The Battle for Control of OpenAI
Court testimony in the second week of Musk v. Altman has laid bare a previously obscured chapter of OpenAI's history: Elon Musk's aggressive campaign to seize absolute control of the company in 2017, long before his departure in 2018. According to OpenAI president Greg Brockman's testimony, Musk demanded majority equity, the right to choose a majority of board members, and the CEO role during negotiations over converting the nonprofit into a for-profit entity. The conflict culminated in Musk storming out of a meeting, grabbing a painting of a Tesla that cofounder Ilya Sutskever had brought as a goodwill gesture, and later attempting to poach Sam Altman to lead a rival AI lab at Tesla.
The revelations, reported by MIT Technology Review from the courtroom, directly undermine Musk's stated justification for the $134 billion lawsuit: that OpenAI's leaders deceived him into donating $38 million by promising to keep the company a nonprofit for humanity's benefit. Instead, Brockman's testimony, supported by emails and journal entries, suggests that Musk was the primary driver of for-profit conversion and that he walked away only after failing to secure unilateral authority over artificial general intelligence (AGI). The outcome of the trial could restructure or even unwind OpenAI's 2025 conversion into a public benefit corporation, imperiling its IPO plans at a valuation approaching $1 trillion.
The Haunted Mansion Party and the Failed Equity Split
The turning point, Brockman testified, came in the summer of 2017 after OpenAI's Dota 2 AI defeated the world's best human players. Musk hosted a celebration at his "Haunted Mansion" near San Francisco, where actress Amber Heard served whiskey amidst confetti and cups. Under cross-examination by Brockman, Musk's email surfaced: "Time to make the next step for OpenAI. This is the triggering event." Weeks earlier, Musk had indicated that a major public achievement would signal the moment to create a for-profit entity to raise capital for building AGI.
Over the following six weeks, the cofounders held intense negotiations. Brockman and Sutskever proposed an equal equity split among the founding team. According to Brockman, Musk fell silent, then said "I decline" before standing up and storming around the table. "I actually thought he was going to hit me," Brockman told the jury. Musk grabbed the Tesla painting Sutskever had brought as a token, walked out, and the negotiations collapsed. Brockman later wrote in his electronic journal that accepting Musk's terms would mean handing "unilateral, absolute control, potentially, over the AGI." The journal was pulled up by Musk's lawyer Steven Molo, who attempted to paint Brockman as motivated by personal wealth—highlighting entries where Brockman asked "what will take me to $1B?" and later admitting a breakaway would feel "morally bankrupt."
Zilis Testifies: Musk Tried to Poach Altman for Tesla AI
Shivon Zilis, a former OpenAI board member and mother of four of Musk's children, took the stand and disclosed that as early as December 2017, Musk had concluded OpenAI was unlikely to build AGI. In an email to Zilis, Musk said he was pivoting to building an AI lab at Tesla. Zilis's testimony revealed that Musk asked Andrej Karpathy, an OpenAI scientist he had already recruited to Tesla, to "send a list of top OpenAI people to poach." More strikingly, Musk personally attempted to recruit OpenAI CEO Sam Altman to lead that prospective lab. The plan never materialized, but it underscores the depth of Musk's efforts to control AI development outside of OpenAI.
Zilis, who joined OpenAI's board in 2020 while also working at Tesla and Neurolink, admitted she did not disclose her romantic relationship with Musk to the board until 2022. Under questioning, she maintained her allegiance was to "the best outcome for AI for humanity." Meanwhile, former CTO Mira Murati and former board member Helen Toner testified via deposition about the brief firing of Altman in 2023, citing his alleged history of lying. Murati's text messages from that period, entered as evidence, showed Altman's frantic attempts to understand his ouster and regain control—a pattern that echoes the current trial's theme of governance instability.
A Critical Week Ahead for OpenAI's Future
The trial resumes next week with testimony from Ilya Sutskever and Microsoft CEO Satya Nadella. Closing arguments will follow, and a jury will deliver an advisory verdict to guide the judge's decision. The stakes are enormous: beyond the personal reputations of Musk, Altman, and Brockman, the verdict could force OpenAI to restructure its corporate governance, potentially undoing the public benefit corporation status that enabled its exponential growth. The company's IPO, expected within months, hangs in the balance. Meanwhile, Musk's own conglomerate—xAI, now a division of SpaceX—is targeting a $1.75 trillion valuation in its own public offering as early as June.
The testimony has already shifted the narrative from a simple nonprofit breach to a complex saga of ambition, control, and high-stakes competition. For the AI community, the case raises fundamental questions about how AGI development should be governed. If a founder like Musk can use courts to retroactively reshape a company's structure, the precedent could chill investment in AI startups and delay the deployment of transformative technologies. Yet if the court finds that OpenAI's leaders acted in bad faith, the industry may see increased regulatory scrutiny and demands for transparency in nonprofit-to-profit conversions. Whatever the outcome, the trial has stripped away the veneer of altruistic mission statements, revealing the raw human drives that power the race to superintelligence.
Comments