1 hour ago
1
For two men whose companies promise to remake civilisation, the opening day was strangely pedestrian. The microphones in a federal courtroom in downtown Oakland kept cutting out. A presentation slide vanished from the screens. “We can’t hear you,” Judge Yvonne Gonzalez Rogers said, before drily observing that the federal judiciary would happily accept additional taxpayer funds.
Sitting under the fluorescent lights this week were Elon Musk, the world’s richest person, and Sam Altman, chief executive of OpenAI. Their tech empires are worth, between them, more than $US3 trillion ($4.2 trillion). Yet here they were, fidgeting through technical hitches a few metres apart, separated by years of broken friendship, as the most consequential trial of the AI age finally got under way.
Musk v Altman is, on its face, a contract dispute over a now decade-old start-up. In substance, it is a referendum on the architecture of the modern AI industry – its mission, its money, and the small constellation of men who insist they alone can shepherd humanity through the most powerful technology ever built.
It will most likely hold implications for AI going forward, and for Australia and other countries where the technology is tearing through workplaces and society more broadly.
Corporate litigation lawyer Andrew Staltman predicted the spectacle would be “crazy and nasty”, “the landing of the Hindenburg on the deck of the Titanic”. So far, that’s proven to be accurate.
The $US150 billion question is whether Musk can convince a nine-person jury – and ultimately Gonzalez Rogers – that Altman and OpenAI president Greg Brockman defrauded him by morphing the 2015 non-profit research lab into what he calls a “wealth machine” now valued at $US852 billion, with IPO ambitions of up to $US1 trillion. Musk wants last October’s for-profit conversion unwound. He wants Altman and Brockman gone.
“It is not OK to steal a charity,” Musk, in a black suit and matching tie, told the jury on Tuesday. If OpenAI prevailed, he warned, the case would set a precedent for “looting every charity in America”.
“I was a fool who provided them free funding to create a start-up. I gave them $38 million of essentially free funding to create what would become an $800 billion company.”
Elon MuskOpenAI’s lead counsel William Savitt told the jury that the trial was merely work of a jealous competitor. “We are here because Mr Musk didn’t get his way at OpenAI. He quit, saying they would fail for sure. But my clients had the nerve to go on and succeed without him.”
Savitt added, when ChatGPT exploded onto the scene in late 2022, “that’s when the sour grapes kick in.”
How we got here
The story, as Musk tells it, begins with Larry Page. The Google co-founder, he testified, once shrugged off the prospect of AI wiping out humanity, calling Musk a “speciesist” for caring about human survival.
“I thought, what would be the opposite of Google?” Musk testified. “An open source, non-profit.”
OpenAI was duly founded in 2015 as a charity “unconstrained by a need to generate financial return”. Musk seeded it with between $US38 million and $US44 million, recruited chief scientist Ilya Sutskever and came up with the name. Then it fell apart.
Court documents show Musk wanted to fold OpenAI into Tesla, as he put it in one 2018 email, “as a cash cow”. Later he sought majority control of a for-profit arm, only to be rebuffed. He left the board in 2018.
A year later, OpenAI launched a “capped-profit” subsidiary and took $US1 billion from Microsoft. Last October, the for-profit became OpenAI Group, a Delaware public benefit corporation. The non-profit kept a 26 per cent stake; Microsoft took 27 per cent. Musk filed suit in 2024, then folded his AI company xAI into SpaceX, valuing the combined entity at $US1.25 trillion.
Inside the courtroom
Musk’s first day on the stand was a peculiar mix of biography, ideology and grievance. It was a tour through Zip2, PayPal, SpaceX, Neuralink and Tesla, then a warning that AI “could make everyone prosperous, but it could also kill us all”. The Verge’s technology reporter Elizabeth Lopatto judged him “more petty than prepared”.
By Wednesday, the gloves came off. Musk delivered the line of the trial: “I was a fool who provided them free funding to create a start-up. I gave them $38 million of essentially free funding to create what would become an $800 billion company.”
Savitt then confronted Musk with his original 2015 pledge to donate $US1 billion to OpenAI. Did he actually contribute anywhere close?
“I contributed my reputation,” Musk replied. “These things all have value.” Pressed on tax breaks, on his own emails proposing a for-profit structure, and on his demands for majority control, he grew testy. “Your questions are not simple. They’re designed to trick me.”
