The $852 Billion Betrayal? Inside the Trial Exposing OpenAI's Founding Fractures

On April 28, 2026, a nine-person jury in the United States District Court for the Northern District of California began hearing testimony in what may be the most consequential corporate governance trial in artificial intelligence history. Elon Musk, who contributed roughly $45 million to found OpenAI as a nonprofit in 2015, is suing CEO Sam Altman, President Greg Brockman, and Microsoft for $130 billion in damages, alleging they converted charitable assets into private wealth [1][2].

The trial, expected to last approximately four weeks, has already produced a trove of evidence — personal diaries, frantic text messages, internal memos — that lays bare a decade of rivalries, power plays, and competing visions for the organization that built ChatGPT [3].

The Financial Arc: From Nonprofit to $852 Billion

OpenAI's valuation trajectory is central to the case. Founded in 2015 with modest nonprofit ambitions, the organization was valued at $29 billion in early 2023, jumped to $86 billion after the ChatGPT phenomenon in October 2023, surged to $157 billion by late 2024, and reached $852 billion following a March 2026 funding round [4][5].

Source: CB Insights, PitchBook, press reports

Data as of May 10, 2026CSV

That growth was fueled by successive rounds of capital that would have been impossible under a pure nonprofit structure. Microsoft invested $13 billion across multiple tranches beginning in 2019 [6]. By 2026, OpenAI had raised a cumulative $56.5 billion, including a $40 billion round in 2025 backed by SoftBank, Microsoft, and others [5].

Source: Crunchbase, press reports

Data as of May 10, 2026CSV

Musk's central contention is that this financial transformation was never authorized by the founding agreement — that Altman and Brockman enriched themselves by converting what was meant to be a public-interest research lab into a commercial enterprise [2].

The Founding Disputes: 2017 Negotiations

The trial's first two weeks focused heavily on a six-week period in August and September 2017, when the four co-founders — Musk, Altman, Brockman, and Ilya Sutskever — debated OpenAI's future structure [7].

Evidence showed the founders explored multiple options: converting entirely to for-profit status, creating a for-profit subsidiary, merging with Tesla, and even issuing cryptocurrency [7]. Musk, according to Brockman's testimony, was "not adverse to the creation of a for-profit subsidiary" but insisted on majority control [8]. When an equity-sharing proposal did not give Musk the largest stake, Brockman testified that Musk said "I decline," stood up, and "stormed around the table so aggressively that Brockman thought Musk was going to physically hit him" [1].

A 2017 diary entry from Brockman, entered into evidence, reads: "This is the only chance we have to get out from Elon" [9]. OpenAI's attorneys argue this shows the other founders felt constrained by Musk's desire for dominance rather than by disagreements over mission.

The Named Rivals and Their Stakes

The trial has identified several key figures whose personal conflicts shaped OpenAI's direction:

Sam Altman — CEO since 2019, architect of the commercial pivot. Former board members Helen Toner and Tasha McCauley testified that Altman displayed a "pattern of lying" and frequently "put words in other people's mouths" [10]. Altman's equity stake has not been publicly specified in trial documents, though OpenAI's corporate structure historically did not grant him direct equity — a fact that became contentious during the November 2023 board crisis.

Greg Brockman — OpenAI President, who testified his stake could be worth nearly $30 billion despite investing no personal cash [4]. His personal journals from 2017 and 2023 are now part of the trial record.

Ilya Sutskever — Co-founder and former chief scientist, who held roughly $4 billion in vested equity when OpenAI was valued at $29 billion in November 2023 [11]. At the current $852 billion valuation, that stake would imply a paper value exceeding $100 billion [11]. Sutskever departed in mid-2024 to found Safe Superintelligence (SSI), now valued at $32 billion [11].

Mira Murati — Former CTO who served briefly as interim CEO during the November 2023 crisis. Her testimony was among the most damaging to Altman: she stated under oath that Altman lied to her about whether a new AI model required safety board review before release [12].

Shivon Zilis — Former OpenAI board member and mother of Musk's children, whose testimony revealed that Musk offered Altman a seat on Tesla's board and that internal communications showed Musk was working to recruit OpenAI employees to his own ventures while still serving on OpenAI's board [13][14].

The November 2023 Crisis in New Light

Text messages entered into evidence reveal the chaotic hours of Altman's brief ouster. When Murati broke the news that both she and Altman were being replaced, she texted him: "They want a new ceo in place." The replacement was "rando twitch guy" — a reference to Emmett Shear [15]. Altman replied asking "more time for what?" and later expressed fear that the board wanted to transfer OpenAI's intellectual property to Anthropic [15].

Board members Toner and McCauley testified that their decision to fire Altman was based on information shared by both Murati and Sutskever regarding "Altman's leadership style, management decisions and resistance to oversight" [10]. Murati told the court that "OpenAI was at catastrophic risk of falling apart" under Altman's leadership and that she "was concerned about the company completely blowing up" [12].

The ouster lasted five days. Altman texted the board: "Still don't want me?" before his reinstatement [15]. The episode resulted in the removal of Toner, McCauley, and Sutskever from the board, consolidating Altman's control.

Safety vs. Commercialization: What the Evidence Shows

Murati's testimony provided the most specific allegation connecting internal conflicts to safety decisions. She stated that Altman told her OpenAI's legal department had determined a new AI model did not need safety board review before release. When Murati verified this with General Counsel Jason Kwon, "what Jason was saying and what Sam was saying were not the same thing" [12]. Asked directly whether Altman was telling the truth, Murati answered: "No" [12].

