OpenAI has proposed giving the U.S. government a 5% stake in the company, and CEO Sam Altman wants the same arrangement applied to every leading American AI developer. The Financial Times reported Thursday, citing two people familiar with the talks, that Altman raised the concept in early discussions with the Trump administration as a way to defuse mounting political pressure on the industry.
The stakes here are structural, not symbolic. A 5% OpenAI stake would be worth roughly $42.6 billion at the company’s post-money valuation of $852 billion, set after its record funding round closed in March. Under Altman’s broader pitch, Anthropic, Google and Meta would cede similar 5% stakes to a government vehicle modeled on a sovereign wealth fund, giving Washington a direct financial claim on the sector’s biggest players rather than just regulatory authority over them.
Government equity in frontier labs would be a new category of leverage. Regulators can slow a company down with rules and reviews. An equity stake gives the government a seat at the table on valuation, governance and dividends, entangling the state’s financial interest with the industry’s growth. That distinction, ownership versus oversight, is what makes this proposal different from the export controls and safety reviews that have defined AI policy so far this year.
It is not clear whether Anthropic, Google or Meta would agree to Altman’s version of the arrangement. The FT said the White House, OpenAI, Anthropic, Google and Meta did not immediately respond to CNBC’s requests for comment, and CNBC has not confirmed independently that a term sheet or formal proposal exists beyond the framework described by the FT’s sources.
The timing lines up with a stretch of friction between AI labs and Washington. Anthropic disabled access to its Mythos and Fable models last month to comply with a government export control directive, then said Tuesday it had resolved policymakers’ safety concerns and would restore access. Regulators have grown more focused on cybersecurity vulnerabilities in frontier models and on the competitive threat from Chinese open-source models that are closing the capability gap at a fraction of the cost.
Altman has been building toward this proposal for more than a year. CNBC reported last month that he first pitched the idea of a government stake to the Trump administration in early 2025, and in April OpenAI proposed a “public wealth fund” to hold assets tied to AI company growth and distribute the economic upside to the public. The current pitch narrows that concept into a specific number: 5% per company, structured through a government-held vehicle.
The precedent for this kind of deal already exists. The Trump administration took a 10% stake in Intel last August after an $8.9 billion investment in the chipmaker’s common stock, and has since taken positions in IBM and several quantum and critical mineral companies. Trump said in May he should have negotiated for a larger share of Intel, a comment that raises the stakes for any AI company negotiating a fixed percentage now rather than later. He has separately called government ownership stakes in AI companies “a beautiful thing” that would make Americans “partners in this revolution.”
The open question is whether the other three labs see this as a fair trade or a precedent they would rather avoid setting. If OpenAI moves alone, it gains political cover but hands the government a permanent economic interest in its outcomes that Anthropic, Google and Meta would not share, an asymmetry likely to shape how each company responds to Washington’s next round of AI-specific pressure.
Reported by CNBC on July 2, 2026, citing the Financial Times.
