The ChatGPT maker has quietly pitched the US government a slice of its equity worth an estimated $42.6 billion. The idea would rope in Anthropic, Google and Meta too, and it lands at a tense moment for the company.
OpenAI has held early-stage talks about handing the US government a 5% ownership stake in the company, a proposal that would turn the world's most valuable AI startup into a partial public asset. The Financial Times first reported the discussions on July 2, citing two people familiar with the talks, and CNBC, Bloomberg and Reuters have since matched the account.
The number sounds abstract until you attach a price to it. At the $852 billion valuation OpenAI locked in during its March funding round, a 5% stake works out to roughly $42.6 billion. That is the figure Chief Executive Sam Altman reportedly raised with President Donald Trump, Commerce Secretary Howard Lutnick and Treasury Secretary Scott Bessent.
Not a one-company deal
Altman's pitch reaches well beyond his own company. According to the reporting, the structure he floated would ask every leading US AI developer to contribute a matching 5% slice of equity into a single government-run vehicle. That means Anthropic, Google and Meta would each be expected to cede the same share.
None of those companies has signaled it would participate. One source told reporters that the administration and Anthropic have not discussed a government stake at all. So while Altman is describing an industry-wide fund, right now it exists mostly as a concept attached to a single company's willingness.
The model behind the plan is the Alaska Permanent Fund, the state vehicle that converts oil revenue into an annual dividend paid to residents. Applied to AI, the logic runs like this: the government holds equity across the frontier labs, the value of that equity grows as the industry grows, and returns flow back to the public. Altman has argued that giving ordinary Americans a financial interest in the technology is the fairest way to spread its gains.
Why now
The timing is the part worth sitting with. This proposal did not surface in a vacuum. It arrived six days after OpenAI delayed the full public launch of its GPT-5.6 model at the government's request, with Lutnick reportedly warning Altman against shipping the model without prior approval.
It also followed a bruising month for OpenAI's closest rival. Anthropic spent much of June with its Claude Fable 5 and Mythos 5 models switched off worldwide under the first US export controls ever applied to an AI model rather than to a piece of hardware. Access was restored only days before the OpenAI news broke.
Read those events together and a picture emerges. Washington has been flexing hard on frontier AI, through export curbs, staggered release requests and national security reviews. The FT framed OpenAI's equity offer as an attempt to "secure good relations with the administration" and blunt political blowback. An equity stake, in that reading, is a peace offering.
The precedent already exists
A government taking a piece of a private company would have sounded outlandish a couple of years ago. It doesn't anymore.
The Trump administration converted CHIPS Act grants into a 9.9% stake in Intel last August, buying in at $20.47 per share. AMD and Nvidia separately agreed to hand over 15% of their China chip revenue in exchange for export licenses. The government has also taken positions in IBM and several quantum and critical-mineral firms during Trump's second term. OpenAI would extend that playbook from semiconductors into the model layer itself.
Vice President JD Vance has said the president prefers equity over cash payouts, which fits the shape of what Altman is proposing.
Bernie Sanders thinks it's far too small
Not everyone views the 5% figure as generous. Senator Bernie Sanders, whom Altman also spoke with, has dismissed the offer as a watered-down version of real public ownership.
In June, Sanders filed the American AI Sovereign Wealth Fund Act, which would seek 50% of the voting shares of major US AI companies through a fund his office valued at around $7 trillion. His stated goal is a $1,000 annual dividend for every American. Against that backdrop, Altman's 5% is the smallest number anyone has publicly attached to the public-ownership debate, a ceiling from the industry set against a floor demanded by its critics.
A long road, and a lot of unknowns
The talks are preliminary, and the obstacles are real. The FT reported that any formal arrangement would likely require an act of Congress, which alone could stall or reshape the entire idea. OpenAI declined to comment to the FT, and the White House did not immediately respond.
There are structural puzzles too. OpenAI is years from a public listing, so a government equity stake in a still-private company is a novel construction. Altman first pitched the concept to the administration in early 2025, and talks have continued behind the scenes for more than a year, which tells you how unsettled the specifics remain.
What is clear is the direction of travel. The company laid out its thinking in an April policy paper titled "Industrial Policy for the Intelligence Age," which proposed a public wealth fund that could invest directly in AI labs and distribute returns to citizens. The 5% stake is that abstract proposal starting to acquire a concrete number, a set of names in the room, and a reason to move now.
Comments
Join the discussion and share your perspective.