Musk-Altman Trial Surfaces Allegations About OpenAI Governance and Business Practices

After three weeks of testimony, closing arguments, and hundreds of internal documents entered into evidence, a nine-person federal jury in Oakland began deliberating on May 19 whether OpenAI's transformation from a nonprofit research lab into the world's most valuable AI company constitutes a breach of charitable trust [1]. The advisory verdict — which Judge Yvonne Gonzalez Rogers can accept, reject, or modify — could trigger up to $134 billion in disgorgement and the removal of CEO Sam Altman and president Greg Brockman [2].

The case, formally Musk v. Altman et al., has become more than a dispute between two billionaires. It is a stress test for California nonprofit law, a window into how Silicon Valley governance actually operates, and a potential precedent that could reshape the legal architecture of AI development for decades.

The Founding Promise and Its Paper Trail

OpenAI was incorporated in December 2015 as a 501(c)(3) nonprofit. Its founding charter committed the organization to developing artificial general intelligence (AGI) "for the benefit of humanity" and sharing its research openly [3]. Elon Musk and Sam Altman served as co-chairs; other founding board members included Ilya Sutskever, Greg Brockman, and several prominent AI researchers.

Musk's legal team introduced early communications showing the contours of a founding bargain. In 2015 emails disclosed during discovery, Musk and Altman discussed a structure "so that the tech belongs to the world via some sort of nonprofit, but the people working on it get startup-like compensation if it works" [4]. Musk's attorneys argued these communications constitute a binding agreement — or at minimum, a charitable trust obligation — that the nonprofit structure would remain paramount.

OpenAI countered that no formal written contract between Musk and the organization ever existed specifying permanent nonprofit status. Altman testified that "he never promised Musk to keep company a nonprofit" and that the structure was always understood to be subject to evolution as capital needs grew [5].

The Capped-Profit Conversion and Who Benefited

In 2019, OpenAI created a "capped-profit" subsidiary, OpenAI LP, where investor returns were limited to 100x their original investment. The nonprofit remained the sole controlling shareholder, with a majority of board members barred from holding financial stakes in the for-profit entity [6].

Then in October 2025, OpenAI completed a full restructuring into a Public Benefit Corporation (PBC) called OpenAI Group PBC. The newly formed OpenAI Foundation retained approximately 26% of the company's equity — an amount worth roughly $130 billion at the pre-restructuring $500 billion valuation, and now worth approximately $221 billion at the current $852 billion valuation [7].

The individual financial beneficiaries are at the center of Musk's case. Greg Brockman's equity stake in the for-profit entity is worth approximately $30 billion, according to evidence presented at trial. Musk's attorneys highlighted a personal diary entry in which Brockman weighed proposals and wrote about how he could "get to $1 billion for himself" — he ultimately received a stake worth thirty times that amount [4][8].

Microsoft invested approximately $13 billion across three rounds between 2019 and 2023, securing a 26.79% ownership stake that is now worth over $228 billion on paper [6]. The removal of the profit cap during the PBC conversion eliminated any ceiling on Microsoft's returns.

Source: CB Insights, TechCrunch

Data as of May 1, 2026CSV

The Internal Communications

The trial surfaced hundreds of previously private communications. Among the most significant:

The Brockman Diary (2017): A personal journal entry reading "This is the only chance we have to get out from Elon" was presented as evidence that founders viewed Musk's departure as an opportunity to restructure [4].

Musk's Threat Email: An email from Musk to Brockman stating "By the end of this week, you and Sam will be the most hated men in America. If you insist, so it will be" was entered into evidence, though the precise context remains disputed [4].

Early Governance Warnings: Within a year of founding, co-founders Brockman and Sutskever warned in internal emails that the proposed governance structure would give Musk "unilateral absolute control over the AGI" and that "once the company was on track toward AGI, the company will be much more important than any individual" [4].

The 2023 Senate Discrepancy: Musk's attorneys challenged Altman over his 2023 testimony before the Senate Judiciary Committee, where Altman stated he had "no equity in OpenAI" without disclosing indirect financial interests through companies in which he was invested [8].

Self-Dealing Allegations

Musk's legal team pressed Altman on business relationships between OpenAI and companies in which Altman holds personal investments — including payments company Stripe, chip startup Cerebras, and fusion energy company Helion, where Altman held a significant stake and served as board chair until recently [1][8].

A structural issue compounds this concern: Altman serves as CEO of both the nonprofit foundation and the for-profit subsidiary. Since the foundation theoretically oversees the for-profit's chief executive, opposing counsel questioned whether Altman could realistically terminate his own employment. Altman acknowledged having "no present intent of doing so" [8].

Altman testified that his conflicts were "fully disclosed" and he "recused himself from transactions as appropriate" [8].

The Legal Standard Under California Law

California nonprofit law holds that all assets of a corporation organized for charitable purposes are deemed held in trust for that declared purpose. In a conversion, these assets must be "disgorged and devoted to a charitable purpose" [9].

The most relevant precedent is the Blue Cross of California case from the 1990s, where the state intervened when the health insurer transferred assets to a for-profit subsidiary. The California Supreme Court required distribution of over $3.2 billion — the entirety of the nonprofit's assets — to charitable healthcare foundations [9].

