Altman Testifies Musk Sought to Transfer OpenAI Control to His Children

In a federal courtroom in Oakland, California, two of the most powerful figures in technology are locked in a dispute over something that began as a charitable venture and became one of the most valuable companies on Earth. The Musk v. Altman trial, now in its third week before Judge Yvonne Gonzalez Rogers, is ostensibly about $38 million in donations. But the real stakes — control over artificial general intelligence, the integrity of nonprofit law, and a company valued at $852 billion — are far larger.

"You Can't Just Steal a Charity"

Elon Musk's central argument is simple to state and complex to prove. He claims that the roughly $38 million he donated to OpenAI between 2015 and 2018 formed a charitable trust requiring the company to remain a nonprofit in perpetuity [1]. Of the 26 claims Musk originally filed in 2024, only two survived to trial: breach of charitable trust and unjust enrichment [2]. Musk is seeking more than $130 billion in damages and structural remedies, including the unwinding of OpenAI's 2025 conversion to a public benefit corporation [3].

"You can't just steal a charity," Musk told the jury repeatedly during three days of testimony in the trial's first week [4]. His attorneys argue that Altman and OpenAI president Greg Brockman used his donations for "unauthorized commercial purposes," enriching themselves and investors like Microsoft while abandoning the mission that attracted Musk's funding in the first place.

The "Hair-Raising Moment": Musk, Control, and His Children

Sam Altman's testimony on May 12 delivered what may be the trial's most vivid image. Altman told the jury that in early negotiations over OpenAI's structure, Musk proposed taking "total control" of any for-profit OpenAI entity, with a promise to reduce that control over time [5]. When co-founders asked what would happen to his controlling stake if he died, Musk suggested it could pass to his children — a response Altman described as a "hair-raising moment" [6].

"I don't think any one person should have control of the world's first AGI," Altman testified, referencing an email he wrote to co-founders at the time. "In fact, the reason we started OpenAI is so that wouldn't happen" [5]. Altman said Musk initially proposed holding 90 percent of OpenAI's equity, a figure other co-founders considered unacceptable [7].

This testimony is supported by emails entered into the court record. Altman also told the jury he was "extremely uncomfortable" with the idea of Musk serving as CEO, and that he doubted Musk would voluntarily relinquish power over a successful company based on his experience with startup founders [8].

Musk's attorneys have not directly disputed that these conversations occurred but have framed them as normal negotiations over governance structure, not evidence of authoritarian intent.

The Documentary Record: Emails, Diaries, and Text Messages

The trial has produced an unusually rich set of internal communications. Several key documents have been entered into evidence:

Greg Brockman's diary. OpenAI's president was compelled to read his private journal entries aloud in court. In a November 2017 entry, Brockman wrote: "It'd be wrong to steal the non-profit from him. To convert to a b-corp without him. That'd be pretty morally bankrupt" [9]. He also acknowledged, "Can't see us turning this into a for-profit without a very nasty fight" [9]. These entries cut both ways — they suggest the founders knew conversion was contentious, but also that they were wrestling with the ethics of excluding Musk.

Altman's 2017 email to Musk. In an email introduced by Musk's attorneys, Altman wrote, "I remain enthusiastic about the non profit structure!" — a statement Musk's legal team uses to argue there was a binding commitment to the nonprofit model [10].

Shivon Zilis's text messages. Zilis, a former OpenAI board member and mother of several of Musk's children, testified with text messages and emails showing that Musk, while still on OpenAI's board, was working to recruit top talent away from the company to Tesla [11]. Her testimony also revealed emails from 2017 and 2018 in which the idea of converting OpenAI to a for-profit was discussed — with Musk's involvement [11].

Musk's pre-trial text to Brockman. Days before the trial began, Musk texted Brockman: "By the end of this week, you and Sam will be the most hated men in America. If you insist, so it will be" [9]. The message was introduced as evidence of Musk's combative posture.

Following the Money: From $38 Million to $852 Billion

The financial trajectory of OpenAI is central to understanding both sides' motivations. When Musk departed the board in 2018, the company had raised modest sums and had no commercial products. Today, the numbers tell a different story.

