Elon Musk and OpenAI executives face intense questioning in high-stakes trial
A no-nonsense federal judge and top litigators are putting Silicon Valley's biggest egos under oath in a trial that could reshape how AI companies are governed.
Share
Add us on Google by Editorial Team May. 10, 2026Elon Musk took the witness stand in an Oakland, California, federal courtroom and called himself “a fool.” Specifically, a fool for funding OpenAI, the organization he co-founded in 2015, which has since become one of the most valuable AI companies on the planet.
The trial, presided over by Judge Yvonne Gonzalez Rogers, pits Musk against OpenAI in a dispute over whether the organization betrayed its original nonprofit mission when it pivoted toward a for-profit structure.
A courtroom where nobody gets a pass
Judge Gonzalez Rogers has intervened multiple times to correct Musk’s courtroom behavior, emphasizing procedural decorum. Musk’s testimony has been described as emotionally charged.
Greg Brockman, another OpenAI co-founder, has also faced intense questioning. The trial has featured two top litigators dissecting the organizational decisions that transformed OpenAI from a research nonprofit into a company now reportedly valued in the hundreds of billions.
The 2017 power play that started it all
Documentation presented during the trial revealed that in 2017, Musk attempted to take control of OpenAI. When co-founders denied him that opportunity, he withdrew his funding.
The trial has also surfaced evidence that Musk actively tried to recruit OpenAI researchers for his other ventures.
Musk left OpenAI’s board in 2018.
Why this trial matters beyond the courtroom
OpenAI was founded with the idea that artificial general intelligence should benefit humanity broadly. The organization now operates with a capped-profit subsidiary, has taken billions from Microsoft, and competes head-to-head with Google, Anthropic, and Musk’s own xAI.
Musk didn’t just walk away from OpenAI. He went on to launch xAI, a direct competitor building its own large language models.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.