OpenAI CEO Sam Altman has reportedly discussed with President Donald Trump a proposal to give the US government a 5% equity stake in OpenAI. The Financial Times reported the talks last Thursday, bringing renewed attention to an idea Altman and others have floated in different forms.
Background and related proposals
Altman first described a more expansive version of this idea in 2021, proposing that all companies above a certain valuation (not just AI firms) pay 2.5% of their market value each year into a fund that would distribute annual payments to Americans. In April 2024, OpenAI presented a narrower proposal that resembles the version reportedly under discussion now.
The notion has wider political resonance as well: Senator Bernie Sanders has proposed granting Americans a 50% stake in top AI companies.
Why a stake? Two core arguments
Proponents offer two main rationales. First, AI systems are trained on human-created works — books, films, artwork — but the creators of those works typically receive no payment from the AI companies that use their output. A free equity stake could be framed as a form of belated compensation. Second, dividend payments or fund distributions could function as a social safety net, addressing public fears that AI may disrupt labor markets.
How big would the payout be?
Details of the most recent proposal are limited. Using OpenAI’s post-March funding valuation of $852 billion as a reference point, a 5% stake would be worth about $42.6 billion. Dividing that sum equally among roughly 133 million American households would amount to approximately $320 per household.
Alternatively, the government could treat the stake like other sovereign or public wealth funds: retain the equity and distribute a share of returns over time. That approach could yield larger payouts in the long run, if AI companies eventually achieve sustained profitability.
The company is reportedly delaying an initial public offering while it aims for a $1 trillion valuation — a challenging target given heavy spending on data centers and the fact that OpenAI has not yet posted a profit.
Political and industry incentives
For AI companies, offering a stake or promising dividends could help shift public opinion slightly in their favor. Surveys indicate a majority of Americans distrust companies to use AI responsibly, many oppose local data center construction, and roughly half of respondents are more worried than excited about AI’s growing role.
More immediately pragmatic, maintaining a cooperative relationship with the Trump administration could yield benefits for AI firms. The administration has previously negotiated technology-related arrangements — for example, an equity stake involving Intel and deals over Nvidia’s sales to China. Good standing with the government can matter for issues like whether a company’s models are labeled a supply-chain risk or how much official help they receive in countering Chinese rivals.
For now: narrative more than policy
The author’s central takeaway is that these proposals are currently more of a narrative than a concrete policy. Altman has discussed variations on the idea for about five years and reportedly raised it with President Trump early in his term, yet there is little sign of a detailed plan taking shape. Even less likely to gain traction is Sanders’ far more ambitious 50% proposal.
Nevertheless, the debates reveal how unsettled the future of AI remains. Altman has cited the Alaska Permanent Fund — created in the 1970s to share oil revenues with Alaskans — as inspiration. That model rests on two assumptions: that the resource is shared and that it will eventually be depleted. Altman appears willing to accept the first premise for AI, but not the second: he has promised that AI will generate extraordinary wealth for decades. Whether Americans ever receive a direct payment may be beside the point; the proposal’s practical purpose could be persuading the public that the AI boom will be large enough to share.
