How to Unscramble an Omelet in Silicon Valley: The Musk v. Altman Trial That Will Try

Starting Monday in Oakland, a federal judge will consider Elon Musk’s claim that Sam Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit. At stake is not just $134 billion in potential damages, but whether it matters, legally speaking, that one of the most powerful AI companies in the world was built on a lie.

Musk and Altman co-founded OpenAI in 2015 as a nonprofit research lab, along with Greg Brockman, an AI researcher and entrepreneur, and others prominent in the field, but Musk left the company after a bitter falling out in 2018.

The following year, OpenAI established its first for-profit subsidiary, with investor returns capped at 100 times their investment. This structure would eventually evolve into the nearly trillion-dollar public benefit corporation OpenAI became in 2025. A public benefit corporation is essentially a for-profit company with a mission statement it’s legally required to consider, but not necessarily to prioritize.

This lawsuit, filed in 2024, originally alleged that Altman and Brockman ran a ‘long con,’  conspiring to enrich themselves at Musk’s expense.

On the eve of trial, in a move OpenAI called “evasive,” Musk’s lawyers voluntarily dismissed those personal fraud claims. What proceeds to trial today are two claims that go beyond Musk’s personal grievance: unjust enrichment and breach of charitable trust — essentially, the argument that OpenAI betrayed, not just Musk, but the public it promised to serve.

OpenAI argues Musk was fully aware the research lab needed to evolve beyond its nonprofit structure, because he participated in those early discussions, and even proposed folding OpenAI into Tesla. Now, OpenAI’s lawyers argue, Musk is disingenuously trying to use the courts to kneecap the most prominent rival to his own weaker and more controversial AI venture, xAI.

“Motivated by jealousy, regret for walking away from OpenAI and a desire to derail a competing AI company, Elon has spent years harassing OpenAI through baseless lawsuits and public attacks,” the company posted on its website, where it also offers a timeline that Musk v. Altman et al case watchers will find helpful as they follow what promises to be a barnburner of a trial.

Hundreds of court filings provide a dishy treasure trove of private communications worthy of a telenovela, including some juicy excerpts from Brockman’s personal journal.

He writes about Musk, “it’d be wrong to steal the nonprofit from him. … that’d be pretty morally bankrupt. and he’s really not an idiot.”

Also, “Financially, what will take me to $1B?”

But without a doubt, it is the beef between Musk and Altman that will dominate this show. “They really do not like each other. That part is not fake,” said Charlie Bullock, a senior research fellow at the nonprofit Institute for Law and AI who advises state and federal policy makers on AI governance topics.

Personal spite between Musk and Altman aside, Bullock said, “We’re going to learn a lot over the course of this case and from the conclusion of this case about whether the legal system can meaningfully constrain frontier AI labs.”

This trial, Bullock told KQED, is “sort of the fallback option” in the absence of other checks on bad behavior in the AI space, such as federal regulation.

This trial also promises to put on lurid public display a mini-universe of incestuous business relationships between men famous for rewriting rules rather than following them.

There is, for instance, a well-established law in California about nonprofits, for-profits, and how transitions between the two should be regulated. Whether and how it applies in this case is up to U.S. District Judge Yvonne Gonzalez Rogers in Oakland to determine over the next month.

OpenAI is like nothing that’s come before

Jill Horwitz, a law professor at Northwestern University and faculty director of the Lowell Milken Center for Philanthropy and Nonprofits at UCLA Law, likens OpenAI’s unique structure to “An enormous tail on a tiny dog.”

“The tail is the operating company, which is what everybody thinks of as being OpenAI, and the dog is the nonprofit, and it’s tiny. And it remains to be seen whether that board can be independent enough, because there’s such overlap between the nonprofit board and the for-profit board,” Horwitz said.

She is one of a small number of legal experts in this field who can speak freely about the case. Both sides asked her to serve as an expert witness, she said, and she turned them both down.

“It’s a weird structure. OpenAI isn’t one company. OpenAI is an interconnected group of companies. But it all is supposed to be advancing the nonprofit purpose,” Horwitz told KQED.

In 2018, even as OpenAI was privately contemplating the for-profit restructuring, it voluntarily adopted a new charter that restated and even strengthened its commitment to the public mission articulated at its founding.

