OpenAI quietly changed its governance structure during its 2025 for-profit transition to make it harder to fire CEO Sam Altman, according to court documents released by Elon Musk's expert witness on Sunday as part of the ongoing Musk v. Altman trial. Under the new bylaws adopted in October 2025, a two-thirds supermajority of the for-profit entity's nonemployee directors is now required to fire the CEO, compared to the previous structure that only required a simple majority (over 50%) of the nonprofit board. The analysis was conducted by Columbia law professor David M. Schizer, who took the witness stand last week and noted that under the new structure, Altman would only need two directors besides himself to remain CEO out of OpenAI's eight current directors (seven of which have voting power). The change puts OpenAI in the minority if it goes public, as proxy voting firm ISS reported in 2024 that the percentage of S&P 500 companies using supermajority rules had declined to just over one-third, with most governance organizations viewing supermajority voting as a tool of entrenchment rather than accountability.
