69% of the U.S. already wants to expropriate half of AI companies: layoffs have politicized abundance

By Startup Fortune · July 13, 2026.

The survey was conducted by the firm Verasight, with 1,690 adults, and the headline figure is emphatic: 69% of Americans support AI companies transferring half of their capital to a public sovereign wealth fund. It is not an isolated opinion. 81% want the government to be able to block the launch of a model if it deems it dangerous, and 89% demand that companies be legally required to publish their safety test results. Public support for tough intervention in the AI industry is no longer marginal; it is a majority.

The context explains the number. According to Challenger, Gray & Christmas, the tech sector accounted for nearly a third of all layoffs in the U.S. in the first half of 2026: 139,156 cuts, up 83% from the previous year. And across the economy as a whole, the firm explicitly attributed AI as the cause of 101,743 layoffs, close to 23% of the total, making it for four consecutive months the reason most cited by companies for reducing headcount. Joseph Briggs, an economist at Goldman Sachs, estimates that more than 9% of the U.S. workforce —some 15 million workers— could be displaced over a transition of roughly a decade, with technology, management consulting and graphic design among the sectors hit hardest already, at a pace of 10,000 to 15,000 fewer jobs a month in employment growth.

The political response already has the shape of a bill. Senator Bernie Sanders introduced in June the American AI Sovereign Wealth Fund Act, which would impose a one-time 50% tax in shares on AI companies with more than $200 million in annual revenue from that business, earmarked for a fund of roughly $7 trillion managed by an independent commission confirmed by the Senate. Sanders talks of an annual dividend of about $1,000 per citizen, funded moreover by a significant structural obligation: companies with mixed businesses —think Microsoft or Google, where AI is intertwined with the cloud, search or productivity software— would have to separate their AI arm in their accounts so that public ownership applies only to that portion.

The figure that should truly worry boards is not that Sanders, a democratic socialist in a Republican-majority Congress, has the paperwork ready. It is that, as Fortune has reported, Trump's White House has floated its own version of public ownership tied to AI infrastructure. When the left and the populist right converge on the same idea —that the state should keep a share of AI's capital— the proposal stops being predictable as a guaranteed legislative failure and becomes a risk that cap tables are beginning to price in, as the article itself warns: a startup raising a Series C today does so with 69% of the country demanding half of their company as a backdrop.

Our read: this survey is the political translation, almost literal, of the thesis we have been maintaining about jobs and AI. The destruction of administrative and back-office positions is not a distant hypothesis; it is the explanation companies themselves are giving today in their layoff announcements, month after month. What matters is not that people are angry —that was foreseeable— but that this anger already has a bipartisan institutional vehicle. And here it is worth separating two things the public debate conflates: the abundance AI promises in the long term (curing diseases, extending healthy lifespans, freeing up human time for what people are passionate about) does not arrive automatically by a decree redistributing shares; it arrives if the productivity generated translates into real growth and that growth is redistributed with instruments that do not strangle the investment that made the technology itself possible. A confiscatory 50% tax on capital, imposed by decree and with a political commission deciding which models are "harmful," resolves short-term anxiety but may introduce governance friction that slows precisely the innovation underpinning future abundance. Honesty requires stating both things: social discontent is justified by real layoff figures, and at the same time, expropriating cap tables by law is legally demanding and economically risky if done without nuance. What does seem to be taking hold, beyond whether this particular bill advances, is that the conversation about who captures AI's gains —capital, the displaced worker, or the state as intermediary— is no longer waged only in the offices of OpenAI, Anthropic or Nvidia. It is waged, survey figures in hand, in Congress.