Denver Anti-War Action and the national Purge Palantir campaign read it as the direct result of their pressure, a repeat of 2020, when the company had left the Bay Area for Denver after a similar campaign. They have said they intend to follow the company to Florida.

The second is the documented story, and it is more complicated. Palantir’s own filing did not cite protests as the reason for the move. It instead pointed to Colorado’s 2024 artificial intelligence law and to climate-related risks to its operations in Colorado. A Colorado labor organizer who had been part of the protests told CBS Colorado that he did not think any single thing had driven the company away. And the hardest fact for the victory story to absorb is the one already on the record: the people did not leave, and the timing of the company’s own real estate moves runs counter to any picture of a hasty exit. In the autumn of 2025, only months before the Miami announcement, Palantir took roughly eleven thousand square feet at the Financial House in Cherry Creek and, public records show, spent around $227,000 building it out, upgrading from its longtime downtown offices in the Tabor Center. That space was meant to be the company’s new headquarters. Reporting at the time described the Miami announcement as scrapping those Cherry Creek plans, and sublease listings in the building surfaced by the following spring, so what the company holds there now is unclear. But a firm spending real money on a new Denver headquarters in September was not a firm with its bags packed for Florida, which tells you how fast the decision was reversed. And the destination is its own kind of tell. The global headquarters Palantir moved to is a coworking suite at an Industrious location on Biscayne Boulevard in Aventura, the principal executive office of a company worth hundreds of billions of dollars reduced to a flexible workspace address.

How many people that even involves is itself disputed, and the honest thing is to say so. Palantir reports only a global workforce in its filings, with no state-by-state breakdown, and estimates of its Colorado presence range widely. The Common Sense Institute, citing figures from the workforce-data firm Unify, put the Denver office at about 87 people. Governor Polis said the company’s executives told him it had closer to five hundred employees in the state. Palantir has long run a distributed workforce, so the true number physically in Colorado is genuinely unclear. But the direction of travel is not. No relocation of those employees to Florida was announced. The headquarters that moved was a line in a filing, not a building full of people boarding flights to Miami. Whatever moved to Miami, it was not the part of Palantir that does the work.

I am not in a position to say which account is closer to the company’s true motive, and that is precisely the point. I do not need to. The public record cannot tell us what is in a corporation’s head, but it can tell us with great precision what is in its buildings, and the buildings in Denver still have people in them. The honest version of this story does not allege a hidden motive. It reports the gap between the announcement and the footprint, and lets the reader sit with it.

What it actually costs to leave

If you want to know how hard it is to move real capability off this corridor, you do not have to speculate, because the federal government is attempting it right now, in public, with one of the corridor’s marquee institutions.

U.S. Space Command was provisionally headquartered at Peterson in Colorado Springs. In September 2025, the Trump administration designated Huntsville, Alabama, as the permanent home of the command, reversing a Biden-era decision to keep the command in Colorado. So the address, in the most literal sense, has been ordered to move. And here is what moving it actually looks like.

As of this spring, about twenty people were working out of a temporary office at Redstone Arsenal in Huntsville. The command expects to grow that to roughly two hundred by the end of the year. The permanent headquarters is not expected to break ground until 2027 and not to be finished until around 2031, with another year after that to move people in. Somewhere between 1,400 and 1,700 personnel are expected to relocate over roughly 5 years. To make any of it happen, the command is paying relocation bonuses to convince people to go and retention bonuses to convince others to stay in Colorado Springs until their jobs are physically moved. And the state of Colorado is suing the federal government to stop the whole thing.

That is what it takes to move a headquarters when you actually mean to move it: years of construction, active litigation, bonuses paid in both directions, and a decade-long managed migration of cleared people and classified work. Set that next to a one-sentence post on X announcing a new mailing address in Miami, and the contrast tells you everything about which kind of move is real. Even Space Command, with the full force of a presidential order behind it, is discovering that the corridor does not readily relinquish its capabilities. The flag can be reassigned in an afternoon. The infrastructure takes until 2031.

And notice what is staying behind even as the flag departs. The Space Force units remain at Peterson and Schriever. Cheyenne Mountain is not going anywhere. Local officials in Colorado Springs, by their own accounts, have watched the commercial space sector continue to grow throughout the entire saga. The headquarters designation is leaving. The corridor is not.

Everything that matters sits downstream of a budget line in Washington, not a logo in Denver.

The dependency

The dependency here runs in one direction, and it is not the direction the announcements imply.

A defense company’s address is the most mobile thing about it. It can be changed to chase a tax advantage, to escape a state law, to make a political statement, or simply to follow a founder’s preference, and it can be changed overnight. Everything that actually constitutes the enterprise is far more difficult to move. The cleared workforce, built up over years. The physical proximity to the bases and laboratories that generate the contracts. The supplier relationships measured in decades. The classified facilities that take years to build and certify. These are the assets, and they are tethered to the corridor whether or not the logo on the filing still says Colorado.

This is why I would not write that Palantir left Denver. Palantir’s address left Denver. The company, in the sense that matters, the people and the work, did not go anywhere, the same way BAE’s Colorado division stayed put through a foreign acquisition, the same way the bases and contractors are staying put even as a four-star command begins its slow march to Alabama.

