OpenAI is trying to turn public ownership of AI from a threat into a negotiated gift, and Washington should be careful about what it accepts.
You don't often see a private company offer the government free stock. That is essentially what OpenAI is discussing with the Trump administration, according to CNBC and a Financial Times report published this week. The talks center on a possible equity donation of up to 5%, which would be worth more than $42 billion if measured against OpenAI's reported $852 billion private valuation.
This isn't a bailout. No taxpayer money buys in, OpenAI doesn't get a rescue check, and the company would keep control of its board. The mechanism being discussed is a donation, which lets the White House say the public captured some of the AI boom's upside without having to write a check for it. Trump has already talked about the idea aboard Air Force One, telling reporters that there are concepts where pieces could be given to the American public and the public essentially becomes a partner.
The timing tells you why OpenAI is interested. Sam Altman first raised the idea with Trump in early 2025, and OpenAI later put a version of it into an April policy paper proposing a Public Wealth Fund that could let ordinary citizens share in AI-driven growth through direct account payouts or dividends. CNBC reported on June 5 that talks were still moving with no final deal. That detail matters. This is not a signed agreement hiding inside a press release. It is a live negotiation.
Here's the political backdrop. Senator Bernie Sanders introduced the American AI Sovereign Wealth Fund Act in June, a far more aggressive proposal that would require OpenAI, Anthropic and other large AI companies to hand 50% of their equity to a public fund paid in shares. Axios reported that the bill would apply to companies generating at least $200 million a year from AI, and Tom's Hardware noted that the fund could carry voting rights and board influence.
Next to that, a voluntary stake of up to 5% with no governance strings looks less like generosity and more like OpenAI setting the opening price before Congress or the White House sets one for it.
Anthropic is not playing the same game, at least not publicly. The Financial Times reported that OpenAI's proposal could become part of a wider model for American AI companies, but it is not clear that other labs would accept the same terms. Anthropic has said it is not offering equity to the administration, while floating a broader version of the idea: sovereign wealth funds that hold stakes across the AI industry rather than one company handing Washington a slice of itself.
That difference is not small. A diversified fund behaves differently from a government stake in one lab. Norway's sovereign wealth fund can own small pieces of thousands of companies without becoming dependent on any single one. A fund built around donated OpenAI shares would be much harder to keep at arm's length, especially if the company later goes public or becomes more deeply tied to federal contracts.
The legal path is also unfinished. The Financial Times reported that the discussions are at an early stage and may require congressional approval. There is no normal route for a private AI company to simply transfer equity to the federal government outside a formal investment or statutory framework. Frankly, that is the part worth watching more closely than the headline percentage.
A government stake changes incentives in ways nobody involved is eager to say plainly. Once Washington owns part of OpenAI, it has a financial reason to protect that investment. That could make safety regulation harder, not easier. It could also create pressure to intervene if the company stumbles, because a private corporate problem would suddenly sit partly on the public balance sheet.
You can see why the offer is attractive. OpenAI gets political goodwill at a moment when it is already working closely with the federal government and preparing for the public-market life of a public benefit corporation. The administration gets to tell voters that AI wealth will not sit only with founders, employees and investors. Everyone gets a line they can sell.
But public ownership is not automatically public accountability. If Washington takes equity without voting power, oversight rules, conflict safeguards and a clear answer on who receives the upside, the deal risks becoming political cover for both sides. OpenAI can say it shared the wealth. The government can say it captured the future. The public may still have no meaningful control over either.
This is still a negotiation, and the final number could move before anyone signs anything. The precedent is already visible. If OpenAI gives up equity for political goodwill, every other AI lab raising money at enormous valuations will eventually hear the same question from Washington: what piece of this boom belongs to the public?
Also read: SoftBank pushes for a full $10 billion loan against its OpenAI stake • Robinhood Stops Renting Blockchain Rails and Builds Its Own • ByteDance Is Building Its Biggest Data Center Outside China in Brazil