Board members who ousted Altman testified they believed he was "putting profits over safety" and warned this "could be disastrous" as the technology grew more powerful [7]. However, the trial has not produced evidence of a specific model release that was made measurably more dangerous by these internal disputes. The timeline of GPT-4's release (March 2023) and subsequent safety evaluations remains contested territory.

Independent AI safety researchers have noted the significance of the testimony without reaching consensus. Gary Marcus, a frequent AI critic, wrote that the trial evidence shows "what matters" is whether the governance structure actually functioned to enforce safety constraints — and the evidence suggests it did not [16].

OpenAI's Defense: Musk Wanted Control, Not Safety

OpenAI's legal strategy has centered on Musk's credibility and motives rather than defending the conversion on its merits [7][8]. Key defense arguments include:

• Musk himself proposed merging OpenAI with Tesla in 2017-2018 and wanted to be named CEO of the combined entity [8]
• Brockman testified Musk told him he wanted control of OpenAI partly to finance building a "city on Mars," which Musk estimated would cost $80 billion [8]
• Communications showed Musk imposed a "funding freeze" of $5 million in quarterly funding in August 2017 when negotiations weren't going his way [3]
• Musk attempted to settle the case days before trial, texting Brockman on April 25 to "gauge interest in settlement." When Brockman suggested both sides drop claims, Musk responded: "By the end of this week, you and Sam will be the most hated men in America" [17]
• Musk founded xAI, a direct competitor to OpenAI, in 2023 — raising questions about whether this lawsuit serves competitive rather than charitable interests [18]

The defense paints the picture of a co-founder who left when he couldn't control the organization and then sued when it became extraordinarily valuable without him.

Governance Structures: OpenAI vs. Its Peers

The trial has forced scrutiny of OpenAI's hybrid nonprofit-for-profit structure relative to competitors. OpenAI's current arrangement — adopted after negotiations with California Attorney General Rob Bonta in October 2025 — places the OpenAI Foundation (nonprofit) in control of OpenAI Group PBC (a public benefit corporation), with the Foundation holding 26% equity worth approximately $130 billion [19][20].

Anthropic, founded by former OpenAI researchers Dario and Daniela Amodei, operates as a public benefit corporation with a Long-Term Benefit Trust that holds governance authority — embedding safety constraints directly into corporate structure rather than relying on a separate nonprofit board [21]. xAI, Musk's own company, is a standard for-profit corporation with no nonprofit governance layer [21].

Google DeepMind operates as a division of Alphabet with published AI Principles but conventional corporate governance [21]. The trial has highlighted that OpenAI's structure was unique — and that its purported safety function may have been more theoretical than operational.

Investor Responses and Market Impact

Markets reacted sharply to trial revelations. On the day testimony began, SoftBank shares fell 10%, Oracle dropped roughly 5%, and CoreWeave fell 7%, all companies with major OpenAI-adjacent investments [4]. A Fortune analysis quoted a veteran tech analyst saying "everyone's overreacting" to the trial's impact on OpenAI's business fundamentals [4].

OpenAI itself flagged legal and governance risks in investor documents ahead of its expected Q4 2026 IPO, noting Microsoft's dominant role as both a risk factor and a dependency: "a substantial portion of our financing and compute" comes from a single partner [22]. The disclosure did not specifically address the trial but acknowledged that "lawsuits" from Musk and xAI posed material risks [22].

No major investor has publicly withdrawn from OpenAI. Microsoft's $30 billion stake remains intact, though negotiations over restructuring terms were described as contentious in multiple reports [6][23].

Regulatory and Legal Precedent

California Attorney General Rob Bonta conducted an 18-month investigation into OpenAI's conversion before reaching a settlement in October 2025 [19]. Under the agreement, the nonprofit Foundation retains nominal control, but critics argue the arrangement is "full of holes" — noting the for-profit entity could effectively override the nonprofit through various structural mechanisms [20].

The legal precedent most frequently cited in trial proceedings is the Blue Cross of California conversion in the 1990s, which resulted in more than $3 billion in stock being transferred to foundations when a nonprofit health insurer went for-profit [24]. If the jury finds a charitable trust existed, the remedy could range from monetary damages to structural injunctions that reshape OpenAI's governance.

Judge Yvonne Gonzalez Rogers, who previously presided over Epic v. Apple, indicated early in proceedings that if a formal charitable trust was established by Musk's donations, she would "be inclined to block OpenAI's effort to escape nonprofit control" [24]. This statement preceded the jury trial phase, which will address factual questions about the nature of the founding agreement.

What's at Stake

The financial exposure is staggering. Musk seeks $130 billion in damages [2]. If Sutskever's equity is marked to current valuation, his stake alone could exceed $100 billion [11]. Brockman disclosed his potential $30 billion interest [4]. For former employees who left amid internal disputes, the nonprofit-to-for-profit conversion determines whether their equity vests or is restructured — with billions of dollars in aggregate hanging on the outcome.

Beyond individual wealth, the case tests whether a nonprofit can incubate the most valuable technology of the era and then effectively privatize it. If Musk prevails, future AI labs may avoid nonprofit structures entirely — or regulators may impose stricter conditions on conversions. If OpenAI prevails, the hybrid model survives but with the understanding that nonprofit governance over AI development can be functionally dissolved when commercial interests reach sufficient scale.

The jury continues deliberations into week three. Ilya Sutskever and Microsoft CEO Satya Nadella are both scheduled to testify [11].