When Savitt asked if Grok, an AI assistant built by xAI, lagged ChatGPT, Musk shot back: “Not any more.” The judge had to intervene more than once.
By Thursday, the judge had had enough of the metaphysics. Before Musk took the stand for a third day, she barred any discussion of AI’s existential risk.
“I suspect there are a number of people who do not want to put the future of humanity in Mr Musk’s hands,” she told the lawyers. “But we’re not going to get into that.” For Musk, whose team had been framing the case as a fight to save the species from a runaway charity, it was a real blow: the trial would be about the contract, and only the contract.
Savitt then played a video of Musk’s October deposition in which the billionaire had denied reading a key OpenAI document, contradicting Wednesday’s testimony. He drew out a flat denial under oath that Tesla was pursuing artificial general intelligence, despite Musk’s recent post saying the opposite. Brockman scribbled a note on a yellow pad and slid it to his legal team. Microsoft’s lead counsel Russell Cohen pursued a parallel line: if Musk had known about the OpenAI-Microsoft entanglement since 2020, why wait until 2024 to sue?
Musk’s family-office chief, Jared Birchall, also took the stand, to detail about 60 donations Musk had made to OpenAI along with a failed $US97.4 billion bid for the lab’s assets earlier this year. The bid, Birchall said, was meant to “prevent the diminishment of the value of the foundation”. Under cross-examination, he conceded he had co-ordinated it with one of Musk’s lawyers in this case. OpenAI asked Gonzalez Rogers to strike the testimony.
What is probably the most damaging evidence, on both sides, is on paper. Brockman’s 2017 diary, disclosed in pre-trial discovery, records his concern that taking the non-profit from Musk would be “morally bankrupt”, and that promising a non-profit while quietly building a for-profit was “a lie”. Microsoft chief technology officer Kevin Scott mused in a 2018 email about whether donors realised they were funding “a closed, for-profit thing”.
Yet Musk has his own paper trail. Internal emails show he once wrote that “it might have been a mistake for OpenAI to be set up as a non-profit”. He proposed making it Tesla’s cash cow. He demanded the right to choose a majority of the board.
When Microsoft poured another $US10 billion into OpenAI in 2023, Musk texted Altman: “What the hell is going on?” The deal felt like a “bait and switch”. Altman replied: “I agree this feels bad.” OpenAI later offered Musk equity. He turned it down. “Frankly, it felt like a bribe.”
What’s at stake
If the jury sides with Musk and Gonzalez Rogers concurs, the consequences could be vast. OpenAI is reportedly aiming to file for an IPO as early as the fourth quarter. A finding of liability – let alone unwinding the for-profit – would jeopardise that timing, threaten the equity of thousands of staff, and embolden regulators eyeing the Microsoft entanglement.
The wider stakes go beyond OpenAI. As The Atlantic’s Matteo Wong observed, the AI industry is now shaped by a remarkably small circle of men, each insisting they alone are the rightful steward of humanity’s most transformative technology. Anthropic was founded by ex-OpenAI staff members who didn’t trust Altman’s leadership. xAI was founded by Musk for the same reason.
For Australia – where the Albanese government recently signed a memorandum of understanding with Anthropic – the verdict will matter. UNSW Scientia Professor Toby Walsh, one of the country’s most prominent AI researchers, says the trial cannot be separated from the broader question of whether the industry can be trusted to police itself.
“The commercial pressures are immense,” Walsh says. “The stakes are millions, perhaps billions of dollars for people like Sam Altman.” His regulatory lesson is blunter: “We’ve learnt from social media that tech companies cannot be relied upon to comply.”
The case is arguably sidestepping the question that ought to be at its centre – what AI should be oriented towards, profit or public benefit – and is instead a fight so far dominated by personalities, and pettiness.
The trial resumes next week. Still to come: Altman, Brockman, Microsoft’s Satya Nadella and Tasha McCauley, the former board member who briefly ousted Altman in 2023. Deliberations are expected by mid-May; most of the verdict will be advisory, with Gonzalez Rogers deciding remedies.
Whatever she rules, the irony is unlikely to be lost on those watching: a courtroom drama, in plain English, may yet prove the most decisive check on the AI age.
At least for now.
The Market Recap newsletter is a wrap of the day’s trading. Get it each weekday afternoon.