Consider what Palantir actually is. Its federal contracts reached $970.5 million in 2025, according to USAspending.gov, and in July of that year the Army awarded it an enterprise agreement worth up to $10 billion dollars over the next decade. Its business is the government, and it has bound itself to that customer for years to come. What roots a book of business like that, if not a headquarters sign? It is the cleared, specialized workforce and the federal relationships built up over years, and none of that boarded a plane to Florida when the filing changed.

This is the part worth sitting with, because it changes where the leverage is. A corridor that moves more than thirty billion federal dollars a year, anchored by clearances, contracts, and proximity that took eighty years to assemble, does not answer to a press release and will not be moved by one either. The organizers who spent a year trying to push Palantir out of Denver were right about the company and aimed at the wrong lever. You cannot relocate a machine like this by making a single city uncomfortable, because the machine is not held in place by any one company’s willingness to stay. It is held by federal appropriations and the political choices that route them. The address on a filing is the most movable thing in the entire system, which is exactly why a company reaches for it first when it wants to appear to have changed. Anyone who wants to understand or contest the power in this corridor should follow the money to its source rather than the sign to its new ZIP code.

The corridor was built over eighty years, from the first laboratory above Santa Fe to the missile components rolling off the line in Fort Collins, and it was built in concrete, clearances, and proximity. Those are not things you relocate with a post. So the next time a company on this road tells you it has moved, the useful question is not where the filing now says it lives. The useful question is whether anyone actually packed a box. Usually, if you read the public record carefully, the answer is no.

A disclosure: I lived on this corridor for twenty-four years, in Colorado Springs, Denver, and Boulder. I studied political science at the University of Colorado Boulder, and in 2024 I ran for Congress in Colorado’s Fifth Congressional District, the Colorado Springs-centered district whose economy is anchored by several of the installations described above. I hold no financial position in any company named here.

Sources and method

This piece is reported entirely from publicly available records: local and state reporting in Colorado and New Mexico, SEC filings, federal contract data on USAspending.gov, and company and government disclosures. It documents footprint, not intent, and attributes every characterization to the party that makes it. The corridor’s federal-dollar figure is a deliberately conservative tally, the two New Mexico laboratories’ budgets set beside Colorado’s own reported defense and aerospace contract awards, before military payroll is counted. Key facts are linked inline above. Figures are current as of mid-2026; this is analysis, not investment advice.

On May 25, less than a day after that piece went up, the Bank of Israel cut its policy rate from 4.00% to 3.75%. The cut was the third since November 2025. The Monetary Committee cited falling inflation, the strength of the shekel, and Israel’s strong macroeconomic data. The committee paused at its February and March meetings during the active conflict in Iran. The May 25 decision resumed the normalization path the Bank had been on before the war interrupted it.

Three signals, one direction

The rate cut is the first signal. The Bank of Israel’s own forecast trajectory, published in March, called for two more cuts to 3.5% by early 2027. The May 25 move was the first. The institution is executing on its pre-war path, treating the Iran conflict as a pause rather than a redirection.

The shekel is the second. The currency hit a 33-year peak against the dollar on the day of the announcement, trading at roughly 2.89. Since the March rate decision two months earlier, the shekel had appreciated 8.3% against the USD, 7.2% against the euro, and 7.4% on the nominal effective exchange rate. Currency markets do not move sovereigns to multi-decade strength while pricing them as risky credits.

The inflation print is the third. Annual inflation in April was 1.9%, close to the midpoint of the 1% to 3% target band and within the range for the seventh consecutive month. The Iran war’s supply-side disruption has been absorbed without lasting damage to the inflation environment.

Each signal could move on its own for reasons specific to its domain. Central bank decisions reflect more than risk-premium readings. Currencies move on flow dynamics that may have little to do with credit conditions. CDS spreads can compress on dealer positioning. What is harder to explain is the agreement of all three. Three independent signals reading the same condition through different channels is an inference to a common cause.

The Bank of Israel itself was careful about the framing. The Monetary Committee noted that “geopolitical uncertainty remains significant, both domestically and globally,” and flagged risks of “a renewed acceleration of inflation” tied to geopolitical developments, energy prices, and fiscal pressures. The cut reflects the institution’s read on current conditions, not a declaration that those conditions are settled.

The methodology is theirs

In its May 25 press release, the Bank of Israel addressed Israel’s risk premium directly: “After increasing at the start of Operation Roaring Lion, Israel’s risk premium, as measured by the CDS spread and by the spreads on dollar-denominated and shekel-denominated government bonds vis-à-vis US Treasuries, declined to near its level from before October 7, 2023.”

That is the same analytical framework the original piece used: CDS spreads as the measure of credit risk, with USD eurobond spreads as the cross-check on the illiquidity premium in non-tradable Israel Bonds. These are not tools imposed on the question from outside. They are the standard market-implied measures of sovereign credit, used by the institution under analysis to assess its own risk premium.

The May 25 statement also makes a specific empirical claim. Israel’s risk premium has returned to its pre-October 7, 2023, level. That is the central bank’s own assessment, in its formal monetary policy statement.

