New York's $36 Billion Kalshi Case — The One Claim No CFTC Rule Can Erase

Four Federal Rulings in Nine Days, the Remedy That Reaches the Whole Company, and the Claim No CFTC Rule Reaches Backward to Undo

Part of the MindCast National Prediction Market Litigation Architecture series: The Missing "Gaming" Definition | CFTC Takes On Nine States | The Ninth Circuit Forum Fight | The Rule 40.11 Paradox | State Legislators and the Gaming Boundary

Kalshi runs a federally registered exchange where people buy contracts paying out on real-world events — which team wins, who takes an election, how a season ends. Federal regulators call the contracts derivatives. More than a dozen states call them bets. Courts have split for two years, and four rulings in late July and early August finally decided something.

Nearly every account is scoring those rulings as wins and losses. Reading them that way misses what changed. New York's Attorney General now seeks at least $36 billion from a company the same petition values at $22 billion — a demand exceeding the enterprise itself, aimed at conduct already completed, filed in the county where Kalshi keeps its headquarters.

The Commodity Futures Trading Commission holds one curative instrument: a pending rule that would finally define "gaming," the statutory term the whole fight turns on. No version of that rule automatically extinguishes money claims already accrued. MindCast AI modeled the eight-actor field across three certified simulation runs and registered where the contest moves next.

The full publication is available at https://www.mindcast-ai.com/p/kalshi-ny-unicorn.

The summary below carries the news, the structural diagnosis at the paper's core, the simulation and what it refused to release, the full forward register, and what each institutional reader should do with it.

The News, and Why Sequence Changes the Reading

Four federal decisions landed in nine days. Read individually, each one misleads. Read in order, they rewrite the Commission's position.

July 27 — Minnesota. A federal court blocked the state's new statute banning prediction markets outright. Suing states directly, on the preemption theory the Commission presses nationally, produced exactly the result the agency wanted.

July 29 — Wisconsin. Judge William Griesbach refused the Commission the same relief across twenty-eight pages, holding the agency unlikely to prove that sports contracts even qualify as the financial instruments it claims, and unlikely to win on preemption regardless.

July 31 — New York. Judge Jed Rakoff refused the Commission's emergency request to stop the state Attorney General from prosecuting her own case, finding neither a strong likelihood of success nor irreparable harm.

August 4 — Utah. Judge Robert Shelby entered final judgment for the State, becoming the first federal court to decide the preemption question outright rather than provisionally, and closed the case.

Attribution shifted somewhere inside that week, and the shift matters more than the scoreboard. In earlier losses the Commission was not a party and could attribute outcomes to the registrant's briefing. Griesbach and Rakoff assessed the same theory in cases the Commission itself filed.

New York's contribution is not the setback. Wisconsin delivered that two days earlier with far more reusable reasoning. New York's contribution is the remedy sitting behind the setback — a petition demanding a customer-by-customer accounting, restitution, disgorgement, triple gain, and $100,000 per unauthorized offering, reaching contracts across sports, culture, elections, and other events rather than sports alone.

Four rulings therefore produced four different things: a federal win on new statutes, a merits defeat on old ones, an emergency defeat inside an enterprise-scale case, and the first final judgment in the country.

The Structural Diagnosis: A Timing Problem, Not a Doctrine Problem

Two findings govern the paper, and both cut against how the fight is usually framed. Neither turns on who has the better legal argument.

Old statutes are beating new ones. Sorting the four rulings by the kind of state law each tested makes a scattered record look orderly. Minnesota's brand-new prediction-market ban lost. Wisconsin's pre-existing commercial gambling statute, Utah's criminal gambling code, and New York's general illegality statute all survived.

The mechanism is procedural rather than ideological. A new statute aimed at prediction markets hands a federal judge a clean question — does federal law displace this state law? Nothing else needs deciding first, and clean supremacy questions are the ones preemption answers well. An old general prohibition denies the court that shortcut, because the judge must first construe state law, and construction opens every doctrine favoring states.

The practical consequence cuts against the loudest legislative instinct in the states: passing a new prediction-market statute can convert a stronger posture into a more vulnerable one.

The remedy runs on a clock no rule reaches. Judge Shelby defeated preemption through a statutory provision listing exactly which transactions federal law shields from state gambling regulation. Sports contracts on a registered exchange appear nowhere on that list. Congress wrote the list before the modern prediction-market sector existed, and the omission does the work.

A ruling built that way needs no definition of gaming whatsoever. Finalizing the pending rule answers the reasoning other courts have used and answers nothing in Utah.

New York's demand meanwhile reaches conduct completed since January 2025, while the Commission's instrument shapes conduct going forward. Defend-then-define assumed federal timelines outrun state ones — and Utah moved from complaint to final judgment in roughly six months.

