-1920x1281.jpg)
Chip Somodevilla/Getty Images
A version of this article originally appeared in Quartz’s AI & Tech newsletter. Sign up here to get the latest AI & tech news, analysis and insights straight to your inbox.
The idea of the federal government taking equity stakes in private companies has rarely been attempted outside of crisis, and never quite like this.
The Trump administration is in active discussions about acquiring financial stakes in leading artificial intelligence companies. The idea, which OpenAI CEO Sam Altman first pitched to the administration in 2025, would have AI companies donate equity to the federal government to seed a public wealth fund that distributes returns directly to American citizens.
It is, depending on who you ask, either the most important wealth-sharing proposal in a generation or a bailout dressed up in the language of populism.
Vermont Senator Bernie Sanders has his own version, which is more socialist-leaning in nature. He wants to transfer 50 percent of the equity of the top AI companies into a public fund, giving the government board seats and voting power. The proceeds would flow directly to Americans as something closer to a universal dividend.
The industry has its own vision of what that check might look like. Altman has floated universal basic compute, in which every citizen would receive an allocation of AI tokens they could spend, sell, or donate — a kind of universal basic income with extra steps that gives the AI companies a cut.
Trump has already been here before, in smaller ways. His administration has taken equity stakes in Intel $INTC, minority positions in nine quantum computing firms under the CHIPS Act, and stakes in rare earth miners including MP Materials and USA Rare Earth. The Reuters tracker of these deals includes more than a dozen companies across semiconductors, critical minerals, and advanced manufacturing.
Defenders argue it gives taxpayers a return on federal subsidies rather than simply handing out grants. Critics counter that it distorts competition, rewards political proximity, and turns the White House into something uncomfortably close to a venture capital firm running a $2 trillion deficit.
There is also the Huawei problem in reverse: if Washington takes a meaningful stake in OpenAI or Anthropic, allies and trading partners may start asking the same questions about American AI that the United States asked about Huawei, namely whether a company so entwined with its government can really be trusted to operate independently.
The proposal landed the same week that AI-linked stocks shed more than a trillion dollars in market value, with chip stocks bearing the brunt. For critics, the optics were hard to ignore.
Senator Elizabeth Warren of Massachusetts had been sounding the alarm about this possibility for months. Back in November, as Altman was first floating the idea with the administration, she wrote to White House officials demanding assurances that the government would not bail out OpenAI or any of its competitors. "We have seen this before," she warned, "take on enough debt, make enough risky bets, and then demand a taxpayer bailout when those bets go south so the economy does not crash."
Now she is pressing the Senate Banking Committee to haul in Treasury Secretary Scott Bessent and Commerce Secretary Howard Lutnick to answer for what she sees as dangerously light oversight of an industry accumulating systemic financial risk.
Her concern is structural. Once the government has a financial stake in OpenAI or Anthropic, it also has a powerful incentive to rescue them if things go wrong and to go easy on safety regulation in the meantime.
Bessent, as head of the Financial Stability Oversight Council, would be the official charged with preventing exactly that kind of crash. An equity stake that looks like a dividend program in a boom could look a lot like a taxpayer-funded backstop in a bust.
