This is information reported by the Financial Times on July 2, 2026, and later picked up by Bloomberg, CNBC, CNN, and Forbes: OpenAI is said to have proposed that the U.S. government take an equity stake of around 5% in its capital. Beyond the figure itself, what merits analysis is the idea Sam Altman would appear to be defending: bringing the state into the capital of the entire American artificial intelligence industry, along the lines of a sovereign wealth fund. It is a conception of the relationship between public power and technology that stands in sharp contrast to the path chosen by Europe.
A methodological clarification is essential here, because it is decisive: at this stage, nothing has been finalized. The Financial Times attributes its information to “two people familiar with the discussions,” which it describes as “conceptual” and at an “early stage.” This is therefore not an operation that has been launched, but a proposal under discussion. It is precisely for that reason that it is revealing.
The contours of the proposal, according to the Financial Times
According to the newspaper, OpenAI proposed transferring an approximately 5% stake to the federal government. Based on the post-money valuation of $852 billion reached in the record funding round closed in March 2026, that share would be worth roughly $42.6 billion. Sam Altman reportedly discussed the idea directly with President Donald Trump, as well as with Commerce Secretary Howard Lutnick and Treasury Secretary Scott Bessent.
The most significant element, however, is not OpenAI’s case alone. According to the Financial Times, Altman and other company executives reportedly suggested that the main U.S. AI companies (Google, Anthropic, Meta, and xAI are named) each allocate 5% of their capital to a vehicle inspired by the Alaska Permanent Fund. In practical terms, this would amount to a public slice of the entire AI industry in the United States.
There are, however, two source-related reservations. The list of companies involved remains uncertain: some reports also add semiconductor manufacturers such as Nvidia, Micron, or AMD, and the Financial Times itself notes that it is unknown which companies would agree to contribute. A person familiar with the matter, quoted by Forbes, also says that Anthropic is not holding talks on its own behalf with the administration about any equity transfer. And any implementation would, according to the newspaper, require an act of Congress, a major political and legal hurdle that is still far from being cleared.
The Alaskan model: a non-neutral choice
The reference to the Alaska Permanent Fund is not incidental. Created in 1976 by the Constitution of the State of Alaska, this fund invests part of the state’s oil revenues in diversified assets (equities, bonds, real estate) and has paid an annual dividend to every resident since 1982. It now exceeds $80 billion and is among the largest sovereign wealth funds in the world.
By invoking this precedent, Altman equates the value created by AI with a rent comparable to oil: a form of wealth from which the state would capture a share in order to redistribute it to citizens. The argument, as reported by the Financial Times, is that giving the public a financial interest in these companies would be “the best way to share the gains” from artificial intelligence. This logic has been developing for more than a year: Altman reportedly presented the idea to Donald Trump as early as the beginning of 2025, and OpenAI put forward in April 2026 the concept of a “public wealth fund” designed to let every citizen benefit from AI-driven growth.
An admission about the balance of power between the state and U.S. AI
The context may reveal more about the approach than the stated generosity does. Several reports link the proposal to the growing political pressure facing major AI players in Washington: criticism of their economic weight, social risks, and influence. Ceding a stake to the state would then amount, for the industry, to buying a form of regulatory peace by giving public power a direct interest in its prosperity.
That is where the move becomes ambiguous. Making the state a shareholder also makes it judge and party: it becomes difficult to regulate harshly a sector from which one is drawing dividends. The debate is not limited to one side of the political spectrum either. Senator Bernie Sanders, for his part, is calling for a far more robust public stake (on the order of 50% of the capital of major AI companies), proof that the idea of public ownership of AI cuts across the American political landscape, but with opposing aims: value sharing for some, taking control for others.
This momentum extends a constant of recent American policy: treating AI as a strategic national asset. During his first term, Donald Trump signed an executive order making artificial intelligence a national priority. The proposal reported by the Financial Times pushes that logic one step further: not merely supporting the industry, but taking an equity stake in it.
The exact opposite of European-style sovereignty
The contrast is what best clarifies the significance of the matter. In the United States, sovereignty over AI would be expressed through ownership: the state enters the capital, shares the risks and the gains, and ties its fate to that of national champions. In Europe, by contrast, it is expressed first through rules. The European Union has made the AI Act its central instrument: regulating uses, imposing obligations according to risk level, protecting rights, without taking equity stakes in companies.
Two philosophies are therefore at odds. The American path, as suggested by this proposal, relies on alignment of interests: if the state benefits when AI prospers, it has an interest in letting it prosper. The European path relies on distance: the regulator remains outside the market in order to be able to constrain it. The former risks capture (a shareholder state regulates poorly what enriches it); the latter, often criticized for stifling innovation, preserves the independence of the watchdog.
For European policymakers, the episode matters less for its uncertain outcome than for what it reveals. Faced with the prospect of an American industry backed by its state, Europe will have to decide whether its sovereignty can continue to rely solely on regulation, or whether it too requires some form of investment. At this stage, Altman’s proposal remains conditional: “conceptual” discussions at an “early stage,” according to the Financial Times, and a mandatory passage through Congress. The dividing line it draws between the two sides of the Atlantic, however, is already clear.