The primary market priced this directly

In early January 2026, the State of Israel completed its first international bond offering since the Gaza ceasefire — $6 billion in dollar-denominated sovereign debt across three tranches at 5, 10, and 30-year maturities. The spreads over comparable U.S. Treasuries were 90, 100, and 125 basis points, respectively. The weighted-average spread of approximately 102 basis points represented a 34 percent tightening from Israel’s 2024 sovereign offering.

Demand reached $36 billion — six times the amount sold. More than 300 institutional investors from over 30 countries participated. The underwriting syndicate included Bank of America, Citi, Deutsche Bank, Goldman Sachs, and JPMorgan. Sovereign wealth funds from Abraham Accords signatory countries participated through their flagship investment arms. The bonds listed on the London Stock Exchange Main Market.

Israel’s Accountant General Yahli Rotenberg characterized the offering result as “a return to the spread levels that preceded the war.”

This is the institutional sovereign bond market expressing the same view in different vocabulary. The Bank of Israel measures the risk premium as having returned to pre-October 7, 2023, levels. The credit market prices it at a 63-basis-point CDS spread, tighter than several formally higher-rated peers. The primary sovereign bond market measures it as ten-figure institutional demand at pre-war spread levels with the largest global underwriters.

For the comparison to Illinois, the January issuance provides additional empirical reference. Israel’s tradable 5-year USD eurobond is priced at 90 basis points over Treasuries. The Illinois Treasurer’s February 2026 Israel Bonds renewal — same sovereign, similar effective maturity, but non-tradable — priced at 114 basis points. The gap of approximately 24 basis points between Israel’s tradable and non-tradable USD debt issued weeks apart is the cleanest available empirical signal that the non-tradable structure carries a measurable additional premium.

What the cut compresses

The cut does not refute the Illinois Treasurer’s stated justification for holding $100 million in Israel Bonds. The office defends the position on technical fiduciary grounds — yield over Treasuries, default history, statutory eligibility. The May 25 cut leaves all three grounds formally intact.

What it affects the size of the headline number underneath that framing.

The 63 bp credit-risk component of the 114 bp spread tracks Israel’s CDS spread. As CDS compresses — and the BoI rate cut, the 33-year shekel peak, and continued macro normalization all point in that direction — the credit-risk portion compresses with it. The illiquidity premium, by contrast, is structural. It is set by the non-tradable feature of the instrument and does not move with macro conditions.

In practical terms, the spread Illinois will be able to claim over Treasuries at the next renewal will likely be at or below the 114 bp claimed at the February renewal. With Israel CDS now at 63 bp and pre-war levels around 50 bp, the credit-risk component has perhaps 10 to 15 bp of additional compression available. The illiquidity premium holds approximately constant.

This is the natural arithmetic of macro normalization. The yield framing produces smaller numbers over time, even with no change in policy or process on the office’s end.

The position itself does not change. What changes is the strength of the yield-based case for it.

What remains

The technical fiduciary framework supporting the position stays in place. The press releases will continue announcing the purchases. The yield language will continue appearing. The statutory eligibility framework will continue to apply.

The question the original piece raised is separate from any of that. Why this particular eligible foreign sovereign, given the universe of eligible foreign sovereigns that meet the same statutory tests and pay similar market-clearing spreads? By what process is the selection made? With what disclosure? These are questions that existing law already requires public investment fiduciaries to answer.

The Bank of Israel’s normalization does not affect those questions one way or the other. They are procedural questions about how selection decisions are documented and disclosed. They apply equally to any foreign sovereign instrument held in a state portfolio, regardless of which sovereign it is or what macro conditions surround it.

The original piece argued that the spread the Treasurer’s office cites as taxpayer value amounts to market-clearing compensation for the risks the position entails. The May 25 decision is consistent with that reading. The credit market continues to price Israel as a normalizing investment-grade sovereign. The central bank is managing policy on that basis. And the procedural question about selection — separate from any view about Israel, the conflict, or state foreign policy — is independent of all of this.

Written in a personal capacity. Standard disclaimers from the original piece apply.

Sources: Globes, The Jerusalem Post, Times of Israel, Israel National News, Xinhua, Middle East Monitor; Bank of Israel press releases dated November 24 2025, January 5 2026, February 23 2026, March 30 2026, and May 25 2026 (English text confirmed at boi.org.il); Israel Ministry of Finance and Accountant General statements on the January 7-13, 2026 sovereign bond offering; Arnold & Porter and the joint book-running managers’ deal documentation; tradingeconomics.com and exchange-rates.org for currency and rate history; investing.com for CDS quotes. CDS values in the chart are approximated from public sources; precise daily quotes vary intra-month.

He was right that Amazon was an unsafe place to work in the early weeks of the pandemic. He was right that workers at the Staten Island JFK8 warehouse, a million-square-foot facility with ten- to twelve-hour shifts and a thirty-minute lunch break, could win a union election against the second-largest private employer in the country, even when nearly every major labor organization in America believed they couldn’t. Now, at a warehouse where 5,500 workers still don’t have a contract, he says he is right that Amazon will eventually be forced to the bargaining table. The federal government has ruled in his union’s favor on that point twice. Amazon is in court trying to argue that the federal labor board itself shouldn’t exist.

