OpenAI co-founder and former chief scientist Ilya Sutskever testified Monday in the final stretch of Musk v. Altman that he holds an ownership stake worth about $7 billion in OpenAI's for-profit arm โ making him one of the largest known individual shareholders. He defended his role in Sam Altman's brief 2023 ouster, while criticizing his fellow board members for rushing the execution: "I felt a great deal of ownership of OpenAI. I felt like I put my life into it, and I simply cared for it, and I didn't want it to be destroyed." Greg Brockman's stake came out in the same proceeding: approximately $30 billion, disclosed publicly for the first time.
The financial disclosures are the headline material. Sutskever turned down a $6M/year compensation offer from Google to join OpenAI in 2015; he now holds the $7B stake. Brockman's $30B is the first public confirmation of his holdings. Microsoft has generated $9.5 billion in sales from OpenAI as of March 2025, through an agreement where OpenAI shares 20% of its revenue and Microsoft has priority over OpenAI's nonprofit owner on profits. A 2022 Nadella email surfaced in testimony: "Microsoft will lose 4 bil next year!!!" โ written to his lieutenants about the OpenAI partnership before the for-profit conversion accelerated. Musk's case rests on the allegation that Altman and Brockman violated commitments to keep OpenAI nonprofit; Sutskever's testimony actually undermines that โ he said OpenAI needed "a lot of dollars" to build computers "as big as the human brain," that becoming a for-profit was "the consensus way forward," and that Musk never negotiated any special promises when funding the original nonprofit.
What's new for the AI ecosystem here isn't the gossip โ it's that the financial-incentive picture sharpens enough that conflict-of-interest in AI safety governance stops being abstract. When superalignment leadership and equity ownership concentrate in the same people, the governance structure supposed to oversee AGI development can no longer credibly claim mission-only motivation. The more pointed item from Sutskever's testimony: he named the superalignment team's disbanding (May 2024, shortly after he left) as the loss of "the most important work at OpenAI for the long term." That's a public AI-safety figure with a $7B financial stake in OpenAI saying, on record under oath, that the company moved away from the safety work he believed mattered most. The trial isn't framed as an AI-safety case, but this testimony documents the shape of how lab-internal safety priorities were dropped under commercial pressure. Nadella's "amateur city" framing of the firing itself โ and Sutskever's own criticism that his board colleagues took "legal advice that wasn't very good" โ confirms that the execution went badly even from the people who agreed with the underlying call.
Altman testifies Tuesday May 12 โ his side of the story is the next big block. Musk's legal claim probably weakens on the for-profit-violation theory after Sutskever's testimony; Sutskever's framing instead strengthens a different Musk argument โ that Altman shouldn't be the one running an AGI lab. The financial disclosures (Sutskever $7B, Brockman $30B, Microsoft $9.5B + 20% revenue share + profit priority) are the kind of reference numbers that will reshape the next year of "who governs the labs" debate. For AI safety researchers, policy advocates, and anyone trying to figure out whether self-governance by frontier labs is structurally viable: this hearing is the citation. The math is now public.