OpenAI CEO Sam Altman testified Tuesday in the Musk v. Altman trial, trying to make his core claim clear to the jury: He didn't steal the charity, but Elon Musk abandoned it.
Altman, wearing a blue suit and tie, spoke for about four hours on the witness stand in federal court in Oakland, California. He said Musk, who co-founded OpenAI with him in 2015, did not keep his promises and eventually abandoned the young startup as it tried to navigate an uncertain future.
"We were basically abandoned," Altman testified.
Musk sued OpenAI, Altman and company president Greg Brockman in 2024, accusing them of breaking promises to keep the artificial intelligence company a nonprofit and fulfill its charitable mission. Musk argued that about $38 million he donated to OpenAI was used for unauthorized commercial purposes.
Altman testified Tuesday that he made no promises to Musk about OpenAI's corporate structure.
Much of the trial, which began late last month, centered on a series of contentious negotiations in 2017 and 2018 between Musk, Altman, Brockman and another OpenAI co-founder, Ilya Suzkver.
Executives agreed on the need to raise more money for computing resources and discussed a variety of potential corporate structures, including for-profit options, to help them achieve that goal.
Negotiations ultimately broke down without a clear solution, and Musk left the OpenAI board in February 2018.
Altman testified that Musk's departure left OpenAI employees worried about how the company would be funded, with some worried that Musk would seek "retaliation." But Altman said Musk's exit was also a "morale boost" for some researchers because they had been "dampened" by Musk's management style.
"I don't think Mr. Musk understands how to run a research lab," Altman said.
Musk continued to communicate with Altman, Brockman and Suzkweil in 2018, months after he officially left the startup's board. He said OpenAI had no chance of success.
"My assessment of the probability that OpenAI can compete with DeepMind/Google without a dramatic change in execution and resources is 0%. Not 1%. I hope not," Musk wrote in an email in December of that year. "Even raising hundreds of millions of dollars won't be enough. This requires an immediate investment of billions of dollars a year, or else don't even think about it."
Altman said Tuesday that Musk's comments are "sealed in my memory."
After Musk left, OpenAI established a for-profit subsidiary that is currently valued by private investors at more than $850 billion.
Musk testified in April that OpenAI's for-profit subsidiary had become "the tail wagging the dog" and repeatedly accused Altman and Brockman of trying to "steal from a charity." Altman disputed that claim, testifying that what Musk really cared about was control.
Altman said Musk has a very strong desire to take full control of OpenAI, at least initially. Part of the reason, he said, is that Musk doesn't trust the decisions of others and that Musk made an "early decision" to only work at companies he controls.
"I feel very uncomfortable about it," Altman said.
Musk testified in April that he did initially seek majority control of OpenAI but that his stake in the company would dilute over time. He also said he's not completely opposed to OpenAI having a for-profit subsidiary, but it becomes a problem when it overshadows the nonprofit.
During the negotiations, Musk suggested merging OpenAI with his electric car company Tesla as a way to inject more capital into the company. He offered Altman a seat on Tesla's board to try to convince him to make the move.
Altman said he didn't think it was appropriate and feared the nonprofit would be essentially destroyed in the process.
"Tesla is a car company, it doesn't have the mission of OpenAI," Altman said. "I don't think we can ensure that the mission is carried out."
Musk lawyers try to paint Altman as untrustworthy
Musk's attorney, Steven Morrow, cross-examined Altman in an attempt to paint the OpenAI chief as unreliable and dishonest. He began his line of questioning by asking Altman if he was "completely trustworthy."
"I believe so," Altman said.
"But you're not sure you're completely trustworthy?" Morrow responded.
"I revised my answer to: Yes," Altman said.
Morrow asked Altman about several people who have expressed concerns about his behavior over the years, including Dario Amodei, a former OpenAI employee who later founded the company's main competitor, Anthropic. Morrow said Amodei had accused Altman of misrepresenting the terms of an investment to him.
"Dario accused me of a lot of things," Altman testified.
Morrow also pressed Altman on some of the board members who briefly removed Altman as CEO in 2023. The board said at the time that Altman "had not been consistently candid in his communications with them."
Altman spoke at length about his ouster during his testimony. He said he was "completely blindsided" and left frustrated, angry and hurt in the tumultuous days before his return to OpenAI.
He said he didn't get much explanation as to why he was fired, other than the board claiming he wasn't candid enough with them.
"I've spent the last few years of my life doing this," Altman said. "I watched it being destroyed."
Closing arguments in the trial are scheduled for Thursday, when the nine-person jury is likely to begin deliberations. The jury in this case is advisory, meaning the final outcome will be decided by Judge Yvonne Gonzalez Rogers.