Source: Visual Capitalist / Sacra

Data as of May 1, 2026CSV

OpenAI's valuation surged from approximately $1 billion at the time of its 2019 capped-profit restructuring to $29 billion after the launch of ChatGPT in late 2022, $157 billion in October 2024, and $852 billion following a $122 billion funding round in March 2026 [12][13]. This appreciation — roughly 850x in seven years — is at the core of Musk's claim that charitable assets were diverted for private gain.

Musk's $38 million in donations, while substantial at the time, represents a vanishingly small fraction of OpenAI's total capital. Microsoft alone has invested $13 billion, SoftBank contributed $40 billion in a single round, and total outside investment now exceeds $122 billion [14].

Source: CNBC / OpenAI

Data as of May 1, 2026CSV

Musk's legal team argues the size of the appreciation proves the diversion was massive. OpenAI's attorneys counter that Musk's donations were spent as intended — on research — and that the value was created by subsequent investment and commercial success that Musk chose not to participate in.

The Competitive Overlap Problem

Musk's credibility as a guardian of OpenAI's charitable mission is complicated by his own commercial interests in AI. His company xAI, founded in July 2023, directly competes with OpenAI in building large language models. During trial testimony, Musk admitted that xAI "partly" distills OpenAI's models [4].

The competitive entanglements run deeper. Text messages introduced at trial showed that while still on OpenAI's board, Musk directed efforts to recruit OpenAI employees to Tesla, with at least 11 former Tesla employees subsequently joining xAI [11][15]. Tesla's annual filing revealed that xAI purchased $430 million in Tesla Megapack battery systems in 2025, and Musk redirected GPUs initially reserved for Tesla to support xAI's data centers [15].

Tesla's own AI division works on Full Self-Driving and the Optimus robot program. Neuralink pursues brain-computer interfaces. All four Musk ventures — xAI, Tesla AI, Neuralink, and SpaceX's AI applications — compete with OpenAI for talent, compute resources, and government contracts [15][16].

OpenAI's attorneys have pressed this point aggressively, arguing that Musk's lawsuit is less about protecting a charity than about hobbling a competitor. Altman testified that Musk's departure from the board was a "morale boost" for researchers who felt "demotivated" by his management style. "I don't think Mr. Musk understood how to run a good research lab," Altman said [6].

Nonprofit Law Under the Microscope

The legal questions at the heart of this case extend well beyond Musk and Altman. Under California nonprofit law, the doctrines of private inurement and private benefit prohibit a charity's earnings from flowing to insiders or disproportionately enriching a private company [17]. The question is whether OpenAI's structural evolution violated these principles.

OpenAI was founded in 2015 as a pure nonprofit. In 2019, it created a "capped-profit" subsidiary — a novel structure in which investor returns were limited to 100 times their investment, with the nonprofit retaining control [18]. In 2025, OpenAI restructured again, converting the for-profit arm into a public benefit corporation (OpenAI Group PBC) while renaming the nonprofit the OpenAI Foundation. The Foundation now holds a 26 percent stake in the PBC, Microsoft holds 27 percent, and employees and other investors hold the remaining 47 percent [18].

The University of Maryland's Robert H. Smith School of Business has called the trial a spotlight on nonprofit law, noting that the case raises questions about whether the extraordinary circumstances required for a nonprofit to abandon its purposes actually existed [19]. UCLA Law professors Rose Chan Loui and Jill Horwitz have argued that "the law allows abandoning nonprofit purposes governing its assets only under extraordinary circumstances, which don't exist here" [17].

More than 30 individuals — including Nobel laureates, former OpenAI employees, and legal scholars — submitted a 25-page letter to the attorneys general of California and Delaware asking them to block the restructuring [17]. Among the signatories was Geoffrey Hinton, sometimes called the "godfather of AI," and Harvard legal professor Lawrence Lessig.

The Steelman Case for Musk

Strip away Musk's personal grievances and competitive interests, and a substantive argument remains. Several nonprofit-law experts and former OpenAI insiders have argued, independently of Musk, that the for-profit conversion represents a genuine betrayal of OpenAI's charitable mission.