And on May 4, three weeks before this conversation, Smalls decided he was right that Jeff Bezos, the lead sponsor of the Met Gala, should not be allowed to walk a red carpet thirty minutes from the warehouse without consequence.

“There’s no way I’m going to allow this billionaire to come to New York City and do a fashion red carpet gala thirty minutes away from the warehouse that he refuses to negotiate a contract with,” Smalls says. He jumped a barricade. The NYPD tackled him. He spent twenty-four hours in a Manhattan jail. He is now facing charges for resisting arrest, trespass, and three other counts. They get dismissed if he stays out of trouble for six months. His entire life suggests he has no intention of doing that.

“We put so much fear in Jeff Bezos,” he tells me, “that he didn’t show up on the red carpet.”

The book

In two weeks, on June 1, Crown will publish When the Revolution Comes: A Fight for the Future of the Working Class. The memoir opens with a prologue on what it feels like to walk inside a grocery warehouse, then circles back to his childhood and moves chronologically through to the 2022 union victory at JFK8.

It’s not a manifesto, despite the title. Publishers Weekly called it “plainspoken” and noted that it “leans more toward personal history than political rallying cry.” Smalls confirms as much when I ask what he wants readers to take from it.

“A lot of people see a piece of themselves inside of me and my story,” he says. “Whether it’s evictions, whether it’s divorce, whether it’s having kids, whether it’s getting fired from a job several times, whether it’s hitting rock bottom — all of that is in the book as well. It’s not just about the success that I had at Amazon. It’s also about the failures I had as well.”

That’s the thread he wants readers to follow. The book moves through his father’s stints in prison, an early Amazon firing for two minutes of “time theft” (he was rehired), a Florida community college he didn’t finish, sound-engineering classes he didn’t finish, a basketball dream cut short by a hit-and-run on the job, an ex-wife and twins, and a brief rap career. He did a showcase with Meek Mill in New Jersey, he clarifies, not a tour. Multiple outlets have gotten this wrong. The cumulative effect is a portrait of a working-class American who happened to be standing where the pressure broke first, not a labor icon.

His mother was a member of the healthcare workers’ union SEIU 1199 for more than twenty-five years and never once mentioned it at home. “She wasn’t like an organizer. She just was a rank-and-file member.” When I ask if he ever imagined this life, he laughs. The skill sets he draws on now, he says, are the ones he watched his mother use to hold down their household. “The grit and the grind that I saw my mom go through, it carried over into the things that I do and how I fight back now.”

The Palestine turn

The Chris Smalls who became famous in 2020 was a workplace organizer. The man in front of me in 2026 is something bigger and less easily categorized. He talks about Gaza, Cuba, ICE, surveillance, and the prison industrial complex as if they’re the same fight.

Last July, he boarded the Handala, a vessel operated by the Freedom Flotilla Coalition, attempting to break the Israeli blockade of Gaza with humanitarian aid. Israeli forces intercepted the ship in international waters. Smalls, the only Black activist on board, was singled out for particularly aggressive treatment, by his account and the accounts of others on the ship. He spent five days in Israeli custody, hunger-striking the entire time, before being released without charges at the Jordan border.

He is now banned from Israel, he says, for 100 years.

The flotilla, he tells me, is what ties it all together. “Once again, connecting the dots with Amazon and genocide was something important to me.” Amazon and Google’s joint $1.2 billion Project Nimbus contract provides cloud infrastructure to the Israeli government and military. “It’s the same technology that’s being used to empower ICE over here,” Smalls says, “and the local police departments and the rain cameras and also the government contracts they have with the Pentagon and the military-industrial complex.”

His argument: the line between an Amazon warehouse worker in Staten Island and a Palestinian under occupation runs through the same cloud server.

He invokes the longshore workers’ creed. “There’s a saying in the labor movement from the ILWU. Injury to one is injury to all. And it doesn’t say injury to one is injury to all except for Palestinians.”

That position has put him at odds with the leadership of the AFL-CIO and the International Longshoremen’s Association, both of which he has publicly accused of complicity in the war. Among nationally known U.S. labor figures, he is more or less alone on this.

Surveillance and sacrifice

I ask him about the cost. He has been arrested in New York, detained in Israel, and most recently stopped by ICE at the U.S. border on his return from a humanitarian aid mission to Cuba.

“My phone was taken by ICE,” he says. As we are speaking, he tells me, a comrade of his has just been detained at JFK airport. The agents are asking him questions about Smalls.

“I’m being surveilled,” he says. “I’m being targeted, of course. I’m being doxxed. It comes with the sacrifice of the work that I have to do, unfortunately. But I ask myself if I don’t do the work, then who else is going to do it?”

He grew up, he tells me, in a context where this was always going to be the cost. “I was in handcuffs by the time I was ten years old. The average Black man has probably been in handcuffs at least once by the time they’re in their thirties. I can sit here and say I’ve been in handcuffs a dozen times, and I probably have a dozen more to go.”

