Sam Altman Wants to Give the US Government a $42 Billion Slice of OpenAI — and He Wants Every Other AI Company to Follow

There's a sentence Sam Altman wrote in the Financial Times this week that deserves to be read twice.

"The labs develop the technology, but citizens and their elected representatives must make the rules."

It's a reasonable thing to say. It's also a remarkable thing to say for the CEO of a company that is simultaneously proposing to give the government that makes those rules a $42.6 billion financial stake in its own future.

That is the proposal now confirmed by the Financial Times, Bloomberg, CNBC, and CNN: OpenAI has approached the Trump administration about transferring approximately 5% of its equity to a US government-linked investment vehicle — and Altman has suggested that the arrangement should extend to every other leading US AI company, including Google, Anthropic, Meta, and xAI.

At OpenAI's current valuation of $852 billion — set during its record-breaking $40 billion funding round in March 2026 — a 5% stake is worth approximately $42.6 billion. Multiply that across the top five US AI companies and you're describing a public investment position worth hundreds of billions of dollars. Altman — who sat alongside Trump at the G7 summit in June — has reportedly discussed the idea directly with President Trump, Treasury Secretary Scott Bessent, and Commerce Secretary Howard Lutnick.

It is one of the most unusual corporate proposals in American business history. It may also be one of the most strategically shrewd.

What Is Actually Being Proposed

The mechanism Altman has described is modelled on the Alaska Permanent Fund — a sovereign wealth fund that invests the state's oil revenue and pays annual dividends directly to Alaska residents. The AI version would work similarly: leading US AI companies contribute 5% of their equity to a public investment vehicle, that vehicle holds and grows the position, and returns flow back to American citizens.

OpenAI's April policy paper, titled "Industrial Policy for the Intelligence Age," described the concept explicitly: that returns from the fund could be distributed directly to citizens, allowing more people to participate in the upside of AI-driven growth regardless of their starting wealth or access to capital.

Anthropic has separately published its own version of this thinking, calling for "universal pre-distributive capital accounts" with priority given to workers whose jobs are most exposed to AI disruption.

The proposal is still at an early, conceptual stage. Any formal action would likely require congressional approval, which would significantly complicate the matter. Other AI companies — Google, Meta, Anthropic — have not publicly committed to participating. An unnamed source told Reuters that the Trump administration and Anthropic have not discussed a government stake in the company specifically.

But the discussions are real, high-level, and moving fast.

The Strategic Logic Underneath the Generosity

Read the proposal charitably and it's a genuine attempt to solve a real problem: the benefits of AI are accruing overwhelmingly to a small number of shareholders and employees at a handful of companies, while the disruptions — to jobs, to wages, to entire industries — are distributed broadly across the population. A public wealth fund is one way to rebalance that equation.

President Trump confirmed publicly that he had discussed "concepts where pieces could be given to the American public, where the American public essentially becomes a partner with the companies."

Read it more critically, and the political calculus becomes equally visible. OpenAI and its peers are under growing pressure in Washington on multiple fronts simultaneously. The government has shown it's willing to block AI model releases — OpenAI was required to provide a limited preview of GPT-5.6 exclusively to government-approved partners before wider release. The Anthropic episode demonstrated that export controls on AI models are now a real tool the administration is prepared to use.

The equity proposal reportedly aims to influence regulation, prevent ad hoc government intervention, and ensure structured global AI deployment. In blunter terms: a government that owns a piece of OpenAI has a financial incentive to see OpenAI succeed. That's not generosity. That's alignment of interests — and it's a remarkably direct way to achieve it.

The Financial Times noted a precedent: the Trump administration acquired a 10% stake in Intel, which reportedly led to a significant increase in Intel's valuation. The pattern suggests the administration is open to this kind of arrangement when it serves both parties.

The $3 Trillion Question Underneath All of This

The sovereign wealth fund proposal doesn't exist in a vacuum. It arrives at a moment when the AI industry is grappling with a structural question that nobody has publicly answered yet.

Total AI infrastructure spending in 2026 is estimated at $1.5 trillion. To justify that level of capital expenditure, the AI industry will need to generate approximately $3 trillion in revenue — and that estimate is probably conservative, given rising costs of memory and inference-specific chips.

The gap between what's been spent and what's been earned is the defining pressure on the industry right now. Anthropic is thought to have hit $60 billion in annualised revenue, while OpenAI reportedly earned $13 billion in 2025. Those are impressive numbers. They are also nowhere near the revenue base required to justify $1.5 trillion in infrastructure spending.

The sovereign wealth fund proposal is, in part, a political insurance policy for the period when that gap becomes impossible to ignore. If the government is a financial stakeholder in the outcome, it becomes a partner in justifying the investment, rather than a regulator asking uncomfortable questions about returns.

Global startups raised a record $510 billion in the first half of 2026, with AI dominating the surge. OpenAI and Anthropic alone accounted for a massive share of total capital. The money is concentrated, the infrastructure bet is enormous, and the revenue to justify it is still being built. The administration's willingness to own a piece of that bet changes the political equation significantly.

What Happens If This Goes Through

The practical implications, if any deal reaches completion, are significant in several directions at once.

For American citizens, the optimistic version is a genuine dividend-paying stake in the most valuable industry of the next century — the AI equivalent of Alaskans receiving annual oil cheques. The pessimistic version is a government equity position that delivers modest returns while creating a conflict of interest in regulatory oversight of the very companies the government now owns.

For the AI industry, a government equity stake changes the IPO equation dramatically. Both OpenAI and Anthropic are reportedly preparing for public listings, with some investors believing valuations could exceed $1 trillion. Introducing government equity into that equation fundamentally changes how those companies may be assessed by public markets — and potentially complicates the independence investors typically expect from publicly traded companies.

The ripple effects reach well beyond the US. Countries like Kenya, currently drafting their own AI regulatory frameworks, will find themselves shaped by whatever governance model Washington and its AI partners settle on.

For geopolitics, the forum Altman envisions would function as a kind of certification body: countries join by agreeing to shared rules, companies within those countries undergo regular audits, and access to advanced AI follows compliance. That is, in effect, a proposal for the United States to govern who gets access to frontier AI globally — with American companies, and the American government as their financial partner, sitting at the centre of that system.

Whether a sitting government that has already demonstrated its willingness to use unilateral export controls on AI models is prepared to cede that power to an international forum is, as one analyst put it, at minimum an open question.

The Sentence Worth Coming Back To

"Democratic institutions must not cede their responsibilities to AI labs. The labs develop the technology, but citizens and their elected representatives must make the rules."

Altman wrote that. He also proposed giving the government that makes those rules a $42 billion stake in his company.

Both things can be genuinely true at the same time. The political pressure AI companies are under is real. The desire to share AI's gains more broadly is, in at least some corners of these organisations, sincere. And the financial alignment of interests between AI labs and the government is a legitimate mechanism for reducing the kind of unilateral, ad hoc regulatory intervention that took Fable 5 offline for eighteen days in June.

What the proposal also does, inescapably, is make the world's most powerful government a financial stakeholder in the world's most powerful technology companies — at precisely the moment when those companies most need the government not to slow them down.

That's not a conspiracy. It's just how power works. And it's the clearest signal yet that the AI industry has moved from being a disruptor of existing institutions to being a participant in them — negotiating, lobbying, and now offering equity stakes to stay on the right side of the rules it's simultaneously asking to help write.