As of mid-2026 there is no US federal AI floor coming on a plannable timeline (the White House National Policy Framework for AI of 20 March 2026 is explicitly non-binding and would require Congressional action, and the proposed federal moratorium on state AI laws was not enacted), and the most-watched comprehensive state law retreated rather than advanced (Colorado's SB 26-189, signed 14 May 2026, repealed and reenacted the Colorado AI Act, removing the algorithmic-discrimination duty of care and the risk-management and impact-assessment obligations and moving the effective date from 30 June 2026 to 1 January 2027), so the operative reality is a non-converging state patchwork, and the defensible enterprise posture is to build to the strictest obligation that actually applies to its own deployments and treat the regulatory map as a moving input rather than waiting for a federal floor.

Anchored on: Colorado SB 26-189 repeal-and-reenactment signed 14 May 2026 (Crowell & Moring, Hunton, Troutman, Akin), narrowing the law to disclosure/transparency and delaying it to 1 January 2027 (prior delay from 1 Feb 2026 to 30 Jun 2026 via SB 25B-004, Aug 2025); the White House National Policy Framework for Artificial Intelligence of 20 March 2026, explicitly non-binding and requiring Congress (Holland & Knight); the non-enactment of the proposed federal moratorium on state AI laws (the Senate had earlier removed a ten-year moratorium from the 2025 budget reconciliation bill); and the two broad state laws effective 1 January 2026 — California SB 53 (Transparency in Frontier AI Act), applying only to frontier developers above ~10^26 training operations and developers with >$500M revenue (Brookings), and Texas's Responsible AI Governance Act, broad in reach but intent-based (Norton Rose Fulbright), plus King & Spalding on the new state laws and the executive order signalling preemption intent. Claim is scoped to the planning posture (no plannable federal floor; flagship state law retreated; build to applicable obligations and track the moving map), not to a prediction of the direction of any single statute, and explicitly frames the picture as divergence rather than deregulation (the EU did not retreat; the failed moratorium preserves state authority). The exact Senate vote count on the earlier moratorium is intentionally left qualitative ('removed... near the budget bill') to avoid asserting a figure not directly cited here. VERIFIED 2026-06-02 via Crowell & Moring (SB 26-189 scope changes + 1 Jan 2027 date), Holland & Knight (framework non-binding, Congress required, moratorium not pursued), Brookings (SB 53 frontier-only thresholds), Norton Rose Fulbright (TRAIGA broad + intent-based + 1 Jan 2026), and King & Spalding (state laws effective 1 Jan 2026 + EO signal). 90-day review cadence (31 Aug 2026). Trigger conditions: (1) Congress enacts federal AI legislation with preemptive effect, moving the claim toward Partial or Not holding; (2) Colorado's reenacted law is amended again before 1 January 2027; (3) a major state passes a comprehensive law on the original Colorado model, reversing the retreat reading; (4) the White House framework is replaced by a binding instrument. Siblings: AM-184 (/eu-ai-act-digital-omnibus-what-still-applies/, the regime that delayed but did not retreat), AM-172 (/ai-governance-data-governance-us-frameworks/, the US sector-frameworks companion), and the operators version OPS-090 (/operators/us-ai-laws-small-business/).