OpenAI CEO Sam Altman privately told a group of conservatives this summer that he did not support having the government take shares in the company, distancing himself from a proposal that he and his firm were widely understood to have supported.
For months, OpenAI has been connected to the idea that it is interested in voluntarily ceding shares to the government. Altman has spoken to senior Trump administration officials about the idea, as The Washington Sun first reported in June. The Financial Times reported that OpenAI had proposed giving the government a 5 percent stake in the company, citing Altman’s private comments. The New York Times wrote that Altman had directly pitched President Donald Trump on creating a “government investment fund” seeded with shares in OpenAI. And The Associated Press reported that Altman told Sen. Bernie Sanders (I-Vermont) that he wants the public to have equity in artificial intelligence companies.
At the off-record meeting with conservative economists on July 29, Altman said that he supports some version of redistributing company shares to the American public, according to three people present and notes taken by one participant reviewed by The Washington Sun. But he also was adamant that he would oppose the government taking equity in his firm and would resist any efforts for federal control over the company, those in the room said.
Altman’s comments, previously unreported, provide new insight about the intentions of the leader of one of the most powerful technology companies in the world, as companies like his own and Anthropic navigate public backlash against the rise of AI.
Trending
To those present, Altman appeared confused why anyone could think he wanted the government to take shares in the company.
“To most people in Washington, ‘public shares in OpenAI’ obviously means government control,” said one of the economists who attended the meeting. “But he did not seem to understand that and seemed confused that anyone would think that’s what it meant. He said, ‘That’s not what we’re talking about.’ … I had the sense that he did not see as big of a difference between individual Americans owning shares and the government owning shares as there is in reality.”
The economists spoke on the condition of anonymity to describe the off-record discussion.
In a statement to The Sun, the company said it supported creating a “government-administered public wealth fund,” which it distinguished from one in which the government took a “controlling or ownership stake in AI companies, which we would not support.”
“We think AI should create economic opportunity for everyone,” the statement said. “In April, we published an industrial policy blueprint with several ideas, including a public wealth fund that would let people share directly in the economic gains from AI. We’ve also talked with policymakers in both parties about related ideas, including whether AI companies could contribute equity to help seed a fund, but we are not close to anything specific.”
Several experts expressed confusion about Altman’s position. Theoretically, it would be possible for OpenAI to at some point donate “nonvoting” shares to a private entity overseen in part by government officials, which would limit federal control over the firm. But this would still amount to ceding stakes to the government.
“Everything we’ve seen from Sam Altman and OpenAI over the years implied a pretty conventional government ownership stake in the companies, including voting rights, dividend rights, other kinds of shareholder rights,” said Matt Bruenig, founder of the People’s Policy Project, a left-leaning think tank. “The distinction they are drawing is hard to understand: How does a government-administered public wealth fund that owns equity not have an ownership stake in companies? These sentences are contradictory.”
The position also diverges from that understood by others in Washington who have met with Altman. In a meeting in Sanders’ office in late June, Altman said he had seen the senator’s bill requiring AI giants to turn over 50 percent of their shares to the government and agreed that the government should have some stake in the AI companies, though not 50 percent, according to another person familiar with the matter, who also spoke on the condition of anonymity to recount that private discussion.
By contrast, the economists who attended the July 29 meeting came away from the discussion with Altman believing that OpenAI agreed with their fierce opposition to expanding the government’s control over the private sector.
Some experts have long been skeptical that OpenAI really would be voluntarily willing to turn over what would amount to billions of dollars to the government. Jeremy Bearer-Friend, an AI expert and law professor at George Washington University who helped design the Sanders plan, said there is nothing in corporate law currently stopping Altman from ceding shares to the government.
Shareholders could sue the company if it made that decision, but it would have the freedom to try.
“They already could have done it,” Bearer-Friend said.
The same day Altman met with conservative economic policy experts, the OpenAI executive also briefed a group of economists influential with Democratic policymakers. Several attendees of that second meeting did not recall the equity stakes question being discussed.
Both meetings started with presentations from Altman about the mathematical breakthroughs made by the OpenAI models. To the conservative economists, Altman stressed that OpenAI had the least politically biased model of any of the major AI companies, two of the people at that meeting said. The liberal economists did not remember Altman discussing the political orientation of the OpenAI model.
Author