He pauses. “Martin Luther King was arrested, I believe, twenty-nine times. So I’m trying to break his record.”

You can read that line as bravado, or as something more deliberate: a claim to a tradition, a way of locating himself inside a longer arc of Black American resistance. I think he means it both ways.

The institutional reality

The harder question, the one I keep circling, is what all this adds up to institutionally. Smalls is no longer the president of the Amazon Labor Union. His successor, Connor Spence, led the union through the recent National Labor Relations Board ruling that ordered Amazon to bargain. It was the most significant legal victory at JFK8 since the original union vote. Smalls’s old union publicly described his Met Gala arrest as a “lone-wolf direct action” that was not coordinated with leadership.

The arc that landed him outside that institution is worth naming. The ALU Smalls founded in 2021 was conceived as a rejection of established labor. If the big unions had been capable of organizing Amazon, he argued at the time, they would have done it already. By June 2024, that posture had reversed. ALU members voted 98 percent in favor of affiliating with the 1.3-million-member International Brotherhood of Teamsters. The leadership election a month later was won by Spence, who had founded the dissident ALU Democratic Reform Caucus and sued Smalls in 2023, alleging he had refused to hold officer elections in violation of the union’s own constitution. Smalls did not run.

To grasp the distance he has traveled, it helps to remember how high the rise was. In 2022 he was named to the Time 100, fielded statements of solidarity from Bernie Sanders, AOC, and Elizabeth Warren, and watched Amazon’s general counsel hand him a national platform by calling him “not smart or articulate” in a leaked internal memo. His profile from that period is mostly intact. His institutional standing is not. He has no organizational vehicle of his own anymore; The Congress of Essential Workers, the nonprofit he founded in 2020, appears largely dormant. The Teamsters local he now belongs to answers to international president Sean O’Brien, who delivered a primetime address at the 2024 Republican National Convention. Smalls, whose recent advocacy has centered on Palestinian solidarity and a public attack on the AFL-CIO leadership, has not publicly addressed the contradiction.

When I ask how he balances direct action with the slower, more institutional work of organizing, he is generous about both. “Everybody has a role to play. I don’t think everybody has to do the same thing. Whatever you can contribute to the movement, you do that. My skill sets are different. My skill sets have always been direct action.”

“Coalition building is always going to be hard,” he says. “But one thing people know about me, whether it’s 100 people, 1,000 people, or 10 people, I’m going to go forward. No matter who shows up, no matter how many people we get, we still got to show up and do what we can.”

That’s the most honest articulation of where Smalls is in 2026, I think. He’s not trying to run anything. He’s trying to be the person who shows up when whoever is running things can’t or won’t. Whether one man jumping a barricade is any substitute for the patient, unglamorous work of contract negotiation is the open question. It will define his next decade.

He seems aware of it. When I ask what keeps him grounded, he goes back to the warehouse.

“My feedback was Chris has sided too much with the workers,” he tells me. That was the reason Amazon gave for never promoting him to salaried manager, despite forty-nine applications. “And it’s like, yes, absolutely. I’m siding too much with the workers, because I was a worker before I became an assistant manager.”

“The only thing that matters to me,” he says a moment later, “is seeing people get liberated from being oppressed.”

What comes next

We end where these conversations always end, on the question of the future. I ask if he has anything else he wants to say.

“No,” he says. “I mean, this is great. I’m always glad to spread the word and plant seeds. The revolution starts with yourselves.”

The line works as both a book ending and a personal credo. He’s been saying some version of it for years. When the Revolution Comes is out June 1. The bargaining order against Amazon stands. The contract at JFK8 does not yet exist.

Chris Smalls is thirty-seven years old. He has, by his own count, several dozen more arrests ahead of him before he meets Dr. King’s record.

He has time.

Christopher Sweat is a journalist with GrayStak, which covers political and social dynamics in real time across Washington, DC, New York, and Chicago: local, state, and federal politics from the streets to the institutions. The full video interview with Chris Smalls is available on the GrayStak YouTube channel.

When the Revolution Comes: A Fight for the Future of the Working Class is published by Crown on June 1, 2026.

On January 6, 2026, Palm Beach County Comptroller Mike Caruso announced that the county had purchased another $350 million in Israel Bonds. The new buy brought the county’s position to exactly one billion dollars, making Palm Beach the largest local-government holder of Israel Bonds in the world. Caruso, a CPA by training, was careful with his framing. “Investment policy is not politics,” he said. “It is purely a function of safety, liquidity, and market rate of return. Israel Bonds are paying a higher rate than other allowable investment alternatives, including U.S. Treasuries.”

Three months earlier, the North Carolina Treasurer’s office had done the opposite. After several years of letting the position run off without any new purchases, the office quietly sold its remaining $6.7 million in pension holdings. The Treasurer’s office told a coalition of activists pushing for divestment that the state hadn’t been adding to the position for a simple reason: they weren’t finding it compelling from a risk-and-return perspective.

Two state-level fiduciaries, both bound by statutory investment criteria, reached opposite conclusions about what the math says.

The question that organizes itself is which one is right.

