OAKLAND, CA., 5/13/26 — OpenAI co-founder Sam Altman took the stand Tuesday at the federal courthouse in Oakland in the third week of the trial of Elon Musk’s lawsuit against Altman, Greg Brockman, and OpenAI.
Musk claims that his fellow OpenAI co-founders, Altman and Brockman, betrayed the company’s charitable charter by operating it for profit and their own personal enrichment instead of being a nonprofit developed “for the good of humanity.” Musk contends that they “stole the charity.”
Altman’s direct testimony on Tuesday supported OpenAI’s contention that Musk did not bring the case against them out of concern for humanity but because he was angry that he could not get personal control of OpenAI. On cross, Altman tried to fight off repeated accusations that he is dishonest, untrustworthy, and out for himself.
Altman’s direct testimony
William Savitt, OpenAI’s lead counsel, conducted the direct examination, beginning by asking Altman what he thought about Musk’s claim that Altman had “stolen the charity.”
Altman paused and shook his head before saying, “I found it difficult to wrap my head around.”
He said that the work they have done has been “hugely beneficial” and that he was “very proud of the work of the charity.”
Altman described himself as a “nerdy kid” who grew up interested in science fiction; he said when his family got a computer, “he spent a lot of time on it.”
Savitt walked Altman through events starting in 2015. Savitt moved quickly; the jury has heard much of the story in bits and pieces from other witnesses. Savit instead focused on key points for the defense.
Altman said that company’s success in the 2017 DotA2 competition — a popular multiplayer video game in which OpenAI’s software beat the top human players — led to the realization that they had to raise a great deal of money for additional computer resources. He said that was the only way to rapidly scale the capability of AI.
He described discussions among the founders in August and September 2017 about possible structures that would allow them to raise the money needed. He emphasized that Musk was fine with a for-profit subsidiary but that he wanted a majority of equity and control of the for-profit. He acknowledged that Musk said his control would be temporary, but Altman did not believe him. Altman said he didn’t think that AGI (artificial general intelligence) should be in the control of one person.
He testified that those discussions ended without agreement. He said that neither he nor the other founders made any promises to Musk that they would not pursue a for-profit subsidiary in the future.
Altman said that Musk cut off quarterly $5 million donations to OpenAI after he did not get his way. That left OpenAI with an extremely short financial “runway.”
Altman said that he and Brockman raised approximately $50 million from charitable investors, but it was not nearly enough for the hundreds of millions and billions needed to pursue the nonprofit’s mission.
Altman said that after Musk resigned from the OpenAI board in February 2018, the remaining board approved the establishment of a for-profit subsidiary and an offering of equity to investors. That allowed the company to raise funding to keep moving the company forward, but far short of the capital it wanted.
He described the relationship that OpenAI developed with Microsoft — a co-defendant in the case — that ultimately led to Microsoft’s investment of $13 billion. Through Microsoft’s investments, OpenAI gained enough computing capacity to create frontier AI models and create commercial products.
Altman testified that he kept Musk apprised of OpenAI’s plans until Musk formed a competing company — x.AI — that was also trying to develop AGI.
Musk, according to Altman, was fine with their plans, until OpenAI became successful with the release of ChatGPT.
Altman gave his assessment of where things stand today. He said he was enthusiastic about the growth in the value of the for-profit subsidiary, because the nonprofit (now renamed the OpenAI Foundation) has become one of the wealthiest charities in the world.
He testified that the Foundation has embarked on programs to address intractable health issues like Alzheimer’s and fund AI “resilience” research — things that can help protect people while AI technology advances.
Altman closed his direct examination with an observation about his OpenAI journey, saying, “If I had known how difficult and painful it would be, I would never have tried… But it was awesome,” he added.
He said he’s “extremely grateful” that he did because it is “wonderful to have created one of the largest nonprofits in the world.”
Altman under cross examination
Steven Molo, Musk’s lead lawyer, began his cross by asking Altman, “Are you completely trustworthy?”
After a pause, Altman said, “I think so.”
Molo continued in the same vein, asking, “Do you always tell the truth?” and whether Altman had “misled” or “deceived” those he worked with.
Altman parried the questions without giving direct answers, finally saying, “I believe I am an honest and trustworthy businessperson.”
Molo then listed the trial witnesses — Ilya Sutskever (co-founder and former chief scientist of OpenAI), Mira Murati (former OpenAI CTO and briefly interim CEO), Helen Toner (former board member) and Tasha McCauley (former board member) — who testified under oath that Altman lied or misled or deceived them. Molo asked Altman if he agreed with their assessments.
Altman said he was not aware of what they said but that he did not agree.
Molo continued the same line, citing alleged charges from other associates at OpenAI and companies where Altman had worked previously. He challenged Altman’s truthfulness in his 2023 testimony before the Senate Judiciary Committee when Altman said that he had no equity in OpenAI but did not disclose that he had indirect interest through another company in which he was invested.
After this line of questioning about Altman’s truthfulness, he got Altman to admit that when he and Musk discussed creating OpenAI, Musk was “one of the most respected innovators in the country,” while Altman had only modest success in his attempt to build a company.
Molo focused on the fact that during the September 2017 discussions, Altman was intent — “fixated,” according to Molo — on being the CEO of the nonprofit subsidiary they were discussing. Molo showed Altman an exhibit that contained a communication to Altman from Brockman and Sutskever. They said they did not “trust his judgments throughout the [negotiation] process,” and asked Altman, “Is AGI truly your primary motivation? How does it affect your political goals?” Molo asked what goals that referred to. Altman said that he had been thinking of running for governor of California.
Molo challenged Altman’s testimony that he kept Musk informed of the full scope of his plans to pursue investments from third parties and business partners. He assailed the many business relationships between OpenAI and companies in which Altman has an investment. Altman said his conflicts were fully disclosed and that he recused himself from transactions as appropriate.
The drama of Altman’s firing by the OpenAI board in 2023 provided another opportunity for Molo to challenge Altman’s truthfulness and lack of candor. Altman refused to concede he had been fired for lying to the old board, insisting that it was all miscommunication and a breakdown in trust.
Altman testified that he was angry and hurt by the firing, and at first did not want to go back, but because he loved the mission and what the company had built, he decided he was “willing to run back into the burning building to try to save it,” but only if his conditions for returning were met. One key condition was that the old board leave and a new board be appointed, one that was “stable” and “would work with me.”
Molo pointed out that Alman is currently CEO of both the foundation and the for-profit subsidiary.
Given that the foundation has oversight of the CEO of the for-profit subsidiary, Molo asked Altman if he thought he would be able to fire himself if he stepped out of line. Altman fought against the question for a while and then said, “I have no present intent of doing so.”
Throughout the trial, Altman and Brockman have maintained that creating a for-profit subsidiary was necessary for attracting the capital required to fund OpenAI’s massive computing needs. Molo challenged that stance, saying the nonprofit OpenAI could have raised the capital from philanthropic sources.
“You know that Stanford raised $3 billion, right?” Molo asked.
Altman countered by pointing to the value of the OpenAI Foundation’s ownership stake in the OpenAI Public Benefit Corporation.
“I do not believe I could get $200 billion into a nonprofit” through philanthropy, Altman said.
This story of part of a Bay City News special project following Elon Musk’s lawsuit against Sam Altman and OpenAI. The series explores the case’s ramifications for AI technology and the future of all humankind!