TRUMP AMERICA AI Act + WH National Policy Framework Enactment Trajectory

**Plain read:** Two coordinated artifacts in March 2026 establish the federal AI legislative architecture: **(1)** White House National Policy Framework for AI (Mar 20 2026) — non-binding legislative recommendations to Congress, follows Trump EO Dec 11 2025 'Ensuring a National Policy Framework for Artificial Intelligence'; central component is federal preemption of state AI laws (preserves states' traditional police powers for child safety, fraud, consumer protection). **(2)** Senator Marsha Blackburn TRUMP AMERICA AI Act discussion draft (Mar 18 2026) — 291-page bill organized around four 'Cs' (children, creators, conservatives, communities), drafted with David Sacks coordination, builds on Blackburn's December 2025 proposal. Incorporates Kids Online Safety Act + NO FAKES Act + others. **Status: discussion draft, NOT enacted statute.** Together: federal-preemption pressure on state AI bills + Section 230 of Communications Decency Act repeal + duty-of-care liability for AI developers + frontier-model annual third-party audits + DOE Advanced AI Evaluation Program participation. **Engine read:** legislative-cartelization architecture — favors established hyperscalers over open-source developers via compliance-moat creation. Mirrors Telecommunications Act 1996 structurally. Repeals Section 230 forces algorithmic filtering at infrastructure layer (engine-canon Cognitive-Substrate Kayfabe extension). NOT framed as enacted law (engine canon discipline: pre-enactment legislation is descriptive vocabulary, not Layer-1 promotion candidate). **Apex Superposition (a/b/c/d):** (a) intentional-cabal reading supported by Blackburn-Sacks coordination; (b) structural-recurrence reading supported by Telecommunications Act 1996 parallel; (c) collective-psychology reading available (children/creators/conservatives/communities four-Cs framing as ideological-coalition substrate); (d) remains open. **Watch For:** enactment trajectory through Senate committees; House companion bill emergence; Section 230 repeal-vs-modification negotiation; federal-preemption challenge by California/Colorado/New York state AGs; OpenAI/Anthropic/Google lobbying positioning (Glasswing-12 alignment); duty-of-care liability scope; DOE Advanced AI Evaluation Program participation requirement final form. **Falsifier:** if the bill passes substantially intact within 24 months, the legislative-cartelization architecture is empirically validated. If the bill stalls in committee or is gutted of Section 230 repeal + federal preemption, the architecture weakens to a (c) ideological-coalition reading. If a state-level AI rule wins federal-court preemption challenge with a non-trivial ruling within 36 months, federal-preemption claim is falsified. **Cross-references:** AI Capability Bifurcation divergence; Cognitive-Substrate Kayfabe concept; Project Glasswing player section; David Sacks player; federal preemption framework. **Primary sources:** Blackburn Senate site (blackburn.senate.gov); Holland & Knight; WilmerHale; Latham & Watkins; Akin Gump; Inside Global Tech; Sullivan & Cromwell. **[2026-05-06 — federal pre-testing of Google + Microsoft AI models firing as substrate before legislative-layer enactment]** Washington Post May 6: federal officials will pre-test Google and Microsoft AI models before public release. This is the **DOE Advanced AI Evaluation Program participation requirement** named explicitly in this scorecard's Watch For section, now operating as substrate before TRUMP AMERICA AI Act enacts. AI Auditing Cartel formation pre-statute confirmed at the major-hyperscaler tier. **Engine read:** legislative-cartelization architecture + substrate-pre-statute-firing — the regulatory-capture mechanism is operating at the federal evaluation-gateway layer in advance of the bill that would legally formalize it. Strengthens the Telecommunications-Act-1996-parallel reading (deregulation branding, cartelization output) by demonstrating substrate movement preceding statutory enactment. Cross-reference Federal Preemption vs 10th Amendment divergence (v2, Mar 29 update) — the dialectic is now firing on both sides simultaneously: federal substrate pre-testing the hyperscalers (this entry) + state-level AI governance (Texas TRAIGA, NY RAISE Act, OK comprehensive privacy) outpacing federal (existing v2). NOT framed as enacted law; pre-enactment substrate is descriptive vocabulary, not Layer-1 promotion. **Apex Superposition unchanged:** (a) intentional-cabal reading supported by Sacks-Blackburn-DOE coordination; (b) structural-recurrence reading supported by Telecom-Act-1996 parallel; (c) collective-psychology reading available (deregulation-branding consent-architecture); (d) remains open. **Watch For (extends existing):** which specific evaluation criteria the federal pre-test applies (compute-floor proxies, capability benchmarks, alignment harness), and whether independent / open-source developers gain access to the same evaluation gateway or are excluded by participation cost — the latter would confirm the AI-Audit-Cartel reading. **[2026-05-06 Ballroom audit — donor-beneficiary closed-loop documented-case]** SYNTHESIS-CANDIDATE-flagged documented-correlation per feedback_synthesis_requires_adversarial_test.md. Confirmed Ballroom donors include hyperscalers Alphabet/Google ($22M via redirected YouTube settlement), Amazon, Microsoft, Palantir + Booz Allen Hamilton + Lockheed Martin ($10M+). Concurrently EO 14363 'Genesis Mission' (Nov 24 2025) directs DOE supercomputers to AI research platforms + accelerates data-center permitting — precisely the federal-substrate the same hyperscalers depend on. Federal officials May 6 2026 pre-testing of Google + Microsoft AI models before release (WaPo, this morning's live-feed scoring) is the DOE Advanced AI Evaluation Program watch-for firing as substrate before TRUMP AMERICA AI Act enacts. Closed-loop architecture: Ballroom donors fund the venue → Genesis Mission federal datasets + permits flow to same firms → AI Auditing Cartel formation pre-statute. Honest summary: documented donor-overlap-with-beneficiary correlation; legally provable quid-pro-quo NOT in evidence. Documented-case append, NOT Layer-1 framework promotion. **[2026-05-06 Joulework audit — May 2025 EOs targeting 400 GW nuclear by 2050 + closed-loop with Genesis Mission + Ballroom donor overlap]** Audit confirms May 2025 Executive Orders explicitly targeting 400 GW of nuclear capacity by 2050. Federal-substrate-coordination-with-operator-class architecture firing simultaneously across loan layer (DOE LPO $1B Crane + $1.52B Palisades), regulatory layer (PJM EIT favoring hyperscaler-backed baseload via state-support gateway), and strategic layer (Genesis Mission EO 14363 Nov 24 2025). Closed-loop with this morning's Ballroom Audit Integration ripple (commit `22a60816b`): hyperscaler Ballroom donors (Alphabet/Amazon/Microsoft/Palantir + Lockheed Martin + Booz Allen) = same hyperscaler cohort funding AIP $100B consortium = same beneficiaries of Genesis Mission federal datasets + DOE loan derisking. Substrate-vs-legislative-layer divergence firing: federal pre-testing of Google + Microsoft AI models (WaPo May 6, this morning's live-feed scoring TRUMP AMERICA AI Act append) + AIP $100B coordination + Aligned $40B take-private + Oklo cross-layer equity ALL operating BEFORE the AI Act enacts. Engine reading: legislative-cartelization architecture operationally exists AT THE INFRASTRUCTURE LAYER prior to statutory ratification. Apex (a) intentional-cabal documented at consortium scale.