The answer turns out to be more useful than either side of the public debate has acknowledged. There is a market-implied tool that almost nobody in this conversation has been using, and when you apply it, both of the standard arguments collapse in different directions. A sharper question emerges from the wreckage.

This piece is an attempt to lay out what that tool shows.

The frame that isn’t working

The public debate about state investment in Israel Bonds has been running on two parallel tracks for two and a half years.

The first track is moral. Israel Bonds are general-obligation debt: the State of Israel uses the proceeds for any purpose, including military spending. For people who object to Israeli government policy on civil-rights or international-humanitarian grounds, that fungibility is the issue. The frame is conscience and complicity.

The second track is financial. State treasurers point to coupons, default histories, and spreads over U.S. Treasuries. Israel has never missed a payment on Israel Bonds in seventy-plus years. The yield premium has typically run between 100 and 150 basis points over comparable Treasuries. For a fiduciary operating under a statutory obligation to seek market return, that looks like a defensible investment.

These two arguments do not engage each other. They run on different vocabularies, in different rooms, with different evidentiary standards. The moral argument operates on conscience. The financial argument operates on yield. Neither side concedes anything to the other, and neither side needs to, because they are not actually debating the same thing.

What is missing from both is the market’s own answer to the question. That answer is now available.

What CDS prices say

A credit default swap is an insurance contract on sovereign debt. The buyer pays a periodic premium; the seller agrees to compensate the buyer if the sovereign defaults. The premium, called the CDS spread, represents the market’s collective real-time estimate of how likely default is. It strips out duration, currency, coupon convention, and everything else that complicates a comparison of bond yields. It just prices the probability of getting wiped out.

Israel’s five-year CDS spread, as of mid-May 2026, sits at 63 basis points. That is the market’s price today, on the risk that the State of Israel defaults on its dollar debt within the next five years.

The arc is what you would expect from a sovereign that fought a multi-front war and then settled. The pre-war baseline was about 50 bp. The Hamas attacks on October 7, 2023, pushed the spread into a sustained climb. Moody’s first downgrade took it past 150 bp by February 2024. The peak, roughly 170 bp, came in April 2024, around the Iranian direct strike on Israel. The Fitch downgrade in August 2024 kept things elevated. The Lebanon ceasefire in late 2024 marked the beginning of the compression. Israel’s June 2025 strikes on Iran produced a transient spike. The 12-day U.S.-Israel war against Iran in February-April 2026 produced another spike to around 110 bp, which has now fully unwound. Israel CDS is currently at the low of the entire post-war range.

Markets have priced through the geopolitical risk premium of the past two and a half years. The implied probability that Israel defaults in the next five years, using the standard 40-percent recovery rate assumption, is about 5 percent cumulative, roughly 1 percent per year. That is not a fragile credit.

It is also, more interestingly, a credit being priced more tightly than its formal ratings would suggest.

Israel’s 63 bp CDS sits below Saudi Arabia at 83 bp, Indonesia at 86 bp, Mexico at 88 bp, and India at 88 bp. Each of those countries carries a comparable or formally lower credit rating. The market is pricing Israel as cleaner credit than the rating agencies, by a measurable margin.

This is the first thing that breaks. The argument that Israel Bonds is somehow a financially imprudent investment, the argument that various divestment campaigns have made, and the argument that the North Carolina Treasurer’s office gestured at do not withstand contact with CDS data. The market does not think Israel is a stressed credit. It thinks Israel is a roughly A-tier sovereign, currently priced more like A+ than like Moody’s Baa1.

So one side of the financial debate is wrong on the data.

The other side is wrong too, and for a reason that turns out to matter more.

What 114 basis points actually pays for

When Illinois Treasurer Michael Frerichs renewed $15 million in Israel Bonds this past February, his office announced it the way every state treasurer announces these purchases. The bond paid 4.86 percent on a five-year maturity. That outperformed a comparable U.S. Treasury by about 114 basis points, which the Treasurer’s office characterized as 23 percent additional return for Illinois taxpayers.

114 basis points sounds like a lot. It is, in the absolute. But that framing imports an assumption that does not survive examination: that Treasuries are the right benchmark.

A Treasury is risk-free. Israel Bonds are not. They carry sovereign credit risk and an additional structural feature that ordinary sovereign bonds lack: they are non-tradable. There is no secondary market, no way to sell into a crisis, no daily mark-to-market price. You hold to maturity, or you do not hold.

The 114 basis points are not the free yield. It is compensation for taking on credit risk and giving up liquidity. CDS data lets you see exactly how much for each.

Sixty-three basis points is the credit spread, the compensation any A-rated USD sovereign bond would pay for the credit risk involved. It is the standard market price for accepting that this is not a Treasury. Fifty-one basis points is the illiquidity premium, the compensation for accepting the non-tradable structure, derived by comparing the Israel Bonds yield to Israel’s actively traded USD eurobonds of similar maturity.

What remains after stripping out those two is the excess return, the alpha, that an investor earns relative to a fair-priced equivalent.

That number, on the math, is approximately zero. Possibly slightly negative, since Israel CDS now prices tighter than peers like Saudi Arabia, meaning a fair-priced equivalent would actually pay slightly less than 114 bp over Treasuries.

