OpenAI has proposed giving the US government an equity stake of around 5%, the Financial Times reported on 2 July 2026, citing two people familiar with the discussions. At OpenAI's reported $852 billion valuation — set in a record funding round in March 2026 — a 5% holding would be worth roughly $42.6 billion, which would make the US government one of the company's largest shareholders if it happened. The discussions are described as conceptual and early; by the FT's account they may require an act of Congress, and it is not clear whether the administration will accept. CEO Sam Altman has argued that giving the public a direct financial interest is the fairest way to spread the gains of AI, and has raised the idea with President Trump, Treasury Secretary Scott Bessent and Commerce Secretary Howard Lutnick.

What is actually being discussed

The mechanism is still unsettled. FT-relayed reporting describes a proposal for the US government to receive around 5% of OpenAI's equity, potentially through a public investment vehicle modelled on Alaska's Permanent Fund — the state-level fund, financed historically by oil revenue, that pays Alaskan residents an annual dividend. It is a state model, not an existing federal one. OpenAI's own policy paper, "Industrial Policy for the Intelligence Age," separately argues for a Public Wealth Fund that would give citizens a stake in AI-driven economic growth; but that paper does not itself create a deal or specify the final mechanics of a government stake in OpenAI. The exact legal route — whether a stake would involve reserved, issued or donated equity — and any extension to other leading US AI developers remain open questions. The safest reading is that the equity structure, the approval route and the scope are all unsettled.

The conversations are not new. By CNBC's account, Altman first raised the concept with the administration in early 2025 and has intensified the discussions in recent weeks, against a backdrop of growing political pressure on American AI firms over security concerns and competition from cheaper Chinese models. Trump has spoken about the idea favourably in general terms, describing a government stake as something close to a partnership with the American public.

Washington is already in the equity business

The proposal lands in a Washington already experimenting with equity and equity-like tools. In 2025, the US government made an $8.9 billion investment in Intel common stock, widely reported as roughly a 9.9% to 10% stake and structured as a passive, nonvoting position. Nvidia and AMD separately agreed to give the US government 15% of the revenue from certain AI-chip sales to China. And President Trump signed a February 2025 executive order directing officials to develop a plan for a US sovereign wealth fund. Forbes has tallied more than $26 billion in equity and equity-like positions taken by the government across around thirty deals since the start of 2025, spanning semiconductors, rare earths and quantum computing. Those precedents do not make an OpenAI stake routine, but they make it less isolated than it would have sounded a few years ago.

The conflict at the centre

The reason the proposal is more than a financing curiosity is the position it would put the government in. A state that helps regulate the AI labs and a state that owns equity in them are not the same actor, and combining the two creates a structural tension: a government with a financial stake in a lab has an incentive to see that lab succeed, which sits awkwardly beside the roles that various parts of the government — the White House, Commerce, Treasury, competition and export-control authorities, and Congress — play in setting safety rules, export controls and antitrust policy for the same company. Supporters of the idea argue this is manageable if the stake is passive and nonvoting, as the Intel position was — ownership of the upside without control of the company — and that letting the public share in AI's returns is a reasonable answer to concerns about who captures the value. Critics see equity in a private company as, in their words, nationalisation by another name, and worry that officials holding shares will pull their punches on a firm the government co-owns. Both concerns are real, and which one dominates depends entirely on the structure that a final deal — if there is one — actually takes.

Why it matters beyond OpenAI

If the model spread across the sector, it would change the environment for everyone building on these platforms, not just for OpenAI, because a government invested in several labs at once has a different relationship to the whole industry than one that only regulates it. There is also a sovereignty dimension that reaches outside the US: analysts have noted that a US government stake could prompt other jurisdictions to demand similar arrangements with their own strategic providers, and could make it harder for firms and governments elsewhere to treat American AI providers as neutral infrastructure. And the equity route is not the only one on the table — Senator Bernie Sanders has proposed a very different mechanism, a one-time tax on the largest AI firms' stock paid into a public fund, which would achieve a form of public ownership without the government becoming a negotiating shareholder. The gap between those two approaches is a preview of the debate that any serious version of this idea would trigger.

Key Takeaways

  • The Financial Times reported on 2 July 2026 that OpenAI has proposed giving the US government an equity stake of around 5% — worth roughly $42.6 billion at its reported $852 billion valuation; the talks are described as early and conceptual, may require congressional approval, and it is unclear whether the administration will accept.

  • The mechanism is unsettled — potentially a government-held public investment vehicle modelled on Alaska's (state-level) Permanent Fund, funded by reserved, issued or donated equity. OpenAI's own policy paper separately argues for a Public Wealth Fund to share AI's gains with the public, but does not itself confirm the 5% talks; any extension to other leading US AI developers is an open question.

  • The idea extends an established pattern: in 2025 the US government made an $8.9 billion (~9.9–10%) nonvoting investment in Intel, secured 15% of certain Nvidia and AMD China AI-chip revenues, and — per a Forbes tally — took more than $26 billion across roughly thirty equity deals; Trump signed a February 2025 executive order to plan a US sovereign wealth fund.

  • The central issue is a conflict of interest: a government that owns equity in an AI lab has a financial incentive alongside the roles various agencies play in regulating safety, exports and competition for that same lab. Supporters say a passive, nonvoting stake manages this; critics call it nationalisation. A separate proposal from Senator Bernie Sanders would use a tax on AI-firm stock instead. Nothing is settled.