# AI Action Plan **"Winning the Race: America's AI Action Plan"** (July 23, 2025) is the second Trump administration's governing document for artificial intelligence — roughly ninety policy actions organized under three pillars, produced under AI-and-crypto czar [[David Sacks]], OSTP director [[Michael Kratsios]] and senior policy adviser [[Sriram Krishnan]]. Its premise is a race with [[China]] that the United States wins by building faster and regulating less; its method is to treat safety rules, state laws and "ideological bias" as the obstacles. Everything the administration has done on AI since — three executive orders the day of release, the December 2025 assault on state AI laws, the June 2026 frontier-model order — is an implementation of it. ## The three pillars 1. **Accelerate innovation** — cut federal rules that "hinder" development, review enforcement actions that slow deployment, favor open-weight models, create regulatory sandboxes, fund worker retraining. 2. **Build infrastructure** — streamline permitting for data centers, chip fabs and the energy to run them; build high-security data centers for defense and intelligence; harden critical infrastructure and create an AI incident-response ecosystem. This is where the plan meets [[Chris Wright]]'s grid orders and [[Stargate]]. 3. **Lead internationally** — export a "full-stack" American AI package to allies while tightening [[Chip Export Controls]] on rivals. The plan also told agencies to weigh a state's "burdensome AI regulations" when awarding discretionary funds — the seed of the preemption fight. ## The orders that followed - **July 23, 2025** — three EOs signed with the plan: *Preventing Woke AI in the Federal Government* (procurement limited to models judged "truth-seeking" and "ideologically neutral," the culture-war provision — see [[Algorithmic Bias]]); *Accelerating Federal Permitting of Data Center Infrastructure*; *Promoting the Export of the American AI Technology Stack*. - **December 11, 2025** — *Ensuring a National Policy Framework for Artificial Intelligence*: a DOJ **AI Litigation Task Force** to sue states over AI laws on Commerce Clause and First Amendment grounds; a Commerce list, due in 90 days, of "onerous" state laws (Colorado's algorithmic-discrimination act and the California, Texas and Utah omnibus laws were named); BEAD broadband funds and other grants conditioned on state compliance; FCC and FTC proceedings toward federal standards; and a request that Congress preempt state AI law outright, with carve-outs for child safety, data-center permitting and state procurement. See [[Federal Preemption]]. - **June 2, 2026** — *Promoting Advanced Artificial Intelligence Innovation and Security*: asks frontier labs to submit new models for *voluntary* government testing 30 days before release (cut from a proposed 90 after Sacks and industry objected), directs DOJ to prioritize AI-assisted hacking, and states in terms that nothing in it authorizes "mandatory governmental licensing, preclearance, or permitting." The order that was supposed to be about safety became a statute-like guarantee against it. ## Why it matters The plan settles a fight inside the right. The populist wing — [[Steve Bannon]], parts of the Senate GOP, the state attorneys general who killed the July 2025 moratorium on state AI laws 99–1 — wanted to constrain the labs; the tech wing, in the White House, wanted the labs unconstrained and the states out of the way. The plan is the tech wing's charter, and the "woke AI" order is the concession that lets the populists claim a win. It also makes AI energy policy: the pillar-two commitments are what justify keeping coal plants open and fast-tracking gas for data centers. ## See Also - [[David Sacks]] · [[Michael Kratsios]] · [[Sriram Krishnan]] · [[Donald Trump]] - [[Stargate]] · [[Chip Export Controls]] · [[Federal Preemption]] · [[Algorithmic Bias]] - [[Chris Wright]] · [[Department of Energy]] · [[China]]