Israel Bonds are not a bad investment. They are not a good investment. They are a fairly priced investment. Holding them is neither financially imprudent nor financially generous. The instrument clears the market the same way any A-rated sovereign instrument with restricted liquidity does.

Once that lands, the entire framing of the debate falls apart.

What is actually being decided

The Treasurer’s defense rests on the premise that the position generates value for taxpayers. The decomposition shows that the position generates the standard market price for the risks it carries. No surplus. No discount. Market clearing.

The divestment activist’s critique rests on the premise that the security is somehow financially problematic, the kind of thing prudent fiduciaries should avoid on its risk profile alone. The decomposition shows that the security pays exactly what it should pay.

Both sides have been arguing about whether Illinois should hold an investment that delivers an excess return. The market says no such investment exists. The 114 bp spread is rent on credit and rent on illiquidity, both fairly priced. There is no surplus to defend or to criticize.

What is left, once both financial pretenses come down, is a question that nobody on either side has been asking with any clarity. If the financial case is neutral, why is this particular security being chosen by these particular officials over the dozens of other foreign sovereign instruments that meet the same statutory eligibility tests? That is not a financial question. It is a question about the decision-making process. It is, in other words, a fiduciary-procedure question. And the procedural question is the part of this whole conversation that turns out to have actual structural leverage.

The Illinois architecture

Most public discussion of “Illinois’s investment in Israel Bonds” treats Illinois public money as a single pot. It is not. Several distinct fiduciary regimes operate in parallel, with different oversight, different disclosure rules, and different statutory authority.

The two that matter most for this conversation:

The Illinois State Treasurer’s Office manages approximately $39 billion in state operating funds, including cash from tax receipts, fees, and fund balances awaiting disbursement. Treasurer Michael Frerichs is the sole fiduciary. Decisions are his alone, within statutory eligibility criteria. The $100 million Israel Bonds position lives here. Every purchase comes with a public press release naming the amount, coupon, and maturity. That visibility is why everyone knows about it.

The Illinois State Board of Investment, ISBI, is a nine-member board with fiduciary responsibility for the pension assets of three of the state’s pension systems: the State Employees Retirement System (about $24 billion), the Judges’ Retirement System (under $1 billion), and the General Assembly Retirement System (under $100 million). Frerichs serves on the board as Vice Chair. ISBI does not publish line-item holdings. Whatever Israel's exposure pension money has at this level is not externally verifiable from public filings.

These are different pots. They get routinely conflated in public conversation, and that conflation has been expensive for everyone trying to think clearly. When divestment campaigns target “Illinois’s investment in Israel Bonds,” they almost always mean the Treasurer’s $100 million, because that is the only piece publicly disclosed. The pension systems operate under entirely different rules.

And ISBI is only three of Illinois’s pension systems. The Teachers’ Retirement System (TRS, about $79 billion, 456,000 members) has its own independent fifteen-member board. The Illinois Municipal Retirement Fund (IMRF, about $55 billion, 520,000 members, and the only Illinois pension system that is actually well-funded at 95.8 percent) has its own elected eight-member board. The State Universities Retirement System (SURS, with about $25 billion in assets) has its own eleven-member board. The Chicago Teachers’ Pension Fund (CTPF, with about $13.7 billion in assets) is separate from all of them. Plus the Chicago municipal pension funds, the Cook County funds, the police and fire funds, and the roughly 660 downstate police and fire pension funds.

“The Illinois pension system” is a sentence with at least five major subjects and dozens of minor ones. Effective intervention has to specify which one.

The statutory lever

The Treasurer’s $100 million sits within a statutory eligibility framework that imposes two binding tests on foreign sovereign investments. To be eligible, a foreign sovereign issuer must have met its bond payment obligations for at least 25 consecutive years and must be rated in the three highest classifications by at least two nationally recognized rating services.

Israel passes the first test. The Israel Bonds program has honored payments since 1951. Israel currently passes the second test on the strength of S&P’s A rating and Fitch’s A rating. Both of those agencies still classify Israel in their top three categories: AAA, AA, A.

Moody’s, which downgraded Israel from A1 to Baa1 in two separate moves across 2024, including an unusual two-notch jump from A2 to Baa1 last September, has Israel in its fourth category. Under one common reading of the “three highest classifications” language, Moody’s may not already count toward the statutory eligibility test. Which means Israel’s current statutory eligibility hangs on S&P and Fitch both staying at A or above.

A single-notch downgrade by either of those agencies to BBB+ or below would mean Israel has only one rating service classifying it in the top three categories. The statute would no longer be satisfied. The Treasurer would have no authority to continue holding the position, regardless of his preferences, and the bonds would automatically become ineligible at the next purchase decision.

This is the most consequential fact in the entire conversation, and almost nobody is talking about it.

Both S&P and Fitch have restored their outlooks to stable as of late 2025 and early 2026, following the ceasefires. But the rating-action cycle is reactive to events on the ground, not to outcomes anyone has guaranteed. A renewed conflict, a fiscal deterioration, or a debt-service stress event could prompt another notch move at either agency. If it happens at the wrong agency, the question of whether Illinois holds Israel Bonds gets answered by statute, not by anyone’s political preference.

