By Deepa Seetharaman and Jonathan Stempel
April 27 (Reuters) – The bitter legal fight between Elon Musk and the leading artificial intelligence firm, OpenAI, led by Sam Altman, may come down to a few pages in one executive’s personal diary.
“This is the only chance we have to get out from Elon,” wrote Greg Brockman, OpenAI’s president and a co-founder, in the fall of 2017. “Is he the ‘glorious leader’ that I would pick?”
Brockman’s diary entry is part of the thousands of pages of internal documents revealed in court since Musk, one of the original co-founders of OpenAI, sued the company, its chief executive Altman and Brockman in 2024.
Musk is seeking $150 billion in damages from OpenAI and Microsoft, one of its largest investors, according to a person involved in the case, with proceeds going to OpenAI’s charitable arm.
Jury selection for the trial is planned for Monday in the Oakland, California, federal court, with opening arguments expected on Tuesday.
The documents offer a rare window into egos and personalities that have shaped OpenAI as it evolved from a nonprofit research lab in Brockman’s apartment to a tech giant worth more than $850 billion.
They also shed light on how the CEOs with the most power to shape generative AI think about the technology.
The trial risks complicating OpenAI’s plans for a potential initial public offering by casting doubt on its leadership. A drumbeat of unflattering disclosures could also intensify Americans’ growing pessimism about AI technology more broadly.
The case centers on Musk’s claim that OpenAI, Altman and Microsoft betrayed OpenAI’s original mission as a nonprofit to benefit humanity by forming a for-profit entity in March 2019, 13 months after Musk left the OpenAI board.
Musk said the defendants kept him in the dark about their plans, exploited his name and financial support to create a “wealth machine” for themselves, and owe damages for having conned him and the public.
He also wants OpenAI to revert to a nonprofit, for Altman and Brockman to be removed as officers, and for Altman to be removed from its board, among other measures.
OpenAI’s lawyers counter that Musk is motivated by a compulsion to control OpenAI and prop up his own AI lab xAI, which he founded in 2023 shortly after OpenAI launched ChatGPT and sparked the AI boom.
The company says Musk was involved in discussions to create OpenAI’s new structure and demanded to be CEO. Microsoft, also a defendant, denies that it colluded with OpenAI and says it teamed up with OpenAI only after Musk left.
HEAVY HITTERS EXPECTED TO TESTIFY
Heavy hitters in Silicon Valley including Musk, Altman and Microsoft CEO Satya Nadella are expected to testify in person. Shivon Zilis, a former OpenAI board member who is also mother to four of Musk’s children, is likely to be a key witness, with OpenAI lawyers arguing that she funneled information about OpenAI to Musk.
The trial comes at a sensitive time for both sides.
OpenAI faces unprecedented competition from rivals including Anthropic, and is spending billions on computational resources. It is also preparing for a potential blockbuster IPO that could value the company at $1 trillion, Reuters has reported.
Musk’s companies face similar pressures. His xAI, now folded into his rocket company SpaceX, trails far behind OpenAI in usage. SpaceX also plans to go public this year in what could be the biggest IPO ever.
According to court papers, Musk gave about $38 million of seed money to OpenAI between 2016 and 2020, mostly before he left the board.
In 2019, OpenAI restructured as a for-profit unit governed by the nonprofit. That let it accept money from outside investors while being accountable for the nonprofit’s original mission.
Last fall, OpenAI overhauled its structure again to become a public benefit corporation, in which the nonprofit and other investors including Microsoft hold stakes. The nonprofit holds a 26% stake as well as additional warrants if OpenAI hits certain valuation targets.
Musk’s lawyers calculated damages by multiplying OpenAI’s valuation and a portion of the nonprofit’s stake that could be attributed to Musk’s contributions. His team says between 50% and 75% of the nonprofit’s stake can be attributed to Musk.
A ‘MANHATTAN PROJECT FOR AI’
Musk and Altman co-founded OpenAI with a goal of developing AI to benefit humanity and fend off rivals such as Google.
Altman approached Musk about the idea in May 2015, branding it the “Manhattan Project for AI,” court documents show.
Musk’s involvement helped OpenAI land top researchers like now-former chief scientist Ilya Sutskever.
By mid-2017, Musk began questioning OpenAI’s viability, at one point holding back promised funds after clashing with Altman, Brockman and Sutskever, according to court filings. One source of tension was that Musk wanted to be CEO, emails show, which made other co-founders uneasy.
Around the same time, Brockman appeared frustrated by Musk’s stance, and wondered if turning OpenAI into a profit-making venture could also make him rich.
“Financially, what will take me to $1B?” he wrote in his diary. “Accepting Elon’s terms nukes two things: our ability to choose (though maybe we could overrule him) and the economics.”
Musk’s lawyers highlighted the entry to show that OpenAI’s leaders were more motivated by profit than the mission.
By January 2018, Musk appeared to have given up.
“OpenAI is on a path of certain failure relative to Google,” Musk emailed.
In late 2022, OpenAI launched ChatGPT.
(Reporting by Deepa Seetharaman in San Francisco and Jonathan Stempel in New York; Editing by Ken Li, Noeleen Walder and Nick Zieminski)
Comments