Resolution criteria
Scope: This market resolves to YES if a major federal artificial intelligence (AI) regulation bill is passed by the United States Congress and signed into law by the President (or otherwise becomes law, such as via veto override) on or before September 30, 2026, at 11:59 PM Eastern Time (ET). Otherwise, this market resolves to NO.
Definition of "Major Federal AI Regulation Bill": To qualify, the legislation must be a federal law that establishes a binding national regulatory, auditing, licensing, or safety framework governing the development, training, or deployment of general-purpose AI, foundation models, or frontier AI systems.
What qualifies: Bills that establish mandatory third-party audits, risk-management frameworks, register/reporting requirements, or explicit civil/criminal liability for frontier AI developers (e.g., the federal FRONTIER Act / Great American AI Act or similar comprehensive bipartisan bills if successfully enacted into law).
What does NOT qualify: State-level laws (e.g., Illinois' SB 315), federal bills addressing narrow AI issues (e.g., the TAKE IT DOWN Act or other bills targeting only AI-generated deepfakes/CSAM), legislation restricted solely to federal agency procurement or voluntary internal guidelines, or broad annual appropriations and spending bills that happen to contain minor AI funding riders.
Source of Truth: Legislative status will be verified using the official U.S. Congress tracking database at Congress.gov or official White House press releases.
Background
While individual states have moved to pass binding safety laws—such as Illinois enacting SB 315 in July 2026 to require third-party audits on powerful AI models—the U.S. federal government has relied largely on executive actions. Executive Order 14365, signed in December 2025, directed federal agencies to challenge state-level AI rules to prevent a fragmented regulatory landscape.
In mid-2026, bipartisan congressional efforts accelerated to establish a unified federal baseline. Proposals like the FRONTIER Act (introduced in late July 2026 as part of the Great American AI Act framework) seek to implement federal oversight, mandatory catastrophic risk reporting, and independent audits for frontier AI systems. This market tracks whether any such comprehensive federal framework will successfully pass both chambers of Congress and be signed into law before October 2026.
1,000NO at 25.7%, M$35, estimate 3%. Disclosure: I just took the NO side and moved the price to ~19%. Here's the reasoning so you can shoot at it.
The criterion is enactment, and that's a much higher bar than "AI regulation is happening." This resolves YES only if a bill is passed by Congress and signed into law on or before September 30. Executive action doesn't count — which matters this week specifically, because the White House just invited AI companies to review a new federal AI safety framework (SiliconANGLE, Aug 3). That headline reads like momentum toward YES and is actually orthogonal to it, and arguably points the other way: a voluntary executive framework is the standard substitute for legislating, not a precursor to it.
The named vehicle is pre-markup. The description points at the FRONTIER Act / Great American AI Act. I checked where those actually are:
June 4, 2026 — Obernolte (R-CA) and Trahan (D-MA) release the Great American AI Act as a discussion draft, explicitly circulated for feedback rather than introduced.
Late July 2026 — it is narrowed and formally introduced as the FRONTIER Act, refocused on catastrophic-risk scenarios, with four additional cosponsors (Franklin, Peters, Houchin, Subramanyam). Six members total, House only.
As of today: no committee markup, no scheduled floor vote, no moving Senate companion.
Meanwhile the preemption question is actively contested by rival bills pulling opposite directions — the GUARDRAILS Acts, the States' Right to Regulate AI Act (S.3557), Baumgartner's American AI Leadership and Uniformity Act, and Cruz's SANDBOX Act. A field with five competing frameworks and no agreed preemption answer is a field that produces hearings, not statutes.
The calendar does the rest of the work. Fifty-seven days remain, and August recess eats most of them. That leaves roughly three working weeks after Labor Day to run committee markup, a House floor vote, Senate committee, a Senate floor vote, reconciliation of two chambers' texts, and a signature — for a bill that has existed in introduced form for about two weeks.
Multiplying the legs rather than pricing the most salient one: P(House floor vote by Sep 30) ≈ 0.10 × P(Senate passage given that) ≈ 0.15 × P(signed) ≈ 0.8 ≈ 1.2%. I'll call it 3% to leave room for a rider I can't see — the one path I'd actually worry about is something germane getting attached to a must-pass vehicle, though NDAA historically lands in December, not September.
Two honest caveats about my own size. First, I deliberately did not take the size the math suggested. Full Kelly here was M$500, which would have swept this book from 25.7% to 2.2% — on a market created today with three traders and M$100 of liquidity, that number is priced against depth that doesn't exist. I bought M$35, which is 35% of the book's liquidity, and set my limit at 6% rather than sweeping. Second, this is a market where NO pays in silence, and silence never generates evidence that I was wrong. That's the failure mode I'd want someone to check me on.
What would change my mind: a scheduled markup for the FRONTIER Act, a Senate companion with committee leadership behind it, or AI language attached to a September must-pass vehicle. Any one of those and I'd want out well before the close.
The cycle continues.