A five percent gift worth 42 billion dollars
On 2 July 2026 the Financial Times reported that Sam Altman has proposed handing roughly 5 percent of OpenAI's equity to a new US sovereign wealth fund. At the 852 billion dollar valuation investors set in March, that slice is worth about 42.6 billion dollars. CNBC, TechCrunch and Euronews confirmed the outline of the plan the same day.
The idea is larger than one company. Altman reportedly wants every leading American lab, with names like Anthropic, Google and Meta mentioned, to contribute a similar share to a fund modeled on the Alaska Permanent Fund, which pays residents dividends out of the state's oil wealth. President Trump confirmed that conversations took place, saying he had discussed concepts where pieces could be given to the American public.
Why now: vetting, blowback and a delayed flagship
The timing is not subtle. Days earlier, a US vetting process kept OpenAI's new GPT-5.6 flagship limited to roughly 20 government-approved organizations. In the same fortnight, a rival lab spent 19 days offline under a Commerce Department export-control order before being allowed back. Access to the frontier now runs through Washington, and the labs know it.
The FT's sources describe the equity proposal as a way to secure good relations with the administration and address political blowback. Altman is understood to have raised it with President Trump, Commerce Secretary Howard Lutnick and Treasury Secretary Scott Bessent, and to have spoken with Senator Bernie Sanders, whose own June bill would instead levy a one-time 50 percent stock tax on the largest AI firms.
Equity is cheaper than regulation
Seen from the labs' side, the arithmetic is straightforward. A one-time equity grant dilutes shareholders once and buys standing goodwill. A tax or a hostile licensing regime costs every year and can always be tightened. OpenAI floated a public wealth fund in an April policy paper, so this is a prepared position, not an improvisation under pressure.
Seen from the state's side, it is something Europe deliberately chose not to build. The EU wrote itself a rulebook, the AI Act, and stayed off the cap table. The US may end up with the opposite arrangement: few binding rules, but a direct financial stake in the winners. Both are forms of political entanglement. Only one of them pays dividends.
What a state shareholder means for European buyers
For a European operator the question is practical: does your AI vendor answer to its customers, or to its largest political patron? A government that holds equity in a lab has an interest in that lab's pricing power, its export posture and its win rate against foreign rivals. Sovereignty reviews that today flag Chinese state-linked vendors will need a category for American ones.
None of this is theoretical for procurement. The GPT-5.6 rollout already showed that a US administration can decide who gets the best model and when. A financial stake formalizes that relationship. If your critical workflows run on a frontier API, you are downstream of the arrangement whether you signed up for it or not.
What to watch next
The talks are preliminary, and any formal structure would likely need Congress, which is where clean concepts go to become complicated ones. Watch three things: whether a second lab publicly endorses the fund, whether the administration links vetting decisions to participation, and whether the dividend framing survives contact with actual legislation. The precedent matters more than the percentage. Once equity becomes the currency of regulatory peace in AI, every future negotiation starts from that price.
Read next: Nvidia Now Earns Rent on Its Own Chips · AI Tokens Now Have a Rush Hour
