Senior Trump administration officials have been weighing whether the federal government should take equity stakes in major artificial intelligence companies, according to people familiar with the discussions. The idea was under review even before the White House’s latest export restrictions on Anthropic added fresh tension to the AI sector.
The conversations remain early and no decision has been made. But the fact that they reached senior Cabinet officials signals that the administration is at least exploring a more interventionist role in the industry than many companies had expected.
According to people familiar with the talks, Treasury Secretary Scott Bessent and Commerce Secretary Howard Lutnick have floated different paths for any government-held equity. Bessent has favored directing the value toward Trump Accounts, while Lutnick has preferred placing it in a sovereign wealth-style fund.
The proposal is still a difficult one for much of the AI industry. OpenAI has previously advanced a version of the idea, but many other major companies have been far more skeptical. Executives at firms including Microsoft and Meta have recently signaled little enthusiasm for the concept.
The debate comes as the White House prepares for a possible meeting with AI leaders that President Donald Trump said would happen soon, with the stated goal of discussing how the industry can “give back something to the public.” But the administration’s decision to impose export controls on Anthropic’s latest models could make any such meeting more combative.
Anthropic’s restrictions have also complicated the broader policy environment around AI. Trump’s push for a public contribution from AI firms now overlaps with a crackdown on one of the sector’s most prominent players, raising questions about whether the administration is using economic leverage as part of its strategy.
Industry reaction has been mixed to negative. Caleb Max, president and chief executive of the National Artificial Intelligence Association, said before the Anthropic move that what AI companies need is clear federal rules and safeguards, not more cash infusion. Other observers doubt that companies would voluntarily agree to a structure that could amount to handing over a slice of future upside to the government.
The legal and financial hurdles are substantial. AI firms are valued much higher than companies that have previously taken government support in exchange for stakes, which would make any public equity position expensive. Some critics also question how such a deal would work in practice, given that many AI companies are not yet paying out profits in the form of dividends.
On Capitol Hill, the idea is not getting an easy reception. Sen. Cynthia Lummis, a Wyoming Republican, called it hard to understand. Sen. John Kennedy of Louisiana also voiced distrust of AI executives during a recent Senate hearing.
The administration has not commented publicly on the discussions, and neither OpenAI nor Anthropic responded to requests for comment, according to the report.
The concept of a government stake in AI companies fits into a broader pattern of Trump-era efforts to extract public or strategic value from private-sector deals. Lutnick has also been involved in discussions over government stakes in companies tied to critical mineral supply chains, and has floated the idea of a national and economic security fund.
For now, though, the proposal appears to be more of an open question than a policy in motion. The administration is still weighing whether to pursue the idea, how to structure it, and whether the industry would ever agree.