Sam Altman has offered the U.S. government a 5% stake in OpenAI, worth roughly $42.6 billion against the company’s $852 billion March post-money valuation, according to a Financial Times report Thursday citing two people familiar with the talks. The pitch reportedly extends to similar 5% slices of Anthropic, Google, and Meta, bundled into a sovereign-wealth vehicle designed, in the FT’s framing, to “secure good relations with the administration and address political blowback.”
The number isn’t arbitrary. It’s precisely half of the 10% the Trump administration took in Intel last August via an $8.9 billion common-stock investment, and it arrives after Trump said in May that he should’ve asked for more of Intel. It’s also a fraction of the 50% Steve Bannon has publicly demanded the government extract from AI firms. Bannon called the 5% offer “tip money.”
That framing captures the bind. OpenAI is trying to price political risk. The MAGA right wants a larger cut; the administration wants a win it can narrate to voters; and Altman wants the whole conversation resolved before the confidential S-1 OpenAI filed with the SEC in June becomes a public roadshow. Anthropic filed its own confidential S-1. Both companies are trying to IPO into an environment where the president has already told reporters that government ownership of AI firms would be “a beautiful thing,” making Americans “partners in this revolution.”
None of this is new for OpenAI’s policy shop, which published a paper in April 2025 calling for a “Public Wealth Fund” that would give citizens “a stake in AI-driven economic growth.” Altman first pitched the sovereign-wealth concept to the Trump administration in early 2025. What’s changed is the machinery around him: the American AI Sovereign Wealth Fund Act, introduced in June, would pull in firms above a $200 million annual AI-revenue threshold, and Trump confirmed the same month that he’d discussed “concepts where pieces could be given to the American public.” Anthropic has floated its own alternative, a “digital dividend.”
The backdrop is coercive in ways the equity talks don’t quite acknowledge. Last month Anthropic disabled access to its Mythos and Fable models to comply with an export-control directive; on Tuesday it said policymakers’ safety concerns had been resolved and access would be restored. The sequencing is the story. A regulator turns off a product, a regulator turns it back on, and the largest labs are simultaneously offering the same regulator a permanent equity position.
Any deal likely requires congressional approval, and safety researchers have already flagged the obvious conflict: a government that owns 5% of the frontier is a government with a fiduciary reason to soften its own rules. The 2008 TARP framing had the opposite valence, equity taken as the price of a rescue. Here the labs are offering it as the price of admission.
Sources
- https://techcrunch.com/2026/07/02/openai-proposed-donating-5-of-its-equity-to-a-us-sovereign-wealth-fund/
- https://www.cnbc.com/2026/07/02/openai-proposes-us-government-own-5percent-stake-to-address-political-blowback.html
- https://www.bloomberg.com/news/articles/2026-07-02/openai-proposes-giving-the-us-government-a-5-stake-ft-says
- https://www.reuters.com/technology/artificial-intelligence/openai-proposes-handing-trump-administration-5-stake-ft-reports-2026-07-02/
- https://www.techtimes.com/articles/319571/20260702/openai-offers-washington-426-billion-stake-experts-warn-safety-rules-face-conflict-risk.htm