The strongest version of this argument, articulated in the letter to attorneys general, holds that "the proposed restructuring would eliminate essential safeguards, effectively handing control of, and profits from, what could be the most powerful technology ever created to a for-profit entity with legal duties to prioritize shareholder returns" [17].

Public Citizen, a consumer advocacy group, has argued that the restructuring payments must go to a new, independent charitable enterprise — not one controlled by the same people who run the for-profit [20]. Critics point out that while the OpenAI Foundation nominally appoints all board members of the PBC, the overlap in leadership and the financial incentives of a $852 billion company could easily compromise the Foundation's independence [21].

Bloomberg Law published an analysis concluding that OpenAI's conversion "sheds nonprofit purpose without justification," noting that the company has not demonstrated the kind of financial distress or mission impossibility that typically justifies such a transformation under California law [22].

These arguments have force regardless of whether Musk is the right plaintiff to bring them. As one legal scholar noted in a Lawfare analysis, the case raises the question of "why OpenAI's corporate structure matters to AI development" — not just to Musk's portfolio [23].

The Steelman Case for OpenAI

OpenAI's defense rests on three pillars. First, the company argues that Musk was aware of and participated in discussions about converting to a for-profit model before he left the board. Emails from 2017 and 2018, introduced by both sides, show that the for-profit idea was in active discussion during Musk's tenure [11][10].

Second, OpenAI contends that the shift was necessary because building advanced AI requires capital at a scale no nonprofit can generate. The company's compute costs alone now run into billions of dollars annually, and attracting top researchers requires competitive compensation [14].

Third, OpenAI argues the new structure preserves the charitable mission through the Foundation's governance rights. The Foundation appoints all PBC board members and can replace them at any time. The 26 percent stake — worth approximately $130 billion at current valuation — would make it one of the most well-endowed philanthropic organizations in the world [18]. California Attorney General Rob Bonta reviewed the restructuring and signed off on the arrangement [21].

Precedent and the Broader AI Landscape

The outcome of this trial could shape how AI organizations structure themselves for years to come. The case has already attracted attention from other nonprofits weighing commercial conversions.

Historical precedent offers some guidance. When Blue Cross of California transferred assets to a for-profit subsidiary in the 1990s, it gave more than $3 billion in stock to two foundations. When Health Net converted from nonprofit to for-profit, it established the California Wellness Foundation with $300 million in cash and 80 percent of the new company's stock [21]. OpenAI's arrangement — giving the Foundation 26 percent of a company valued at $852 billion — is larger in absolute terms but arguably less protective in relative terms.

California Attorney General Bonta has signaled that his office will continue to monitor the arrangement. A coalition of more than 50 organizations, led by the San Francisco Foundation, has called on Bonta to protect "up to $300 billion of OpenAI's charitable assets from a for-profit takeover" [24]. Whether Bonta's approval of the restructuring will survive legal challenge — and whether Musk's lawsuit provides an additional lever — remains an open question.

What Comes Next

Closing arguments are scheduled for Thursday, May 15. Because the case is being tried in equity rather than at law, Judge Gonzalez Rogers will issue the binding ruling, though she has indicated she will weigh the nine-person advisory jury's recommendations [3]. A decision is expected by mid-May.

If Musk prevails, the court could order the unwinding of OpenAI's corporate restructuring — a remedy that would send shockwaves through the AI industry and Silicon Valley's investment ecosystem. If OpenAI prevails, the verdict would affirm that nonprofit boards have broad discretion to adapt their structures to changing circumstances, potentially encouraging other research organizations to pursue similar conversions.

Either way, the trial has already accomplished something unusual: forcing a public accounting of how a small research nonprofit became one of the most valuable private companies in history, and whether the people who built it honored the promises that got it started.

Microsoft CEO Satya Nadella, who testified on May 11, told the court that Musk never raised concerns about OpenAI's direction directly with him during the years of Microsoft's deepening partnership [25]. Altman, for his part, told the jury he could not comprehend Musk's framing: "It feels difficult to even wrap my head around that" [2].

The jury, the judge, and the broader public will soon have to decide which version of this story they find more convincing. The answer will matter far beyond this courtroom.