The Regulator Wants to Be a Shareholder

Five percent of OpenAI is no longer a symbolic gesture. If the idea reported by the Financial Times ripens into an actual deal, the United States government could end up holding a direct slice of one of the most important AI companies in the world. According to the report, OpenAI has discussed handing the US administration 5% of its shares, part of an attempt to deal with mounting political pressure over a single question: who benefits from the enormous profits of the AI boom. Reuters carried the report but stressed that it could not independently verify it. So the precise wording has to stay careful. This is a discussion, or a proposal, not a signed deal.

The number itself explains why the story is big. According to the Financial Times, OpenAI is valued at roughly $852 billion, a private valuation of a company that is not publicly traded and has no official market capitalization. A 5% stake at that valuation is worth about $42.6 billion on paper. That is large enough to look like real public participation, and probably small enough not to threaten control of the company. There is still no official explanation for the choice of that specific number, which is why it is worth reading as a political signal at least as much as a financial offer.

The Political Question Behind the Offer

The proposal arrives after months in which the pressure on AI companies moved from academic debate into the political arena. AI companies build their products on broad human knowledge, consume vast amounts of electricity and chips, reshape the labor market, and raise money at valuations once reserved for mature public giants.

As the money grows, it becomes harder to settle for the promise that the technology will benefit everyone someday. The political question gets more direct. If AI generates profit on a historic scale, will the public get a share of it, or only carry the risks?

OpenAI has already sketched one possible answer. In a policy paper published in April, the company proposed creating a Public Wealth Fund, a vehicle that would let citizens benefit directly from the economic growth AI produces. The paper’s own words are that “returns from the Fund could be distributed directly to citizens, allowing more people to participate directly in the upside of AI-driven growth, regardless of their starting wealth or access to capital.” The official document does not commit to transferring 5% of OpenAI itself, but it does carry the view that citizens deserve a share of AI’s economic upside, even if they never invest in the stock market.

Trump, Sanders, and the Same Argument From Two Directions

This public interest is coming from two political poles at once. President Trump said in June that he expects leading AI companies to “give back something to the public,” a phrasing Reuters described as a possible hint at a government stake in the firms. His own words were that he would meet the top executives shortly and that “if we do that, the public will become very rich.”

From the other side, Senator Bernie Sanders put forward a far broader proposal: a sovereign wealth fund financed by a one-time 50% tax, paid in stock, on the largest AI companies. Sanders frames it as a way to transfer part of the wealth and power of the AI industry to the American public, with the shares held in a public trust that would also carry voting rights and board seats.

Seen against that backdrop, the proposal attributed to OpenAI looks like an attempt to get ahead of the political pressure. Five percent is a lot of money, but it is also very far from Sanders’ model. For OpenAI, it may be a way to show that it shares the upside with the public without giving up meaningful control of the company.

The State Is Already Moving Closer to the Tech Companies

The idea of the US government holding a stake in a large technology company used to sound highly unusual. It sounds less so today. The Trump administration already took a stake of 9.9% in Intel, and Reuters has reported unusual arrangements with Nvidia and AMD, under which the companies agreed to pass 15% of the revenue from certain AI-chip sales to China back to the government.

None of this means OpenAI is on its way to being nationalized. It does show that the line between regulation, industrial policy, and government ownership is getting less clear. When AI is treated as strategic infrastructure, the state no longer settles for setting the rules from the outside. It also looks for ways to participate economically in what it sees as a future national asset. That reframing is the same one that is quietly turning routine AI oversight into a boardroom-level problem, close to the gap between what AI does inside an organization and what leadership can actually see and report on.

What the Public Would Actually Get

This is where the complexity starts. A government stake can sound like an elegant solution, but the details will decide whether the public actually gains. Would the shares be held in an independent fund? Would citizens receive dividends? Would the government hold voting rights? Could a regulator that owns shares in a company still oversee it without a conflict of interest?

There is also an international dimension, and it is not merely theoretical. A Forrester analyst told Reuters that a government stake taken before an IPO could ease investor concerns about regulation in the United States, but could also trigger similar demands from other countries. Reuters separately reported that Anthropic and the US administration have not discussed taking a government stake in the company, a reminder that the idea is still far from an industry standard. Anthropic has pointed instead toward a tax-financed digital dividend rather than handing over equity, which is its own answer to the same question of who ultimately audits and benefits from a critical AI vendor.

At this stage the answer is still unclear. But the proposal itself teaches something important about the current moment. AI companies can no longer settle for promising that the technology will benefit everyone in the future. As their valuations rise and their influence widens, they will be asked to explain who receives a share of the wealth they create, who oversees them, and what the public gets in return for the risk it is asked to carry. And the next question is broader still. If a model like this advances in the United States, will other countries demand that AI companies grant them a similar arrangement as a condition for operating in their territory?