Trump Eyes U.S. Government Stake in AI to Win the Global AI Race

Washington is pursuing direct equity positions in frontier AI companies alongside a sweeping national-security directive, marking the most aggressive assertion of U.S. federal power over artificial intelligence in the country's history.

Lead

President Donald Trump signed National Security Presidential Memorandum-11 (NSPM-11) on June 5, 2026, ordering the U.S. military and intelligence community to absorb frontier artificial intelligence at wartime speed — and simultaneously confirmed that his administration is in advanced negotiations to acquire a direct government equity stake in leading AI developers, including OpenAI. The twin moves represent the sharpest pivot in U.S. technology governance since the Cold War nuclear buildup, fusing Trump AI policy with a bet that Washington must own a share of the infrastructure powering the AI global race.

What Happened

NSPM-11 revokes the Biden administration's National Security Memorandum-25, which had imposed a more cautious governance framework on AI within the national security enterprise. In its place, the Trump directive sets four binding pillars — Adoption, Adaptation, Assurance, and Accountability — each backed by explicit deadlines.

Within 90 days, the Secretary of Defense must update Pentagon Directive 3000.09 governing autonomous weapons systems, with mandatory annual reviews thereafter. Within 120 days, the Defense Department, the Director of National Intelligence, and all agencies with intelligence community elements must overhaul procurement processes to allow rapid onboarding of advanced AI models from multiple commercial vendors.

The memorandum also empowers agencies to terminate contracts with AI companies that repeatedly restrict government use of their systems — a provision widely interpreted as a direct response to legal friction between the Pentagon and Anthropic over deployment guardrails. A classified annex is due within 90 days.

Alongside the operational directives, NSPM-11 establishes an AI National Security Strategic Reserve, a standing roster of top non-governmental AI experts the government can mobilize in strategic crises, modeled loosely on reserve structures used in nuclear and cybersecurity domains.

The Equity Stake Question

Concurrent with NSPM-11, the White House confirmed it is in discussions with OpenAI about a possible U.S. government AI stake in the company. The proposed structure is novel: rather than a taxpayer-funded acquisition, OpenAI would donate equity to the federal government to seed what the company calls a Public Wealth Fund — an entity outlined in OpenAI's April 2026 policy proposal designed to "invest in diversified, long-term assets" and allow citizens to participate directly in AI's economic upside. OpenAI is currently valued at more than $850 billion by private investors and is preparing for an initial public offering as early as this year.

Trump framed the rationale publicly: "There are concepts where pieces could be given to the American public, where the American public essentially becomes a partner." The model reflects a populist arithmetic — if national security tech generates historic wealth while displacing workers, the state wants a claim on the upside rather than absorbing only the downside.

No final terms have been agreed, and the structure remains subject to change.

Precedent Already Set

The conversations with OpenAI are not without precedent. The Trump administration converted a portion of its CHIPS Act commitments into a 9.9% equity stake in Intel, an $8.9 billion position, during the president's second term. The government has since taken direct equity positions in at least 10 companies spanning semiconductor, quantum computing, and critical minerals — asset classes Washington views as strategic. An AI equity stake would extend that logic to the software layer of the national security tech stack.

Strategic Context: The AI Global Race

The underlying driver of both NSPM-11 and the equity talks is competitive urgency. U.S. defense and intelligence officials have characterized the AI global race — principally against China — as a contest that will determine military dominance, intelligence superiority, and economic leverage for decades. China's government-directed AI ecosystem integrates commercial and state actors at a speed and scale that U.S. officials argue American procurement processes were never designed to match.

NSPM-11 directly targets that gap. By requiring multi-vendor AI procurement, establishing the Strategic Reserve, mandating high-security AI computing infrastructure, and removing contractual friction that slows deployment, the administration is attempting to compress into months what traditional acquisition timelines measure in years.

The equity stake dimension adds another lever: if Washington holds financial stakes in frontier AI developers, it gains both an information advantage — visibility into roadmaps and capabilities — and a structural alignment of incentives between corporate success and national interest.

Geopolitical Dimension

The memorandum's reach extends beyond U.S. borders. Allied intelligence-sharing frameworks — particularly the Five Eyes partnership — will need to accommodate new standards for AI-processed intelligence. The requirement to terminate contracts with companies that restrict government use could reshape how allied defense ministries procure the same American AI platforms.

Export control policy, already restructured under the Trump administration's Diffusion Rule revisions, interacts directly with NSPM-11: the same frontier models now being fast-tracked for Pentagon deployment are the ones subject to controls limiting foreign access. The dual-track logic — accelerate at home, restrict abroad — is the defining architecture of U.S. Trump AI policy in this period.

Outlook

NSPM-11 sets a compressed timeline. The 90- and 120-day clocks running from June 5 will produce measurable outputs — updated weapons-autonomy doctrine, reformed procurement rules, and a classified strategic annex — before the end of the third quarter of 2026. The equity discussions with OpenAI and, potentially, other frontier developers are likely to intensify ahead of OpenAI's anticipated IPO, as the structural window for a donated equity arrangement narrows once the company is publicly listed. Whether the Public Wealth Fund concept survives political and regulatory scrutiny remains the central open question — but the direction of U.S. government AI stake policy is now clearly set toward direct federal participation in the AI industry, not merely regulation of it.