Both findings point the same direction. The Commission's position is deteriorating on schedule rather than on merit, and no available federal action resets the clock.

The Interstate Finding: A Hub, Not a Template

Uniqueness raises an obvious objection. If headquarters jurisdiction is what makes New York work, forty-nine states cannot copy it, and the case teaches them nothing.

Answering the objection produces the paper's most consequential call. A copyable instrument would produce fifty separate lawsuits rising and falling alone. A unique instrument produces dependence — New York's outcome moves every other state's position without any of them holding the same tool.

What travels is everything the proceeding produces. The petition demands an accounting identifying each customer and itemizing bets placed, money lost, and gains received. Per-customer attribution is precisely the capability every other attorney general currently lacks, and a court order producing it creates a template requiring no headquarters nexus to copy. Confidentiality may block sharing the underlying data; the design travels regardless.

Infrastructure to carry the transmission already exists. Thirty-seven states and the District of Columbia filed a joint brief against Kalshi in April, and fifteen states addressed prediction-market legislation in 2026.

Other states will therefore copy accounting fields, remedy design, product distinctions, and enforcement language. Headquarters nexus stays behind, and the paper prices what that asymmetry does to every actor holding an unexercised claim.

The Evidentiary Weight

Courts reached the paper's governing thesis without reaching for it, which is corroboration a framework cannot manufacture for itself.

MindCast filed the definitional-gap argument with the Commission on April 17, 2026, as a public comment on the very docket that became the June proposed rule. The argument held that the agency's refusal to define gaming was the structural fault line driving national litigation rather than a housekeeping omission.

Judge Torres decided the New York federal case through that gap on July 7, eighty-one days later.

Griesbach and Shelby then went further than the thesis predicted, defeating preemption on statutory structure that never reaches the definition at all. The publication reports the correction against MindCast's own prior framing rather than around it — a court can now defeat federal preemption without the definition ever mattering.

Independent judicial arrival at a published thesis, followed by a documented correction to that thesis from the same record, is what a falsifiable framework looks like in operation.

The Simulation: Eight Cognitive Digital Institutional Twins, Three Certified Runs

Every structural claim routed through the MindCast AI Proprietary Cognitive Digital Twin Foresight Simulation (MP CDT FS) engine before finalization. Eight institutional twins carried the field — the Commission's enforcement arm, its rulemaking apparatus, the New York Attorney General, the registrant, the multistate coalition, federal judicial construction, and two state classes sorted by posture. All eight were frozen at an August 7 evidence cutoff and released only after passing the engine's integrity gate.

Three results carry the analysis, and each concerns discipline rather than output.

The engine blocked eight predictions and published every one. Whether attorneys general deliberately shift instrument form failed on a four-docket evidence base. Whether staffing shortages drive the Commission's rulemaking delay failed on an unpopulated record. Which direction uninvolved states' incentives point failed because two channels resolve in opposite directions and the net could not be established.

One blocked prediction cuts against MindCast's own thesis. The claim that agency habit rather than statutory structure generates the losses degraded during simulation after the Utah judgment and lost release eligibility. A registry that retires its own governing mechanism when the record turns is the one whose forward numbers carry weight.

Two runs disagreed on the governing mechanism, and the paper reports the disagreement rather than resolving it.Remedy scope and constraint geometry both survived every test, so the engine branched instead of averaging incompatible readings into a synthetic score.

Whether money or doctrine arrives first is the contest's decisive uncertainty. Manufacturing a single number would have concealed it, and concealing it would have made every downstream forecast less useful.

The Forward Calls

Every prediction below carries a band, a window, a named settlement source, and a falsifier, so each can fail in public. Bands rather than point estimates reflect genuine uncertainty rather than false precision, and the full falsifiers appear in the publication.

Primary predictions

• Retrospective New York exposure survives every available prospective federal instrument. 78–88%, by June 30, 2027.

• New York's outcome is cited in other states' filings, with no state replicating the instrument. 74–86%, by June 30, 2027.

• The Commission's own proposed rule is cited against it as ambiguity evidence or classification concession. 72–84%, by January 31, 2027.

• The next federal merits ruling turns on statutory structure rather than the missing definition. 70–82%, by March 31, 2027.

• The Second Circuit decision revises more of the register than any other scheduled event. 70–82%, at decision.

• The New York proceeding returns to state court. 70–82%, by December 31, 2026.

• Kalshi lists new non-sports contract categories before seeking any state gaming licence or settling across multiple states. 68–80%, by June 30, 2027.

• The Commission's options narrow after the next merits ruling regardless of which side wins. 65–78%, by March 31, 2027.

Interstate predictions

• A qualifying New York accounting triggers a comparable data demand by another state within 180 days. 70–85%, conditional.

• No complaint with five or more state plaintiffs emerges. 81–92%, by June 30, 2027.

