OAKLAND, Calif. — In 2015, a group of idealists made a promise to humanity. They would build artificial general intelligence — the most transformative technology in history — and they would give it away. No shareholders. No profit motive. No masters but mankind itself. Elon Musk put his name, his credibility, and $38 million behind that promise.
On April 7, 2026, his lawyers filed six pages that suggest he wants it back — all of it.
The document, docketed quietly in federal court in Oakland as Plaintiff's Amended Notice of Remedies, is less a legal brief than an act of demolition. Musk is asking a jury and a judge to do something American courts almost never attempt: to reverse time. To strip Sam Altman from his board seat. To remove both Altman and Greg Brockman as officers of OpenAI's for-profit division. To claw back every dollar of equity and compensation they extracted. And, most audaciously, to unwind OpenAI's entire October 2025 restructuring — peeling back the layers of a company now valued at roughly half a trillion dollars until what remains is the bare nonprofit that Musk says he originally funded.
He says he wants none of the money for himself.
That detail matters. It transforms the suit's optics from revenge to reformation — from a billionaire's grudge into a constitutional question about whether founding promises, made under oath to donors, to engineers, and to the public, can simply be discarded once the economics change.
The answer, OpenAI insists, is yes — and necessarily so. The nonprofit couldn't raise the capital to compete with Google, with Anthropic, with the sovereign wealth funds now flooding into artificial intelligence. Altman's team notes, with some relish, that Musk himself once floated a for-profit structure in earlier board discussions. On April 6, OpenAI's chief strategy officer Jason Kwon escalated, writing to the attorneys general of California and Delaware accusing Musk of orchestrating "anti-competitive behavior" in coordination with rivals — and implying the $79–134 billion damages demand is designed not to vindicate charity law but to kneecap a competitor.
The demand itself remains the case's most vulnerable limb. Musk's experts calculated that OpenAI gained $65–109 billion and Microsoft gained $13–25 billion from his founding contributions. Judge Yvonne Gonzalez Rogers, presiding in Oakland, has already called the methodology "not particularly persuasive" — numbers, she suggested, pulled from the air. Legal analysts at Darrow estimate that even a decisive Musk victory would likely yield restitution of roughly $38 million: the amount he gave, returned. Prediction markets in late March priced his odds of winning at just 28 to 36 percent.
And yet the filing is not, at its core, about money.
Musk's lawyers invoke California corporate law — specifically §5142(a), granting broad standing to remedy charitable trust breaches — alongside precedents in which courts removed directors from failing charities and installed receivers to rebuild them. Their argument is elemental: OpenAI accepted tax exemptions, attracted engineers willing to work for below-market wages, and harvested public goodwill on the explicit promise it would never serve private enrichment. Then, having banked all of that, it pivoted. Microsoft took a stake now worth north of $135 billion. Altman and Brockman accumulated equity in an entity they were expressly prohibited from profiting from.
"Those proceeds belong to the charity," the filing states, "not to its faithless stewards."
What happens after April 27 — when jury selection begins in a courtroom twenty minutes from Silicon Valley — will reverberate far beyond this case. A wave of AI ventures structured as nonprofits-with-commercial-subsidiaries is watching. Venture capitalists who backed that architecture are watching. Microsoft, whose Azure cloud deal with OpenAI runs through 2032, is watching.
The smart money still bets on settlement before summer. Both sides have too much to lose at trial and enough leverage to deal.
But if it goes the distance, twelve Oakland jurors will be asked to decide something no court has ever quite decided: whether the promise to build a god and give it to humanity was a binding contract — or merely a fundraising pitch.
Jury selection begins April 27. The future of artificial intelligence governance may begin the day after.
not investment advice
Sources: https://storage.courtlistener.com/recap/gov.uscourts.cand.433688/gov.uscourts.cand.433688.459.0.pdf