Trump weighs Intel-style ownership stakes in OpenAI and Anthropic

President Donald Trump said he might pursue an ownership deal with OpenAI and Anthropic, the two most closely watched names in artificial intelligence.

The idea is reportedly aimed at sharing AI-driven wealth with the American public. It would also turn Washington from regulator into shareholder of the companies it is supposed to be overseeing.

The Intel blueprint

The template already exists. In August 2025, the Trump administration converted $8.9 billion in federal CHIPS Act funding into a roughly 10% equity stake in Intel.

That stake was structured as non-voting. The government got a piece of the upside without a seat in the boardroom, at least on paper.

According to the research findings, Trump first raised the idea of exploring equity partnerships with AI firms in June 2026. His latest comments suggest the concept has not gone away.

Altman is already on board. Anthropic, not so much

OpenAI CEO Sam Altman has been the most enthusiastic participant. He reportedly proposed a 5% government stake in OpenAI to Trump as early as 2025.

Based on a recent valuation of about $852 billion, that 5% slice would be worth approximately $42.6 billion. For comparison, that is several times the $8.9 billion the Intel arrangement involved.

Altman’s pitch reportedly also involved talks with Commerce Secretary Howard Lutnick and Treasury Secretary Scott Bessent.

Anthropic is telling a different story. The company has denied having any discussions about government stakes.

Instead, Anthropic is considering a “digital dividend” funded by taxes on the AI sector. The goal sounds similar: spread the gains from AI beyond shareholders. The mechanism is not, since a tax-funded dividend keeps the government out of the ownership structure entirely.

The IPO wrinkle

The timing complicates matters. Both OpenAI and Anthropic filed for confidential IPOs in 2026, according to the research findings, even as these equity conversations played out.

Notably, the voluntary AI safety accords signed in September 2026 did not include any equity arrangements. Whatever Trump has in mind would sit outside that existing framework.

What this means

For investors, the biggest question is how an Intel-style deal would be priced and structured. In Intel’s case, the stake came from converting funding that was already committed. It is less clear what the government would be exchanging for a stake in OpenAI or Anthropic, neither of which relies on CHIPS Act money in the same way.

If the stake were simply granted, as Altman’s proposal seems to imply, existing shareholders would see their ownership diluted. With OpenAI’s valuation sitting at about $852 billion, even a single-digit percentage represents a large transfer of value.

The Intel stake was non-voting, and a similar structure would likely be needed to avoid the appearance of direct state control over frontier AI development.

The things to watch are concrete. First, whether Trump’s comments turn into a formal proposal with terms attached. Second, whether Anthropic’s digital dividend idea gains traction as an alternative. Third, how either outcome shows up in the companies’ IPO paperwork once those filings become public.