Some legal scholars have argued that OpenAI's conversion should require a minimum payment of $30 billion to charitable purposes, based on the control premium (typically valued at 20-30% of enterprise value) that the nonprofit holds over the for-profit entity [10].

Successful prosecutions of "mission drift" against tech nonprofits by state attorneys general remain rare, though the California AG's office has maintained an active investigation into OpenAI's conversion since January 2025 [11]. Meta sent a letter to Attorney General Rob Bonta in December 2024 arguing that permitting the conversion would set a precedent with "seismic implications for Silicon Valley" by allowing startups to enjoy nonprofit tax advantages until profitability [11].

The Board That Approved the Conversion

Of OpenAI's original founding board members, none remain in their original governance capacity. Musk departed in 2018. Ilya Sutskever left after the November 2023 board crisis. The board that reinstated Altman in November 2023 was itself reconstituted entirely — the previous members, including Helen Toner and Tasha McCauley, were replaced as a condition of Altman's return [12].

Source: OpenAI, TechCrunch

Data as of May 15, 2026CSV

Toner and McCauley — two of the board members who voted to fire Altman in November 2023 — testified for Musk's side during week two. They stated that Altman was "untrustworthy, withheld information from the board and sometimes lied" [1]. Seven of the nonprofit's current eight board members also serve on the for-profit's board. The nonprofit hired employees only a month before the trial started and performs work only in grant-making rather than AI research [1].

Twelve former OpenAI employees filed an amicus brief in April 2025 stating that OpenAI had "abandoned its nonprofit roots" and that Altman "was a person of low integrity who had directly lied to employees about the extent of his knowledge and involvement" in forcing departing employees to sign lifetime non-disparagement agreements [1].

The Steelman for OpenAI

OpenAI's defense rests on several arguments that deserve substantive consideration.

First, every major AI safety-focused organization has converged on similar hybrid structures. Anthropic operates as a public benefit corporation. Musk's own xAI is a for-profit company. DeepMind was acquired by Google (Alphabet) outright. The argument that nonprofit status is incompatible with frontier AI development is not unique to OpenAI — it reflects an industry-wide consensus that the capital requirements of training frontier models (now measured in tens of billions per training run) exceed what charitable donations can sustain [13].

Second, Musk left OpenAI's board voluntarily in 2018. Altman testified that after Musk's departure, the nonprofit was "left for dead" and required the for-profit restructuring to survive as a competitive research organization [14]. Musk then founded xAI as a direct competitor in 2023, raising the question of whether his lawsuit represents a principled governance challenge or a competitive maneuver.

Third, Musk himself admitted during testimony that xAI "distills" (learns from) OpenAI's models — an acknowledgment of competitive motivation that OpenAI's lawyers emphasized during closing arguments [15].

Fourth, Altman testified that no single person should govern AGI development, and that Musk had sought 90% equity and majority control of OpenAI — a characterization that, if true, suggests Musk's original vision was not purely charitable either [16].

How OpenAI Compares to Peer Institutions

Two prominent models offer contrast. The Allen Institute for AI (Ai2), founded by the late Paul Allen in 2014, operates as a fully nonprofit research institute. It releases model weights, training code, and training data publicly and has scaled competitively without a for-profit conversion [17].

The Mozilla Foundation offers a closer structural parallel: it wholly owns the Mozilla Corporation, a for-profit subsidiary. But Mozilla has maintained clear separation between foundation governance and corporate operations, added diverse board members with AI and community expertise, and retained approximately $1.4 billion in reserves without converting the foundation itself [18].

What differentiates OpenAI's path is the scale of capital raised ($13 billion from Microsoft alone), the removal of the profit cap, the dual-role CEO structure, and the foundation's reduced role to pure grant-making rather than research oversight [1][9].

What Happens If Musk Wins

The potential remedies are extraordinary. Judge Gonzalez Rogers retains sole authority over the final liability ruling and has opened a parallel remedies phase to consider: removal of Altman and Brockman from their positions; up to $134 billion in disgorgement from OpenAI and Microsoft; and unwinding OpenAI's 2025 PBC recapitalization [2].

Musk initially sought $150 billion in damages [19]. The practical consequences of unwinding a $852 billion company with thousands of employees, major government contracts, and a pending IPO are difficult to overstate. No expert witness testimony quantifying the full unwinding cost has been publicly reported, though OpenAI's forensic accountant testified during the trial [2].

Microsoft's $13 billion investment, now worth over $228 billion on paper, would face uncertainty. The companies recently restructured their partnership, with OpenAI capping Microsoft's revenue share payments — a deal premised on the legitimacy of the for-profit structure [6].

What Happens Next

The jury's verdict is advisory only. Judge Gonzalez Rogers will issue the binding ruling on liability. Even if the jury sides with Musk, the judge may fashion different remedies — or reject the verdict entirely. Even if the jury sides with OpenAI, the California Attorney General's separate investigation into the conversion remains active [11].

Musk left for China during closing arguments, prompting his lawyer to apologize for his absence — a moment that may or may not influence the advisory verdict [20]. The spectacle of the trial — with its diary entries, threatening emails, and mutual accusations of dishonesty — has overshadowed what remains a genuinely difficult legal question: when a nonprofit becomes valuable enough to attract billions in investment, who has the right to decide its future?