Elon Musk spent three days on the witness stand in Oakland, California, beginning April 28, 2026, as a key witness in the federal trial against OpenAI. Presiding over the case is Judge Yvonne Gonzalez Rogers. The amount at stake: $130 billion in damages, to be directed entirely to OpenAI's charitable arm. Not a cent would go to Musk personally.
Steven Molo, leading Musk's legal team, told the nine-member advisory jury that Sam Altman and Greg Brockman “stole a charity.” Molo compared the situation to a museum looting its own paintings. OpenAI was founded in 2015 as a nonprofit research lab in Brockman's apartment. According to Bloomberg and Reuters data, it is now valued at over $850 billion. Already trades as a pre-IPO tokenized asset on Binance. Musk invested at least $44 million in the organization during its early years.
What Musk Accuses Altman Of
The core argument is straightforward, though openAI promised to remain a nonprofit. Then came the commercial arm, then Microsoft with over $10 billion in capital, then plans for a public listing. Betrayal of the founding mission, says Musk. An inevitable evolution to compete with DeepMind and the sector's biggest players, counters OpenAI.
The breaking point, in court, has a specific name: the Microsoft investment. Too large for a traditional donation. “Microsoft has its own motivations, different from those of a nonprofit,” Musk testified. From that point he requested an internal legal investigation. In 2018 he left the board, and years later, he founded xAI.
Bill Savitt, OpenAI's attorney, captured the whole case in a single sentence during his opening statement: “We are here because Mr. Musk did not get what he wanted from OpenAI. He left saying they would surely fail. My clients had the courage to go forward without him.” Hard to counter without solid evidence on the other side.
The Sharpest Moments of Cross-Examination
Savitt's cross-examination produced exchanges that were difficult to dismiss. The harshest: Musk admitted he had not read the fine print of a 2018 term sheet that explicitly discussed the possibility of raising $10 billion from future investors. Savitt's response was blunt: “It's a four-page document.” No excuse that holds.
OpenAI Valuation 2019-2026 (USD billions)
Sources: Crunchbase, Bloomberg, Reuters
OpenAI Valuation 2019 — 2026 (USD billions)
* 2026 figure as of the ongoing trial in Oakland. IPO has not yet taken place.
Sources: Crunchbase, Bloomberg, Reuters
There is more. xAI, the AI company Musk founded after leaving OpenAI, used OpenAI's own models for training purposes. Musk confirmed this in court without hesitation: “It's standard practice to use other AI systems to validate your own.” Judge Rogers remarked that she found it “ironic” that Musk is building a for-profit AI company while arguing that the technology could pose an existential danger to humanity. Musk did not respond.
Rogers had already rebuked him on day one. Before the jury entered the courtroom, Musk had posted on X against OpenAI and Altman. Rogers warned him directly: one more post like that and a restraining order follows. The tone of the trial was set before anyone had even been sworn in.
What OpenAI Now Risks
Musk is seeking three things. First: OpenAI returns to a nonprofit structure. Second: Altman and Brockman removed from all operational roles and governance. Third: $130 billion in damages directed to the company's charitable arm. The advisory jury does not issue a verdict; it only guides the judge. The final decision rests with Rogers.
For OpenAI the stakes are tangible. An IPO is in preparation. The company's valuation exceeds $850 billion, per Bloomberg data, and Microsoft's investment underpins much of its infrastructure.
In this season of major tech listings, an adverse outcome could reshuffle the cards in unpredictable ways. Microsoft is named as a co-defendant. The charge: facilitating a breach of the charitable mandate through its investments in the commercial arm. Microsoft's lawyers pushed back on a specific point: Musk was aware of the company's relationship with OpenAI as early as September 2020, years before filing suit, and his claim may be time-barred.
Court reconvenes Monday, May 4. The tribunal was not in session Friday, May 1: the jury received explicit instructions not to discuss the case or conduct research over the weekend. The next witnesses expected are Greg Brockman and Stuart Russell, a computer science professor at UC Berkeley and one of the world's foremost AI safety experts.
Judge Rogers has already rejected Musk's request to introduce testimony about the “extinction risk” posed by artificial intelligence: “This is not a trial about AI safety risks,” Rogers clarified, with a discernible note of dry wit. What Russell will be permitted to say, within the limits the court grants him, is the concrete data point to follow next week.
