Musk v. Altman: Three Weeks of Trial, Everything That Happened

The trial pitting Elon Musk against Sam Altman and OpenAI is nearing its close. Three weeks of hearings in Oakland have surfaced unexpected revelations, explosive testimonies, and contradictions that embarrass both sides. Here is the complete record.

Background: a 2015 promise becomes a 2026 lawsuit

The story begins with OpenAI’s founding in 2015. Sam Altman persuaded Elon Musk to co-found an AI lab that would operate entirely as a nonprofit: no shareholders, research published in open source, technology that would « belong to the world. » Musk invested $38 million on that basis.

Ten years later, OpenAI is valued at $852 billion and is preparing a public offering, with Microsoft having invested $13 billion. Musk’s lawsuit alleges breach of founding mission, unjust enrichment, and an anticompetitive arrangement with Microsoft.

Week 1 (April 28 – May 2): Musk on the stand

The trial opened on April 27 before Judge Yvonne Gonzalez Rogers. Musk spent three days on the witness stand more than seven hours in total, repeatedly returning to a single line: « You can’t just steal a charity. »

Cross-examination produced the week’s sharpest moment. Musk admitted that his chatbot Grok, built by xAI, was trained by distilling OpenAI’s models. He downplayed the admission, calling it « standard practice in the industry, » but the contradiction was set in stone: the man suing OpenAI for the commercial exploitation of AI is feeding his own models with their work.

The defense also flagged that Musk founded xAI in March 2023, just days before publicly calling for a six-month moratorium on AI development. A moratorium that would have slowed his direct competitors at the exact moment he was starting up.

Week 2 (May 5 – 9): OpenAI fires back

Greg Brockman, OpenAI’s co-founder, opened the second week by dismantling Musk’s account. According to Brockman, Musk was not betrayed by the creation of a commercial subsidiary. He had in fact demanded absolute control over any for-profit structure and left when it was denied. Brockman described a tense meeting in which Musk, furious at not getting his way, stood up and paced around the table.

Then came the trial’s most unexpected witness: Shivon Zilis, a former OpenAI board member and the mother of four of Elon Musk’s children. She revealed that Musk had actively tried to recruit Altman to lead an AI lab at Tesla, even offering him a seat on Tesla’s board. The implication is significant: Musk did not leave OpenAI over an ideological disagreement. He wanted Altman working for him. When Altman refused, the relationship collapsed.

Zilis also disclosed that she had a personal relationship with Musk while sitting on OpenAI’s board, and had not declared it to the organization a fact that became public only in 2022 through a Business Insider report, adding yet another layer of undisclosed conflicts of interest to the case.

Week 3 (May 11 – 13): Nadella and Altman take the stand

Microsoft CEO Satya Nadella testified on May 11. He stated that Musk had never reached out to him to raise concerns about Microsoft’s investment in OpenAI. He also revealed an internal anxiety: in an email, he had written that he did not want Microsoft to become « the next IBM » while OpenAI became the next Microsoft a reference to how IBM ceded control of the PC era to Microsoft.

Under cross-examination, Nadella acknowledged he was unaware of any full-time employees working for the OpenAI nonprofit before March 2026, nor of any grants, research, or open-sourced technology it had produced. This admission weakens the argument that OpenAI honored its founding mission in any meaningful way.

Sam Altman took the stand on May 12. He introduced himself as « an honest and trustworthy businessperson » and systematically challenged Musk’s claims.

« The framing of ‘stealing a charity’ – I find it hard to even wrap my head around it, it’s so disconnected from reality. » – Sam Altman

Most critically, Altman reversed one of the trial’s central assumptions. He testified that Musk had initially sought 90 percent of the equity in any commercial entity built around OpenAI, along with total control of the organization. « My belief is he wanted to have long-term control, » Altman said. He added that in his view, Musk « didn’t understand how to run a good research lab. »

Altman also described his surprise firing in November 2023, before being reinstated under pressure from employees and investors as an « incredible betrayal » that was « very public and very painful. »

What the judge must decide

Did OpenAI breach its nonprofit mission? Does the shift to a commercial structure with Microsoft’s investment constitute a violation of the founding commitments?

Unjust enrichment? Can Musk claim damages on his initial donations, given that the nonprofit appears never to have functioned substantively?

Anticompetitive arrangement? Does the Microsoft-OpenAI partnership violate competition law?

The nine-person jury plays an advisory role only. The binding decision rests with Judge Gonzalez Rogers, expected by mid-May 2026.

What this trial is really about

Three weeks of hearings make one thing clear: both sides embody the contradictions they ascribe to the other. Musk attacks OpenAI for abandoning the public good while his own AI model trains on their data. OpenAI defends its philanthropic mission while engineering an $852 billion IPO.

This is not a moral trial. It is a power struggle over who controls AI, and in whose interest. Whatever Judge Gonzalez Rogers decides, the ruling could set a precedent for every tech organization that presents itself as mission-driven before pivoting to profit.

That is the real stakes far beyond the egos of the people involved.

Sources: CNBC, MIT Technology Review, NPR, ABC7, GeekWire, Fortune, Al Jazeera, Axios (April–May 2026)