OpenAI Offered Washington a 5% Stake Worth $42 Billion. It's Not a Gift.

Sam Altman has spent the past year telling regulators that AI will eliminate jobs faster than the economy can absorb the shock. This week, according to the Financial Times, he offered a fix: give the American public a 5% stake in OpenAI itself, worth roughly $42.6 billion at the company's $852 billion valuation. Altman reportedly pitched the idea directly to President Trump, Commerce Secretary Howard Lutnick, and Treasury Secretary Scott Bessent, framing it as a way to let ordinary citizens share in the upside of a technology that OpenAI itself keeps warning could hollow out their jobs.

The mechanism is borrowed from a 50-year-old idea. Alaska has paid residents an annual dividend from its oil wealth since 1976 through the Alaska Permanent Fund, and OpenAI wants to build something similar for AI: a Public Wealth Fund that would hold equity stakes handed over by AI companies and pay out dividends to Americans. Critically, OpenAI isn't proposing to go it alone. The pitch explicitly calls for Anthropic, Google, and Meta to contribute matching 5% stakes, turning a single company's PR maneuver into a proposed industry-wide tax paid in equity rather than dollars. Notably absent from the ask are the chipmakers, Nvidia and AMD, who have arguably captured more of the AI boom's value than any single model developer.

The timing is not an accident. OpenAI has spent 2026 fending off a steady drumbeat of criticism over its valuation, its governance structure, and its trillion-dollar IPO ambitions, while simultaneously publishing policy papers calling for four-day work weeks, higher corporate taxes, and levies on companies that replace workers with AI. A sovereign-wealth-style stake handed to the government does double duty: it's a hedge against the political backlash of being the company that automates away millions of white-collar jobs, and it's a preemptive answer to lawmakers who are already asking harder questions. Senator Bernie Sanders, for one, has floated a far more aggressive version of the same idea — a 50% government stake across all major AI firms, not 5%.

Trump's own reaction, per reporting, was cautiously receptive: he called the concept "interesting" and said his administration would look into arrangements where "the American public essentially becomes a partner with the companies." That's a notably different posture than the deregulatory, hands-off stance the administration has taken on AI policy elsewhere, and it hints at how much political pressure has built up around a handful of AI labs commanding valuations that rival entire national economies. Any real version of this proposal would almost certainly require an act of Congress, and getting Anthropic, Google, and Meta to voluntarily hand over billions in equity is a very different proposition than OpenAI floating the idea in a private meeting.

Strip away the populist framing and the proposal looks less like generosity and more like insurance. A government that owns 5% of OpenAI has a direct financial stake in the company succeeding, which makes aggressive regulation, antitrust action, or a forced breakup considerably less appealing. It also gives OpenAI a talking point every time critics point out that its technology displaces workers or its valuation dwarfs the market caps of companies that actually turn a profit. Whether the other AI giants play along, or whether this dies in a congressional committee, the fact that a 5% government stake is even being discussed at OpenAI's scale says something about how fused AI policy and AI ownership have become in 2026.