TL;DR
Trump says he will meet AI companies next week on a “partnership” plan to give Americans an equity stake. Altman first pitched the idea in early 2025. Sanders wants 50%. Critics warn government ownership would undermine AI regulation.
President Donald Trump said on Thursday that he will likely meet with AI companies at the White House next week to discuss what he called a federal government “partnership” that would let the American public profit from the industry’s success. “There are concepts where pieces could be given to the American public, where the American public essentially becomes a partner with the companies,” Trump said, adding that the programme could include sending company dividends to Americans.
The idea is not new. OpenAI CEO Sam Altman first pitched the concept of government stakes in major AI firms to the Trump administration in early 2025, and has raised it again in recent weeks. But the president’s public endorsement, combined with a parallel and far more aggressive proposal from Senator Bernie Sanders, has thrust the question of who owns AI’s economic upside into the centre of Washington’s fractured AI policy debate.
Two proposals, very different teeth
Altman’s version is voluntary. OpenAI’s April policy document proposed a “Public Wealth Fund” that would give every citizen a stake in AI-driven economic growth. Under this framing, OpenAI could donate equity to the government to seed the fund. The pitch positions the company as a willing participant, not a target.
Sanders’ version is not voluntary. The senator announced plans for the AI Sovereign Wealth Fund Act, which would impose a one-time 50% tax on the stock of the largest AI companies, paid in shares rather than cash. The government would manage the resulting fund, use its voting shares to secure board representation, and block decisions it deems harmful. The bill targets firms like OpenAI, Anthropic, and xAI. The legislation is widely considered unlikely to pass, but it has shifted the boundaries of what Washington is willing to discuss.
The gap between the two proposals is vast. Altman is offering a gift. Sanders is proposing a seizure. Trump appears to be somewhere in the middle, publicly warming to the concept without committing to either version’s mechanics.
The conflict-of-interest problem
The most pointed criticism of any government equity arrangement came from Nat Purser of the advocacy group Public Knowledge, who cautioned against any setup that makes the government less inclined to impose or enforce safety regulations because doing so might diminish the value of its own holdings.
The concern is not abstract. Three days before Trump’s profit-sharing comments, he signed an executive order asking AI companies to voluntarily submit frontier models for government testing up to 30 days before public release. The word “voluntarily” is doing a lot of work in that sentence. If the government holds equity in the companies it is supposed to regulate, the incentive to keep that testing voluntary, rather than mandatory, becomes even stronger.
Anthropic has said it is not involved in discussions with the administration about granting equity to the government. No other company has publicly confirmed participation.
Precedent, but not quite
The Trump administration has already taken equity positions in Intel, IBM, and several quantum and critical minerals companies during the president’s second term. But those stakes came with specific industrial policy objectives, typically tied to subsidies or procurement contracts. An AI equity programme would be structurally different: it would aim to redistribute wealth from an entire category of companies to the general public, not to secure supply chains.
The legal question nobody has answered
The most fundamental obstacle is mechanical. How would a private AI company transfer equity to the federal government? OpenAI, currently valued at more than $850 billion by private investors, is restructuring from a non-profit into a for-profit entity and preparing for an IPO as soon as this year. It closed a record-breaking funding round in March co-led by MGX, backed by Abu Dhabi’s sovereign wealth fund. Anthropic has filed confidentially for an IPO. xAI merged into SpaceX, which is about to conduct the largest IPO in history. Each has a different corporate structure, different investor base, and different fiduciary obligations. A one-size-fits-all equity transfer mechanism does not exist.
Sanders’ bill would use the tax code, but a 50% stock seizure raises obvious constitutional questions. Altman’s voluntary donation model avoids those, but depends entirely on companies choosing to participate, which means the fund would be as large or as small as the industry decides.
The California approach, announced by Governor Gavin Newsom in a separate executive order last month, takes a different path entirely. Newsom is exploring mandatory equity grants for workers displaced by AI, a state-level capital pool funded by taxes on AI productivity gains, and voluntary profit-sharing programmes incentivised by tax breaks. None of these require the federal government to hold shares directly.
What happens next
Trump said the meeting with AI companies will happen “next week.” No formal agenda has been announced. The planning remains at the discussion stage, with no companies confirmed and no legal framework in place.
The broader context is a country that has spent 18 months unable to agree on basic AI safety rules, let alone ownership structures. The EU has its AI Act. The US has a stalled executive order, a voluntary testing regime, and now two competing proposals to give the public a cut of the profits. Whether any of it results in Americans actually receiving dividends from AI companies, or whether the conversation simply makes the companies look generous while changing nothing, will depend on details that nobody has written yet.
