Musk Recounts Losing Trust in Motives of OpenAI’s Other Leaders

Elon Musk spent hours Wednesday telling a jury about his falling out with the other leaders of OpenAI over what he described as their desire to “get rich” off a project that was supposed to serve the public good.

In his second day of testimony during a closely watched trial against OpenAI, the serial entrepreneur detailed his increasingly volatile relationship with fellow co-founders Sam Altman and Greg Brockman beginning in 2017, when he said he became skeptical of their motives.

“It is okay to have a for-profit that is helping the nonprofit, but it can’t become the main thing. It can’t far exceed the nonprofit,” Musk said on the witness stand in federal court in Oakland, California. “That’s really what’s at issue here.”

Musk has alleged that Altman and other leaders at OpenAI enriched themselves by abandoning its altruistic principles and converting to a for-profit company with billions of dollars in support from Microsoft Corp.

OpenAI and Altman have accused Musk of harassment and say the real goal of the lawsuit is to undercut competition with his own startup that he co-founded in 2023, xAI.

The stakes in the trial are high — maybe even existential — for OpenAI because of what Musk is seeking: As much as $134 billion in damages, removal of Altman as CEO and Brockman as president, as well as an unwinding of OpenAI’s for-profit conversion that was completed in October.

Read More: Musk Testifies He’s Suing OpenAI to Stop Altman’s ‘Looting’

Under questioning from his attorney, Musk reviewed a document showing a 2017 proposal for the billionaire to have majority control over a new for-profit entity, if it were created. He acknowledged that was under consideration but stressed such an arrangement would have been temporary.

“If I’m providing almost all the money here and it’s a for-profit, it seems like initial control would make sense,” he said. But as more investors rolled in, he said, that “would mean I did not have control over time.”

In his own statement Tuesday to the jury in federal court in Oakland, California, OpenAI attorney William Savitt argued that Musk “wanted to turn OpenAI into a full-on for-profit company and take absolute control of it.” But, Savitt said, “the other founders refused to turn the keys of artificial intelligence over to one person.”

‘Final Straw’

Musk told jurors Wednesday the “final straw” in his falling out with OpenAI’s other leaders — Altman, Greg Brockman and Ilya Sutskever — came in 2017 when they seemed to turn on him.

Musk read aloud an excerpt in an email from Sutskever: “You’ve shown to us that absolute control is extremely important to you.”

Musk testified that he felt the co-founders “were frankly not being entirely honest with me” after there had initially been a consensus on how to structure the startup.

“They had gone back on what they had agreed on previously,” he said. “I felt this was somewhat disingenuous and what they really wanted to do was create a for profit where they had as much shareholder ownership as possible.”

Nonprofit Status

Musk rejected the idea that he wanted to wield total control of OpenAI indefinitely and said he remained committed in its early years to maintaining the company’s status as a nonprofit research outfit.

He characterized the discussions in 2017 about creating a possible for-profit entity for the AI startup as mere “brainstorming.”

Musk also acknowledged that he directed one of his representative to filed paperwork to create a for-profit entity known as a B-Corp, “but it turned out not to be” needed.

The world’s richest person ended up leaving OpenAI’s board in 2018, after he had contributed $38 million to the startup. Microsoft invested its first billion dollars in OpenAI the next year, followed by $12 billion more.

‘Having Your Cake’

Musk’s attorney, Steven Molo, asked him why xAI was created as a for-profit.

“You can go either way, create something as a nonprofit or a for-profit,” he said. “You just can’t create something as a nonprofit and turn it into a for profit. That’s having your cake and eating it too.”

Musk first sued OpenAI in February 2024 in California state court. He later dropped that case and pursued a new complaint in federal court and added Microsoft as a defendant.

Musk pushed back on what will be one of OpenAI’s key trial arguments: that his lawsuit was filed after the three-year statute of limitations.

He said despite his concerns about Altman and Brockman in 2018 and 2022, he didn’t believe at the those times that they would proceed with “stealing” the charity.

“Thinking that someone might steal your car is not the same as someone who has stolen your car,” Musk testified.

“I would have filed a lawsuit sooner if I had thought they had stolen a charity sooner.”

Cross Examination

Musk grew irritable and defensive under cross examination from Savitt, frequently saying that he couldn’t provide simple yes or no answers because Savitt was attempting to mislead the jury.

US District Judge Yvonne Gonzalez Rogers intervened during the questioning a handful of times and asked Musk to refrain from providing long-winded answers.

“Your questions are not simple,” Musk told Savitt. “They’re designed to trick me, essentially.”

“The classic reason why you cannot always answer a yes or no question: If you ask a question ‘have you stopped beating your wife?’” Musk said.

Gonzalez Rogers interjected, and told Musk: “We’re not going to go there,” drawing laughter from in the courtroom.

To contact the reporters on this story:
Madlin Mekelburg in Austin at mmekelburg@bloomberg.net;
Isaiah Poritz in San Francisco at iporitz@bloomberg.net

To contact the editors responsible for this story:
Ben Bain at bbain2@bloomberg.net;
Stephanie Gleason at saerts1@bloomberg.net

Seth Fiegerman, Peter Blumberg

© 2026 Bloomberg L.P. All rights reserved. Used with permission.