WASHINGTON: The Trump administration is in active discussions with OpenAI about a government equity stake in the company, President Trump confirmed to reporters aboard Air Force One on June 5, 2026. The talks, which CEO Sam Altman first floated to administration officials in early 2025, are now running in parallel with a separate Senate bill from Bernie Sanders that would force OpenAI, Anthropic and xAI to hand over 50 percent of their stock to a public fund. No terms have been finalized in either track, and sources cautioned that the OpenAI deal may not close.
Trump described the concept in broad strokes. “There are concepts where pieces could be given to the American public, where the American public essentially becomes a partner,” he said, according to CNBC. The talks were first reported by NOTUS on June 4.
How a Voluntary Equity Donation Would Actually Work
The mechanism under discussion is unusual. Rather than the Treasury writing a check to buy OpenAI shares at the company’s roughly $965 billion private valuation, OpenAI would donate equity to seed a federal Public Wealth Fund. The fund concept comes directly from OpenAI itself, which proposed it in an April 2026 policy document. In that paper, OpenAI argued a sovereign fund could invest in long-term assets and let ordinary citizens share in AI’s upside, possibly through direct payments funded by returns.
Altman has been pushing the idea inside the administration for more than a year and revisited it with senior officials this week. The legal pathway is unsettled. There is no standing framework for a private company to transfer equity to the federal government outside of bankruptcy or bailout precedents, and any structure would face questions about valuation, fiduciary duty to existing OpenAI shareholders, and antitrust optics if Washington became a partial owner of one frontier lab while three competitors stayed fully private.
Anthropic publicly confirmed it is not in these conversations, a person familiar with the matter told TechCrunch. xAI has not commented publicly.
What does a government stake in OpenAI mean for founders?
For founders building on or competing with frontier labs, a government stake in OpenAI would tilt the playing field in two directions at once. It would tie OpenAI’s commercial decisions, including pricing and API access, to political oversight and possibly board representation. It would also signal Washington views the largest AI companies as too economically significant to leave entirely in private hands, a category previously reserved for utilities, large banks and defense primes.
The structural shift cuts both ways. A government shareholder could constrain how aggressively OpenAI pursues monetization that hurts smaller developers, but it could also entrench OpenAI’s incumbency by giving the federal government a direct financial interest in its success. Anthropic’s decision to stay out of the talks looks like a bet that independence is worth more than the political cover a public stake might provide.
The Sanders Bill Goes Much Further
Running on a separate track, Sen. Sanders announced the American AI Sovereign Wealth Fund Act in a June op-ed. The bill would impose a one-time 50 percent tax on the equity of OpenAI, Anthropic, xAI and other leading AI firms. The tax would be paid in company stock, not cash. Those shares would seed a public fund, with the federal government taking voting rights and equal representation on each company’s board.
Sanders framed the bill in compensation terms. The models, he wrote, were trained on the public’s collective output, including books, code, journalism, art and conversations, without permission or payment. The fund’s eventual dividends would go to ordinary Americans and fund health care, education and housing. Sanders said he will introduce the bill formally in the coming weeks.
The convergence of a Republican president and the Senate’s most prominent democratic socialist on the principle of public AI ownership, if not the mechanism, is the unusual political story underneath the policy. The Trump-OpenAI track is voluntary, narrow and structured around donated equity. The Sanders track is mandatory, sweeping and built on a tax. Either path would still mark the first time the federal government has taken a direct ownership position in a major tech company in the modern era.
What’s Next
OpenAI has not commented on the equity discussions and the administration has not released a term sheet or timeline. Three things will determine whether either proposal moves from rhetoric to mechanism. First, OpenAI’s IPO path: the company is reportedly preparing a confidential S-1 filing, and any government stake structure would need to be settled before public-market investors set the float. Second, the Sanders bill’s text and committee assignment, which will reveal whether the 50 percent rate and the named companies survive markup. Third, whether xAI joins or rejects the voluntary track, since Anthropic’s refusal already fragments the field.
Founders watching this should treat the Trump-OpenAI talks as a leading indicator of how Washington will engage with the next wave of frontier-lab capital events, including the SpaceX, OpenAI and Anthropic IPOs queued for 2026. For context on Anthropic’s standalone valuation trajectory, see GREY Journal’s coverage of Anthropic’s $900 billion round and the broader policy backdrop in our Trump AI cybersecurity executive order analysis.
