Your family’s $300 stake in OpenAI

Sam Altman has been floating some version of "AI wealth for everyone" since 2021, when he proposed that big companies pay 2.5% of their market value annually into a national fund. That plan never went anywhere. But the idea resurfaced last week when the Financial Times reported Altman is now discussing a 5% government stake in OpenAI directly with President Trump, a much narrower version of the same pitch OpenAI floated back in April.

The math is not exactly staggering. OpenAI was valued at $852 billion after its March funding round, so 5% comes out to roughly $42.6 billion. Spread across the country's 133 million households, that's about $320 apiece if handed out as a one-time equity grant. If instead the stake got folded into something like a sovereign wealth fund — Altman has cited Alaska's oil-dividend model as inspiration — the payout could theoretically grow larger over time, assuming OpenAI ever turns a profit, which it currently does not.

The logic behind these plans has real appeal on both sides of the aisle. Bernie Sanders has pushed a far more aggressive version, calling for a 50% public stake in leading AI firms. The pitch works on two fronts: AI models are trained on human-made books, films, and art without compensating their creators, so equity could function as a kind of retroactive payment. And it offers political cover against the fear that AI will gut the job market, even though economists are split on whether that fear is justified.

For OpenAI, there's an obvious upside beyond good PR. Half of Americans are more wary than excited about AI's spread into daily life, and most don't trust companies to handle it responsibly. A national dividend narrative could soften that. But the sharper motive might be staying close to an administration that's already taken equity stakes in Intel and cut itself into Nvidia's China sales — deals that keep companies in Washington's good graces, which increasingly determines who gets government cooperation on chip exports or protection from being labeled a security risk.

Five years after Altman first raised this idea, there's still no actual policy, just a recurring pitch that resurfaces whenever it's politically convenient. That doesn't make it meaningless, though. It signals that even OpenAI's own leadership isn't sure how the public will feel about AI's payoff, and that the loudest reassurance the industry can offer right now isn't safety — it's cash, however small the check might turn out to be.