For anyone trying to think about leverage in this conversation, that is the lever. It is automatic. It does not require legislative action. It does not require political will. It only requires that the rating agencies do what rating agencies do, and the relevant agencies are S&P and Fitch.

Illinois in the national context

Illinois is one node in a national pattern. Since October 2023, U.S. state and local treasuries have purchased over $1.7 billion in Israel Bonds across more than 35 jurisdictions.

Illinois sits fifth. Palm Beach County alone exceeds the next four combined. The Illinois $100 million is a meaningful position by Treasury standards (about 0.25 percent of the state operating portfolio), but it is small relative to the leading state holders and dwarfed by Palm Beach.

Movement in the other direction is starting too, though the dollar amounts are modest. North Carolina sold its pension position late in 2025. New York City, under former Comptroller Brad Lander, stopped new purchases for the city pension. Active divestment campaigns are tracked by Jewish Voice for Peace and Palestinian-led partner organizations across more than 10 states, with ongoing efforts in New York State, Ohio, Pennsylvania, and Miami-Dade County.

The North Carolina pattern is the most legally durable of these. The Treasurer’s office did not cite human rights. It did not invoke international law. It did not engage the moral argument. What it told the divestment coalition, by accounts from people in those conversations, was that the office no longer found the position compelling on risk-and-return grounds. That is a fiduciary-frame argument, made within the standard investment-policy review process, requiring no legislation. It does not trigger anti-BDS preemption statutes because it is not a boycott motivated by national-origin animus. The fiduciary review concludes that the security is not compelling.

This is, in my reading, the most analytically defensible path forward for state-level officials who want to move away from these holdings. It does not require winning the moral argument. It does not require winning a legislative fight. It requires asking treasurers and pension boards a procedural question they are already required under existing law to answer. What review process governs the choice to hold this security over other eligible alternatives? How is that process documented? What disclosure accompanies the decision?

For state-level Treasurer holdings, that question lands on a single elected official, in a context where the press-release norm has already established a high disclosure baseline. For pension-system holdings, it lands on boards with explicit fiduciary obligations under the state pension code. Either way, the question is procedural rather than substantive. It does not ask the official to take a position on a contested international conflict. It asks them to document and defend their decision-making process, which, under existing law, they are already required to do.

The choice that requires articulation

What CDS data has done is establish, with as much market authority as anything ever gets established, that the financial premise underlying the standard defense of these holdings is not what it appears to be. The 114 basis points over Treasuries is not alpha. It is rent on credit and illiquidity, paid by the State of Israel and collected by the holder, in roughly the amount that any other A-rated sovereign would pay. That is not a critique. It is just what the market shows.

The choice to hold this security, when fair-priced alternatives exist, requires articulation. Not as a confession. As a normal incident of fiduciary review. The Treasurer cannot honestly say this is a high-return investment. The activist cannot honestly say this is a financially imprudent investment. Both have been wrong, in opposite directions, on the same data. What replaces both is a more concrete question about the decision-making process, situated within a body of public investment law that is much better suited to the question than either the moral framework or the financial-prudence framework that has been doing the work so far.

Illinois has $100 million on the table, statutorily authorized, fairly priced, and held in a portfolio whose decision-maker is required by Illinois statute to make those decisions on fiduciary criteria. The question worth asking, in that exact frame, is what the fiduciary criteria are, how the choice among eligible foreign sovereigns is documented, and what disclosure accompanies the decision. That is a question that existing law already requires an answer to.

The most useful thing the CDS analysis does is tell us that, whatever the answer turns out to be, it cannot be “because this is the best return available.” The math does not support that claim. The market does not support that claim. The answer will need to be something else.

Articulating that something else, on the public record and with documentation that survives fiduciary review, is the move actually available here. It does not require winning the argument that nobody has been winning for the past two and a half years. It only requires that officials do, on the public record, what their fiduciary obligations already require them to be doing on the public record.

That is what the spread, properly decomposed, makes possible.

Sources: investing.com daily CDS quotes; Moody’s, S&P, and Fitch rating actions; Illinois Treasurer’s office press releases; relevant pension fund annual reports; Bond Buyer; Times of Israel; Palm Beach County Clerk’s office press releases. CDS trajectory chart shows approximate monthly closes; precise daily values vary intra-month.

Written in a personal capacity. Views are my own and do not represent GrayStak or any other affiliated organization. The author holds no financial positions in Israel Bonds, has no compensated relationship with the Illinois Treasurer’s Office, the Development Corporation for Israel, or any divestment coalition, and has not been paid to write this piece. CDS values are approximated from public sources; precise daily quotes vary. Statutory interpretations reflect the author’s reading and have not been adjudicated in court. The account of the North Carolina Treasurer’s Office’s reasoning is reconstructed from public reporting and descriptions of those conversations by divestment coalitions; the office has not made formal public statements directly addressing Israel Bonds as a category. I have written in other contexts on regional foreign policy and political risk questions; this piece is procedural and analytical rather than positional. Not investment advice.