• Of the first three previously uninvolved states to act, at least two use something other than a lawsuit first. 67–82%, conditional.

• New actions by uninvolved states stay product-specific or harm-specific rather than categorical. 71–85%, conditional.

• Public state exposure data activates a tribal or compact instrument within 180 days. 64–81%, conditional.

Secondary predictions

• Kalshi obtains no state gaming licence anywhere. 72–84%, by June 30, 2027.

• No settlement resolves exposure across three or more states within four quarters. 62–75%.

• Old general statutes outperform new prediction-market bans by at least three to one in surviving preemption challenge. 62–75%, by June 30, 2027.

• Multistate coalition coordination degrades observably as recovery becomes rivalrous. 52–66%, by December 31, 2027.

• At least one court outside Utah adopts the enumeration reasoning as an independent ground. 48–62%, by June 30, 2027.

Scenario-contingent forecasts, with conditions fixed in advance rather than selected once the record arrives:

• A money resolution arrives before any national doctrinal resolution. 45–60%.

• The Second Circuit opinion rests on the savings language or the statutory enumeration rather than definitional status. 55–70%.

Tail risks

• Congress legislates a definition. 8–15%.

• The Supreme Court intervenes on an emergency basis before any appellate court rules. 5–12%.

Read together, the bands describe a Commission losing options rather than losing a case — and a state proceeding whose clock runs faster than any federal instrument aimed at it.

What This Means for Your Institution

Six audiences read the same record and face different decisions. None of them needs to win the national classification fight before acting.

• State attorneys general. Enforcement under existing gambling law has survived where new prediction-market statutes have not, and New York's accounting demand describes attribution infrastructure worth preparing for now rather than after another state builds it. MindCast maps which instrument form survives a preemption challenge and where the evidence architecture transfers.

• Prediction-market platforms and registrant counsel. Geofencing resolves market access and does nothing for disgorgement, restitution, or triple gain on completed transactions. MindCast models retrospective exposure as a separate line from prospective access, and simulates which consequence-side product lines survive a functional screen.

• Institutional allocators and counterparties. Platform-wide relief against a company headquartered in the enforcing county differs in kind from a state geofence, and the branch between money and doctrine determines which risk arrives first. MindCast prices the segmentation, judgment, and finalization timelines against a specific position.

• Tribes and tribal gaming authorities. Friend-of-the-court participation preserves nothing — twenty-three tribes sought that status in Utah and the court denied the motion as moot after ruling. MindCast maps the compact-rights vehicle before the federal record hardens.

• Congressional offices and policy counsel. Judge Shelby's holding turns on a list Congress wrote decades ago, a drafting artifact rather than a policy judgment, and any categorical bill lacking express concurrent-authority language reproduces the same defect. MindCast's constraint-geometry analysis identifies where the savings-clause question decides the outcome.

• Gaming operators and licensed incumbents. Parity arguments strengthen as state enforcement instruments survive and weaken if the Second Circuit reverses. MindCast models which framing a reviewing court credits and when the licensure off-ramp becomes the industry's live demand.

Across all six, the same asymmetry holds. Every actor listed above can move on the record as it stands, while the Commission remains the only participant whose options shrink with each week it declines to act.

The Through-Line

Prediction markets will keep operating. The open question is who draws the boundary and who pays for the conduct that preceded the drawing.

A regulator can absorb losses indefinitely when they land on someone else, arrive slowly, and address conduct a later rule can bless. Nine days in late July and early August degraded all three conditions at once. Wisconsin removed the attribution shield. Utah supplied finality and a route around the Commission's principal cure. New York attached the same demonstrated failure mode to a remedy reaching the company rather than one state market.

Minnesota remains the counterweight and deserves its weight. The Commission's lawsuits work, and work well, against new state prohibitions — which is the one kind of statute states are now least likely to pass.

Institutions holding a position in this fight have a window before the enforcement instrument's shape gets fixed. The window closes on a state-court clock rather than a federal one, which is why the timing question now matters more than the doctrinal one.

MindCast AI runs this architecture — Cognitive Digital Twins, Dynamic Predictive Game Theory, and dated, falsifiable registers — on complex litigation, innovation economics, and geopolitical risk. Related work on this contest: A Boundary Rule with a Functional Core · The CFTC NPRM Is a Litigation Brief · Kalshi Loses Federal Forum · Competitive Federalism: A Field Guide for State and Tribal Regulators · Prediction Markets and the Dual Nash-Stigler Trap · Live-Fire Simulators and Runtime Predictive Infrastructure.

If your institution holds a position in the prediction-market fight, faces the gaming-boundary question, or needs the instrument space mapped before it hardens, outline your matter below and our team will respond with next steps — for suitable matters, a tightly scoped pilot simulation against your decision window.