SAM ALTMAN OFFERS 5% OF OPENAI TO US GOVT

Sam Altman is in talks with the US government about giving American taxpayers a direct ownership stake in OpenAI. The proposal on the table is five per cent of the company. Altman has reportedly suggested that other top AI builders, including Google, Anthropic, xAI, and Meta, would have to grant the government a similar stake. The discussions are still very early, and any deal would require approval from Congress.

What makes this worth paying attention to is not the number itself. Five per cent of a private company that has not yet settled on a for-profit structure is a symbolic stake. But it signals something real about the trajectory of the relationship between AI companies and the federal government, and it fits a broader pattern the Trump administration has already established.

you may like

FABLE 5 IS BACK ONLINE! BUT THERE IS A LESSON HERE

ALEX KARP SAYS AI COMPANIES DONT KNOW HOW UNLIKEABLE THEY ARE

AMAZON DROPS OPENAI FLICK

THE PATTERN THAT CAME BEFORE

The Intel precedent is the cleanest parallel. President Trump used to call for Intel CEO Lip-Bu Tan to resign until his administration took a 10 per cent stake in the company. Trump has since boasted that "America's stake [in Intel] is now over 60 billion dollars". That is up from $8.9 billion in 2025. The stake grew roughly sevenfold, which is the kind of return that tends to get the attention of people who set policy.

The Intel case makes clear that the administration views direct equity as a legitimate tool for asserting influence over strategically important companies. AI is at least as strategic as chip manufacturing, and the regulatory pressure on the industry has been building all year.

THE PRESSURE CAMPAIGN

In June, Trump signed an executive order asking AI companies to share powerful models for voluntary government review 30 days before public release. That same administration ordered Anthropic to block access to its Mythos and Fable cybersecurity models, though Anthropic was recently granted permission to restore user access. OpenAI itself had to roll out a limited preview of its GPT-5.6 model to government-approved partners at the administration's request.

None of these actions required an equity stake to execute. The executive branch already has enough leverage through export controls, procurement rules, and national security authorities to shape what AI companies do. What equity adds is permanence and upside. A stake gives the government a seat at the table that survives changes in administration, and it turns the government into a shareholder that benefits when the company succeeds.

THE BROADER REGULATORY MOOD

Politicians, including Trump's allies, and organisations like the UN are calling for more stringent AI policies. The direction of travel is clear across the political spectrum. The question is not whether AI will be regulated, but what form that regulation takes. Direct equity is one answer. Others include mandatory safety testing, licensing regimes, and liability frameworks that hold companies accountable for downstream harms.

The five per cent proposal is notable partly because Altman did not stop with the US government. He and other OpenAI executives also floated the idea of giving five per cent equity to sovereign funds like the Alaska Permanent Fund. That suggests a broader strategy: treat government as an investor rather than an adversary, and normalise the idea of public ownership stakes in frontier AI companies across multiple jurisdictions.

WHAT THE CRITICS WILL SAY

The obvious objection is that a five per cent stake does not give the government meaningful control over OpenAI. It does not get the government a board seat unless the voting structure grants it one, and it does not override the existing investors who hold larger positions. The equity would be more symbolic than operational, a political gesture rather than a governance structure.

The counterargument is that the gesture itself changes the terms of the relationship. Once the government is a shareholder, it has a direct financial interest in the company's growth and a legal basis for information requests that would otherwise face resistance. Equity creates alignment of incentives in a way that regulatory oversight alone does not.

PROBABLY WON'T HAPPEN

The talks are early enough that the proposal could collapse before it reaches a formal vote. Congressional approval would be required for any deal, and the legislative calendar is crowded. The Intel stake was executed through different mechanisms, so there is not a clean legislative template to reuse.

what to read next

SPACEX BUYS CURSOR AFTER HISTORIC IPO

SWEDISH COURT ORDERS GOOGLE TO PAY KLARNA $1.46BN IN ANTITRUST DAMAGES

EU ORDERS META TO KILL INIFINITE SCROLL

But the fact that the conversation is happening at all tells you something about how the Trump administration thinks about strategic industries. If the AI companies resist, the administration has other tools, as Anthropic and OpenAI have already learned. If they accept, the US government becomes a direct participant in the most important technology race of the decade. Either way, the relationship has changed.