The dossier ran 52 pages. It took roughly a year to assemble. And when its author finally acted on what it contained, the whole thing unraveled in less than a week.

Former OpenAI chief scientist Ilya Sutskever testified Monday that he spent approximately one year gathering evidence that CEO Sam Altman had displayed a “consistent pattern of lying” before joining the November 2023 board vote to remove him. The testimony came during the third week of Elon Musk’s trial against OpenAI and Microsoft — a proceeding that could reshape the most valuable AI company on Earth.

Sutskever confirmed that Altman’s alleged conduct included “undermining and pitting executives against one another.” He told the court he had discussed removing Altman with then-CTO Mira Murati after the two had been talking about his behavior “for a long time.” The document, assembled at the board’s request, was his case — methodical, documented, a full year in the making.

Then the execution fell apart.

‘Amateur city’

Microsoft CEO Satya Nadella, who also took the stand Monday, delivered a blunt assessment of the board’s handling. “It was sort of amateur city as far as I’m concerned,” Nadella testified, adding that he “never got clarity” from anyone about why Altman was fired — despite Microsoft being OpenAI’s largest investor at the time.

Sutskever largely agreed with the diagnosis. He criticized his fellow board members for rushing the process, lacking experience, and accepting “legal advice that wasn’t very good.” Within five days, he had reversed himself, voting to reinstate Altman because he feared the company would collapse. “I felt a great deal of ownership of OpenAI,” Sutskever told the court. “I simply cared for it, and I didn’t want it to be destroyed.”

The tension between the two testimonies is the governance story in miniature. If Sutskever’s year of documented concerns had merit, the board had a legitimate case — and fumbled the most consequential leadership decision in the short history of commercial AI. If the concerns lacked substance, a sitting CEO was briefly removed from the most important AI company on Earth based on grievances dressed up as governance.

The jury may never get a clean verdict on that question. What they have instead is a parade of former insiders, each offering a fragment of the same unhappy picture.

The insiders turn

Murati, who left OpenAI in 2024, testified in a video deposition that Altman had a pattern of “saying one thing to one person and completely the opposite to another person” and accused him of “creating chaos.” Former board member Helen Toner described a “pattern of behavior related to his honesty and candor.” Another former board member, Natasha McCauley, alleged Altman caused “repeated crisis events” through his leadership.

OpenAI has denied all of Musk’s allegations and characterized the lawsuit as harassment driven by personal jealousy. Current chairman Bret Taylor offered a counternarrative Monday, testifying that Altman has “been forthright with me and the other board members and grown OpenAI in ways that have exceeded my expectations.”

Altman is expected to take the stand as soon as Tuesday.

The money underneath

The financial stakes surrounding the trial are difficult to overstate. Sutskever disclosed that his ownership stake in OpenAI’s for-profit arm — valued at roughly $850 billion — is currently worth about $7 billion, making him one of the largest individual shareholders. Musk is seeking approximately $134 billion in damages, the removal of Altman and president Greg Brockman, and the dismantling of OpenAI’s for-profit structure, according to The Guardian. Closing arguments are set for Thursday before US District Judge Yvonne Gonzalez Rogers.

Nadella’s testimony also revealed the financial engine behind the partnership. Microsoft generated $9.5 billion in sales from OpenAI as of March 2025, including a 20 percent revenue-sharing agreement, according to Wired. Text messages presented in court showed Nadella pressing Altman to launch paid ChatGPT subscriptions in early 2023 — “sooner is best” — then following up weeks later about signup numbers.

Sutskever, who left OpenAI in 2024 to found AI startup Safe Superintelligence, appeared subdued on the stand. He barely made eye contact with anyone during his hour of testimony. The man who spent a year building the case against Altman now holds a $7 billion stake whose value depends entirely on the CEO he tried to remove.

Sources