OpenAI's 5% US government stake idea: what it means for you

If you saw headlines about OpenAI offering the US government a stake in itself and wondered what on earth that means, here's the plain version. OpenAI — the company behind ChatGPT — reportedly offered to give the US government 5% of the company. At OpenAI's value of around $852 billion, that 5% is worth roughly $42.6 billion. The pitch, according to the Financial Times on 2 July 2026, is that the public should get to share in the money AI makes. The first thing to know: this is an early idea, not a signed deal.

Owning a stake in a company just means owning a slice of it — and a slice of its future profits and value. If OpenAI hands the government 5%, the government owns 5% of OpenAI. But the interesting part is what happens to that slice. OpenAI reportedly wants it to go into a public fund modelled on the Alaska Permanent Fund. Alaska takes money from its oil and puts it in a fund that pays every resident a cheque each year. OpenAI's idea swaps oil for AI: the government's slice of the AI industry would sit in a fund that pays out to the public. In theory, ordinary people would see some benefit from AI's growth instead of all the gains going to the company and its investors.

Two reasons, mostly. First, OpenAI is under a lot of political pressure — worries about what powerful AI can do, and about a handful of companies getting very rich and very powerful. Offering the government a share is a way to turn critics into partners: it's harder to crack down hard on a company you co-own. Second, cheaper AI models from China are piling on the pressure, so US companies want the government on their side. OpenAI's boss, Sam Altman, talked the idea over with President Trump and senior officials, and also with Senator Bernie Sanders — who, it's worth knowing, wants a much bigger 50% government stake, not 5%.

This is the part to keep clear, because a lot of coverage makes it sound more certain than it is. What's reported is that OpenAI floated the idea and that senior people discussed it. What's not settled is basically everything else. The talks are described as early and 'conceptual'; a deal like this would probably need Congress to pass a law; and the plan assumes other big AI firms — Anthropic, Google and Meta — would give up similar stakes, yet Anthropic and the government reportedly haven't even discussed it. When reporters asked, most of the companies and the White House declined to comment. One thing that makes the idea believable, though: in 2025 the US government already took about 10% of the chipmaker Intel, so owning a piece of a big tech company isn't unheard of. For now, treat it as a serious idea being tested — not something that's happening yet.

It's also worth knowing why this idea is landing now rather than a year ago. Two pressures are squeezing the big US AI companies at once: governments are increasingly nervous about what powerful models can do, and much cheaper AI models from China are undercutting the case that American labs deserve special treatment. Offering the public a financial stake is, in part, a way to answer the question voters are starting to ask — 'if AI makes these companies unimaginably rich, what do the rest of us get?' Whether a government stake is the right answer is genuinely debated — critics worry it ties the state's finances to the companies it regulates — but the question itself isn't going away, which is why you'll likely hear more about ideas like this, not less.