Altman Is Offering Equity in Companies He Does Not Own

The Boldest Offer in Silicon Valley History Might Also Be the Emptiest

Sam Altman wants to give the United States government a piece of the AI industry. According to a new Financial Times report, he has been in serious talks about offering 5% equity stakes from OpenAI and a roster of competitors, including Google's DeepMind, Anthropic, Meta, and Microsoft, to form something resembling an AI sovereign wealth fund for American taxpayers, modeled loosely on Alaska's Permanent Fund.

At OpenAI's current valuation of $852 billion, that 5% slice alone would be worth more than $42 billion. It is an audacious idea. It is also, as Neeta points out in this episode, potentially an offer Altman has no legal standing to make.

Altman does not hold direct equity in OpenAI. His personal wealth from the company comes indirectly, through a web of investments in reportedly as many as 400 companies, at least 10 of which are either in current deals with OpenAI or reported to be in talks. He also does not control Google, Anthropic, Meta, or Microsoft. None of those companies have commented on the proposal, and the silence is conspicuous.

The Intel Playbook and Its Complications

The proposal is not as eccentric as it first sounds, and that is partly what makes it worth taking seriously. Last year, under sustained pressure from the current administration, Intel agreed to hand the US government a 10% stake in the company. The political headwinds for Intel largely eased after that agreement. The OpenAI camp appears to be reading that playbook closely.

The problem is that Intel's shareholders are now suing over the deal. And beyond the legal exposure, the arrangement raises a question that cuts to the core of American economic philosophy: should the federal government hold equity positions in private companies?

Historically, the answer from both parties has been a reluctant no, with narrow exceptions. The government's emergency stakes in automakers after the 2008 financial crisis remain the most cited example, and those were treated as crisis measures, not policy templates. The conservative instinct against state ownership of enterprise has not disappeared, even as the current administration shows more appetite for strategic industrial intervention.

A Triangular Mess at the Top of the AI Industry

While Altman works his sovereign-wealth angle, the relationships between the industry's biggest players are fracturing in real time.

Microsoft, once OpenAI's most important backer and partner, has been drifting. The financial ties remain, but the partnership has frayed. Microsoft grew closer to Anthropic in the interim. Now, something appears to have shifted for Satya Nadella. In a wide-ranging CNBC interview last week, he warned plainly that AI giants should not be allowed to eat the economy. It is a striking pivot for someone who spent years championing exactly those giants.

More provocative still: Microsoft is reportedly evaluating whether to offer DeepSeek, the Chinese frontier model, to its enterprise customers. DeepSeek has been credibly accused by both Anthropic and OpenAI of illegally distilling their models, essentially probing and extracting the fruits of American AI research without authorization. Experts believe DeepSeek has made genuine innovations of its own, but the consensus view is that it would not be where it is today without that illegal foundation. If Microsoft proceeds, it would be routing customer revenue toward a product built substantially on its own partners' stolen work.

Meanwhile, Anthropic is reportedly building a competitor to Microsoft Copilot. Given the lukewarm reception Copilot has received from most enterprise users, a well-executed Anthropic alternative could represent a genuine threat to one of Microsoft's most strategically important software bets.

What This Means Going Forward

The sovereign wealth fund idea surfaces a tension that the AI industry has so far managed to defer: who actually owns the value being created, and who bears the political and social risk if it goes wrong? Governments are starting to ask that question with equity stakes rather than just regulatory pressure, and the AI labs are learning that their trillion-dollar valuations make them targets for exactly this kind of negotiation.

For Altman specifically, the path to a trillion-dollar valuation and a credible IPO runs directly through political goodwill. Offering equity, even equity he may not control, is a way of buying that goodwill in advance. Whether the other companies named in the proposal see it that way is another matter entirely.

The next few months will test whether this proposal gains traction or quietly disappears. Either outcome will tell us something important about how the US government and the AI industry intend to divide the spoils of a technology that is reshaping the entire economy.