OpenAI offers the US government a 5% stake worth $42.6 billion

Sam Altman walked into the White House with a proposal that would have sounded insane a year ago. Hand over 5% of OpenAI to the federal government.

The Financial Times broke the story on July 2. Reuters, Bloomberg, CNN, and Forbes all confirmed it within hours. The talks are real, and they are happening at the highest level. Altman personally pitched the idea to President Trump, Commerce Secretary Howard Lutnick, and Treasury Secretary Scott Bessent.

A 5% slice of OpenAI clocks in around $42.6 billion based on the company’s March valuation of $852 billion. That is not a symbolic gesture. That is the largest single corporate giveaway to the US government in history.

But it is not just OpenAI. The proposal calls for every major American AI lab to contribute 5% of their equity into a sovereign wealth vehicle modeled on the Alaska Permanent Fund. Alaska takes its oil revenue, invests it, and pays dividends to residents. OpenAI wants to do the same thing with AI. Every citizen gets a cut.

The idea has bipartisan appeal. Progressives like the wealth-sharing angle. Republicans like the America-first framing. Trump himself told CNBC the same day that government stakes in private companies are “very American.” He pointed to the Intel deal from last August, where Washington took a 10% stake worth $8.9 billion. That stock has since climbed sharply.

Here is the catch. The FT describes the conversations as “early” and “conceptual.” Implementing this would likely require an act of Congress. And nobody knows if Anthropic, Google, or Meta will play ball. Those companies have their own IPO plans, their own valuations, and their own ideas about who should profit from their work.

Both OpenAI and Anthropic are preparing public listings. OpenAI filed its S-1 after losing $38.5 billion in 2025. Anthropic quietly filed paperwork for what could be the biggest AI IPO on record. A government stake taken before those listings would set the terms for everyone else.

The timing is not accidental. The White House has been tightening its grip on frontier models for weeks. Officials forced OpenAI to limit GPT-5.6 to a small group of government-approved partners. The Commerce Department slapped export controls on Anthropic’s Fable and Mythos models, then lifted them after negotiations. Washington is also finalizing voluntary release standards with the major labs, with an announcement possible within days.

OpenAI floated a “public wealth fund” in a policy paper back in April. The pitch was simple: AI will displace jobs and concentrate wealth. Let the public own a piece of the engine doing the displacing. The Alaska model is the template. Oil made Alaska rich. AI might do the same for the country.

The sovereignty implications stretch beyond US borders. Forrester analyst Indranil Bandyopadhyay warned that once Washington owns equity in the labs, other governments will demand the same arrangement. European and Asian customers of American AI providers will have to reconsider their assumptions about neutrality. If the US military and intelligence apparatus holds equity in OpenAI and Anthropic, claims of independence get harder to maintain.

Trump has made AI dominance a centerpiece of his administration. The Intel stake, the MP Materials deal, the GPT-5.6 access restrictions, the voluntary standards negotiations. The pattern is clear. Washington is not just regulating AI. It is becoming a partner in it.

A 5% stake would make the US government the single largest external shareholder in OpenAI. Larger than Microsoft. Larger than Thrive Capital. Larger than any venture firm on the cap table. The company that built ChatGPT would answer to shareholders who also control the Pentagon, the CIA, and the FCC.