In part, this had to do with the pressure Altman and OpenAI felt to attract top AI researchers, many of whom are concerned about the ethics of unleashing world-changing software on the rest of us. In 2024, 13 current and former OpenAI and Google DeepMind employees took the extraordinary step of publishing an open letter titled “Right to Warn,” calling out their own industry, and asking for protection if they warned the public.

“We are hopeful that these risks can be adequately mitigated with sufficient guidance from the scientific community, policymakers, and the public. However, AI companies have strong financial incentives to avoid effective oversight, and we do not believe bespoke structures of corporate governance are sufficient to change this.”

To this day, it remains unclear whether Altman’s talk about benefiting humanity was anything more than a savvy sales pitch designed to attract top AI talent and allay the concerns of federal regulators. This is one of the key questions trial watchers will be most keen to see answered.

“It’s quite typical for scientific research organizations to do all the hard work of the research before their IP is sold to a for-profit company for practical purposes,” said Rose Chan Loui, founding executive director of the Lowell Milken Center for Philanthropy and Nonprofits at UCLA Law.

What makes OpenAI unusual, Chan Loui said, is how explicitly and repeatedly the AI developer bound itself to promising its AI would be developed safely and for the benefit of all of humanity. “When they opened up to investment and formed the subsidiary, they recommitted to that purpose. They tied themselves even more tightly.”

Anthropic, founded by former OpenAI employees who left over concerns about the company’s direction, has cultivated a reputation as the more safety-conscious, ethically serious player in the AI race, the light gray hat to OpenAI’s dark gray one. Anthropic chose to incorporate as a public benefit corporation from the beginning, rather than a nonprofit, because a public benefit corporation has far more legal flexibility. “Anthropic may be behaving in a way that the public thinks is more charitable, but its legal duties to do so are a lot lower than OpenAI’s,” Horwitz said.

But is Musk the right party to bring this suit?

For legal eagles following this case, it’s curious that Musk is the plaintiff, rather than California’s attorney general, who is the primary legal guardian of charitable assets in the state, where most of OpenAI’s assets are located. But in 2025, Attorney General Rob Bonta negotiated a binding memorandum of understanding with OpenAI. The AG in Delaware, where OpenAI is incorporated, issued a parallel statement of non-objection.

A coalition of more than 30 California foundations and nonprofit organizations, including the San Francisco Foundation and TechEquity, urged Bonta to take immediate legal action to protect OpenAI’s charitable assets, arguing his office had both the authority and the responsibility to do so.

More than 50 organizations also petitioned Bonta to halt OpenAI’s for-profit conversion until he calculated the full market value of OpenAI’s nonprofit assets, estimated at the time at up to $300 billion, and directed OpenAI to transfer that value to independent nonprofit entities.

“It’s not too late for the Attorney General to revisit his agreement with OpenAI,” wrote Catherine Bracy, founder and CEO of TechEquity, an Oakland-based tech accountability organization. “The evidence this trial unearths, especially how OpenAI violated its original charitable mission in pursuit of profit, will likely leave him no choice.”

Chan Loui is among those scratching her head over a basic question: why does Musk get to bring this case at all? “He’s a competitor,” she said.

A personal fraud claim, that Altman lied to him to get his money, might have given Musk the clearest standing as an injured party. But Musk voluntarily dismissed those claims late last week. What remains rests almost entirely on a public interest argument, one that California’s attorney general, not a billionaire with a rival AI company of his own, would typically make.

Chan Loui worries about what it would mean if Judge Gonzalez Rogers effectively threw out that hard-won agreement between the attorneys general and OpenAI, essentially substituting a billionaire rival’s lawsuit for the state’s own regulatory process, whatever its deficiencies.

“You don’t want just anyone, any donor to complain,” Chan Loui said. “We have all this litigation against charities.” She said she sympathizes with those who want OpenAI to recommit as fully as possible to its original ethos, but she worries about what legal precedents this case could set for everybody else.

What’s not in dispute is that this trial will be a riveting spectacle for Silicon Valley, which will be watching this case with a mix of curiosity and fear. Judge Gonzalez Rogers has already proven she will rule against powerful tech companies when she determines the law demands it.

Also, the documents already unsealed suggest that what gets said in that Oakland courtroom may reveal a lot more about how Silicon Valley’s AI elite actually operates than anything previously said or posted in public.

“How much is OpenAI worth? Most of $1 trillion?” Bullock said. “There are ways that you could unscramble this omelet, but it would be extremely difficult, and it would be a massive headache for everyone involved.” He anticipates that whoever ends up on the losing end of this case will appeal.