OAKLAND, CA., 4/30/26 — Elon Musk returned to the stand Wednesday in federal court in Oakland to continue the testimony he began on Tuesday in his lawsuit against Sam Altman and OpenAI.
The lawsuit arises from Musk’s claim that OpenAI — the San Francisco-based artificial intelligence developer that Musk, Altman, Gregory Brockman and Ilya Sutskever founded as a nonprofit corporation in 2015 — has become a profit-seeking commercial operation in violation of its founding charter and the law. The suit also names Microsoft as a defendant, as Musk claims the company played a role in moving OpenAI to a for-profit structure. Sutskever, however, is not a named defendant.
Direct examination
When Musk stepped to the stand as the first witness in his case, his attorney, Steven Molo, asked, “Big picture, what’s this suit about?”
“It’s really very simple,” Musk replied. “It’s not ok to steal a charity.” He added that, if he lost the case, “the entire foundation of charity in the U.S. will be destroyed.”
Molo then led Musk through a discussion of his career, casting him as a visionary concerned with the fate of humanity.
He founded SpaceX, Musk said, “to make life interplanetary.” Molo reminded the jury that SpaceX, a private aerospace company, had rescued American astronauts from the International Space Station when the Boeing Starliner spacecraft failed.
Musk is also the CEO of Tesla, the electric car company whose founding mission was to “accelerate the transition to sustainable energy” after both General Motors and Toyota had canceled their electric vehicle programs.
“We had to make a good electric car with long range and then convince people to buy it,” Musk continued.
He said that Tesla now sells 2 million vehicles per year, as well as solar power systems and storage batteries, all needed for “the sustainable energy transition.”
Musk told the court he founded Neuralink for “a better human-AI symbiosis.” He said Neuralink technology had enabled people with quadriplegia and people with ALS, a neurodegenerative disease, “to use their computers and phones and speak again.”
Asked how he could manage his duties at so many companies, Musk said he worked 80 to 100 hours a week and he “liked to solve problems to help people.” He added. “I don’t have any yachts or anything.”
Molo called out a flattering email Musk received from Sutskever, one of the four founders of OpenAI, in which Sutskever wrote, “It helps that we have the most overwhelmingly competent person in the world helping us.” Asked about the emails, Musk explained, “Really talented people want to work with me.”
Testifying in a relaxed and confident manner — though occasionally too fast for the court reporter to keep up — Musk described his lifelong concern about AI safety and how that worry accelerated as he saw Google going all-in to develop AGI (artificial general intelligence).
Musk said that he believed it would be disastrous if a profit-making entity like Google achieved AGI before anyone else and used it to wield enormous power.
Musk described his extensive contributions to the formation and establishment of OpenAI, contributions that he said went way beyond financial ones. He emphasized his role in tutoring Altman and another OpenAI co-founder Gregory Brockman on how to build a company and his role in recruiting much of the organization’s engineering talent, including Sutskever, an AI scientist Musk said was one of the top three in the world. Musk said that without him “there is no OpenAI.”
Molo walked Musk through a series of exchanges between Musk and the other three founders in September and October 2017. Musk testified that much of the discussion was “brainstorming” about ways that OpenAI could raise more capital.
Among the ideas that were floated was creating a for-profit subsidiary and giving the four founders equity. Musk discussed a term sheet that was sent to him on that topic that proposed he would get 51% of the equity, with the other receiving three smaller amounts.
In that context, Musk discussed a structure that would have initially given him control of the for-profit entity and the ability to appoint most of the board — noting that he wanted this initial control to make sure things “were going in the right direction.” He said he anticipated that after a one- or two-year period, after other investors came on board, he would have the right to appoint four of 12 board members.
Musk said he ultimately grew frustrated with the discussion about equity. He wrote an email that gave the others an ultimatum: they had to decide either to go it alone as a for-profit or stick with the nonprofit structure.
Musk recounted that both Brockman and Altman quickly responded that they were good with the nonprofit structure.
He said that in 2018 he left the board of OpenAI because there were many issues at SpaceX and Tesla. He said Tesla was so close to bankruptcy that he was sleeping at the factory to get the company back on track.
Musk testified that his view about OpenAI’s management went through three stages. In the initial stage, he was confident in the OpenAI team. In the second phase, he said he started to have questions about whether they might be trying “to steal the charity.” In the third phase, he had concluded that the charity had in fact been stolen.
Cross-examination
Under cross-examination by William Savitt, the lead lawyer for Altman, Brockman and OpenAI, Musk answered the questions in a clear and confident — at times wary — voice. He sparred with Savitt, sometimes challenging the questions as complex and designed to “trick” him. He was particularly testy when asked “yes” or “no” questions, frequently noting that a “yes” or “no” answer would be misleading.
Savitt used his cross to challenge several key areas of Musk’s testimony.
First, he questioned whether Musk genuinely cared about OpenAI having a for-profit subsidiary. Savitt bore down Musk’s participation in the 2017 discussions over the proposal to give him equity, which he implied failed not because of Musk’s concern over the for-profit nature of the subsidiary but because the other founders would not give him full control.
Musk’s response was that he had no problem with a for-profit subsidiary as long as it was “in service to” the for-profit. He said his problem was that at OpenAI the subsidiary had become the “main event.” When referencing the September 2017 discussions, Musk said he only wanted control for an initial period so he could make sure it was going “in the right direction.”
Second, in service of the defendants’ argument that Musk waited too long to sue, Savitt challenged Musk’s statements about when the final phase of his “three phases” began. Savitt presented exhibits suggesting that Musk had learned negative information about Altman and Brockman’s activities far earlier than he admitted. He used old emails and texts to imply that Musk knew enough to have brought his claim as long ago as 2019 and had slept on his rights.
Third, Savitt repeatedly challenged Musk’s testimony as inconsistent with his prior deposition testimony, suggesting that Musk was an unreliable witness who could not be trusted to tell the truth. And when Musk said he did not remember a document from seven or eight years ago, Savitt agreed it was hard to recall things that long ago, an observation supporting his position that by suing many years after the events at issue, Musk has slept on his rights.
Musk’s cross-examination did not finish by the time the judge excused the jury so he will return Thursday to complete Savitt’s cross. Microsoft’s lawyer said his cross-examination would be short.