Sam Altman in talks on an OpenAI stake for every American family

Sam Altman, CEO of OpenAI, is negotiating a mechanism that would give every American family a stake in the company worth about $300. The Financial Times reported this on July 3, 2026, citing sources familiar with the matter.

What is being discussed

The details of the scheme have not been made public, but the key figure, $300 per family, sets the financial logic of the talks.

• $300 is the approximate market value of the minimum package for one American household
• The US has about 132 million households, so the total size of the scheme would be on the order of $40 billion
• In 2025 OpenAI completed a restructuring and changed its status to a public benefit corporation (PBC), which technically makes it easier to distribute shares to third parties
• The Financial Times reported the talks on July 3, 2026; MIT Technology Review drew attention to the story on July 6

An old idea in a changed context

Altman has publicly promoted the idea of "people's dividends" from AI since 2021. In his essay "Moore's Law for Everything," he described a concept in which AI-driven productivity gains should translate into direct payments to citizens, modeled on the Alaska Permanent Fund, where state residents receive an annual share of oil revenue.

Back then the idea sounded like distant philosophy. Today the context has changed radically: OpenAI's valuation exceeds $300 billion, and the gap between those who hold stakes in AI companies and the rest of America has become a visible political problem. Some members of Congress are already asking openly: on what grounds are the benefits of AI concentrated among a narrow circle of investors?

A separate debate about the cost of automation is growing in parallel. AI systems trained on publicly available text and data have de facto used the labor of millions of authors, journalists and developers without any compensation. Several lawsuits against OpenAI are built on exactly this argument. Handing shares to a broad public could serve as a political answer to those claims, or at least soften the criticism.

"People whose work and public content trained these systems should get something in return" is an argument growing louder across

American society.

Why it is hard to pull off

OpenAI has no public plan with dates and a legal structure. The scheme raises several fundamental questions.

Who gets it. Citizens, permanent residents or taxpayers? Each option changes the scope and the final cost of the scheme.

In what form. A direct transfer of shares requires every recipient to have a brokerage account. A dedicated national fund would need a legislative basis. Distributing securities at this scale may require SEC approval and disclosure comparable to a full IPO.

What happens next. If most families sell their shares immediately, the scheme turns into a one-time $300 payment rather than long-term participation in the company.

For comparison, in Alaska, where the "resource dividend" model has operated since 1982, payments come from a state fund, not from a private company. That is a fundamentally different legal construct. No major US technology company has voluntarily done anything like this.

What this means

The "$300 per family" discussion is both a business initiative and a political maneuver. OpenAI clearly wants to get ahead of regulatory complaints and avoid becoming the company that reshapes the labor market for millions while enriching only its own investors. Whether the talks turn into an announcement with concrete deadlines remains an open question.