Musk Is Suing OpenAI for $134 Billion. The Trial Starts Monday. Here's What He's Actually Claiming.

Elon Musk donated $38 million to OpenAI in 2015. On Monday, his lawsuit demanding $134 billion in damages goes to trial.

That gap isn't a rounding error. It's the central argument. Musk's legal team claims the distance between those two figures measures how much he was defrauded, that OpenAI's founders accepted his early donation based on an explicit promise the company would remain a nonprofit, and then built a $340 billion commercial enterprise on the back of that promise. The openai lawsuit begins April 27, with jury selection starting at 8 a.m. in a federal courthouse in Oakland, California, before U.S. District Judge Yvonne Gonzalez Rogers.

The defendants are Sam Altman, Greg Brockman, OpenAI, and Microsoft. The expected witnesses include all three individuals plus Satya Nadella. The evidence includes a private diary entry in which one of OpenAI's co-founders wrote the words "it was a lie." The damages number, $134 billion, is what Musk's attorneys say represents the wrongful gains extracted from work he helped finance, and it is no longer money Musk is asking to keep for himself.

Here's what the case actually argues, why the evidence is harder to dismiss than it looks, and what the verdict could mean for the AI industry, regardless of who wins.

This analysis is based on Musk's amended federal complaint, court records from U.S. District Court for the Northern District of California, and reporting from [CNBC](https://www.cnbc.com/2026/01/08/musk-openai-altman-lawsuit-trial.html), [Bloomberg](https://www.bloomberg.com/news/articles/2026-04-07/musk-wants-openai-nonprofit-to-get-any-trial-winnings-from-suit), and [The Washington Post](https://www.washingtonpost.com/technology/2026/04/23/musk-altman-lawsuit-trial-openai/).

What Happened: A $38M Donation, a For-Profit Pivot, and a Federal Courtroom

OpenAI was founded in December 2015 as a nonprofit organization. Musk was one of its primary early backers, contributing approximately $38 million and serving on the board until his resignation in 2018. The founding agreement, as Musk's legal team characterizes it, was clear: OpenAI would develop artificial general intelligence for the benefit of humanity, operate without profit motive, and keep its research open rather than proprietary.

The company shifted direction. In 2019, OpenAI created a capped-profit subsidiary and accepted a $1 billion investment from Microsoft. The for-profit structure grew from there. Microsoft's total investment eventually reached roughly $13 billion. In the first quarter of 2026, OpenAI closed a $122 billion funding round, the largest in venture capital history, valuing the company at approximately $340 billion. ChatGPT became the fastest-adopted consumer application ever recorded.

Musk filed suit in 2024, alleging fraud, breach of contract, and violations of charitable trust law. His core claim: OpenAI's transition to a for-profit model wasn't a legitimate governance evolution. It was a betrayal of promises made specifically to secure his early funding and labor.

The most significant evidence came out of the 2025 discovery process. Greg Brockman kept a private diary. A 2017 entry, cited in Musk's complaint and referenced in Judge Gonzalez Rogers' summary judgment ruling, written while the company was internally debating its transition to a capped-profit structure, reads: "I cannot believe that we committed to non-profit if three months later we're doing b-corp then it was a lie." When OpenAI moved for summary judgment, asking the judge to dismiss the case before trial, Judge Gonzalez Rogers denied the motion. She specifically cited that diary entry as evidence a jury should evaluate.

Other discovery materials include a text message from Altman to Musk, "you're my hero", sent during the period when Musk was still a major donor and board member. There are also reportedly late-night texts from Satya Nadella, who will testify at trial. None of those communications were written with the expectation that a federal court would eventually read them.

Musk's April 2026 amendment to his complaint shifted the requested remedy. According to Bloomberg, he is no longer asking for $134 billion to be paid to himself. He is asking the court to direct any recovery entirely to OpenAI's charitable nonprofit arm, and to require that Altman and Brockman be removed from leadership. OpenAI called it a "legal ambush" in court filings, arguing it improperly expanded the scope of relief in the final days before trial began.

Why This Trial Is About More Than a Billionaire Dispute

The simplest frame for this case is two powerful men, Musk and Altman, who fell out and are now litigating their grievances. That frame is not wrong. It is also incomplete.

The trial's central legal question is whether founding promises made by AI companies carry enforceable weight. OpenAI was not alone in 2015 in positioning AI development as a charitable mission. Dozens of organizations, research labs, and early-stage companies in that period accepted funding, talent, and goodwill based on similar mission language. If a federal jury finds that such promises, even when made informally, even without written contracts, can support fraud claims when the company later changes course, the implications reach every organization that has ever raised money under the banner of "AI for humanity."

The IPO question is equally high-stakes. OpenAI has been preparing to go public, and the for-profit structure is a prerequisite. The $340 billion valuation is built on that structure. Microsoft's $13 billion stake depends on it. Amazon's $50 billion investment, announced at the beginning of 2026, assumes it continues. If the court supports Musk's request that OpenAI return to operating as an actual nonprofit, the entire capital table becomes legally uncertain, not just OpenAI's plans, but its existing investors' positions.

Microsoft's CEO testifying at trial is not a minor detail. Satya Nadella is not a defendant. He is a witness because he was present for key conversations about OpenAI's governance evolution, and because Microsoft's $13 billion investment was made while that evolution was underway. His testimony will be examined for what it reveals about how much Microsoft knew about the shift, and when they knew it. An investor aware that a company's governance structure was being contested while writing billion-dollar checks is a different legal situation than one who had no knowledge.

The broader industry implication is more abstract but more durable. Every AI lab that has ever described its work in terms of human benefit, every research organization that has attracted low-compensation researchers based on mission culture, every startup that has accepted donations or grants from individuals who believed in the nonprofit framing, all of them now have a case law benchmark for how those promises will be treated if the organization later changes course.

The Evidence That Could Define the Verdict

Fraud requires several elements: a false representation, made knowingly, with intent to deceive, that a plaintiff reasonably relied on, causing damages. Each element matters. Here is where the evidence sits on each.

The false representation. Musk claims Altman and Brockman promised OpenAI would remain a nonprofit, and that this promise was a material condition for his contributions. OpenAI argues no binding contractual promise was ever made in writing, and that the company's structure changed through a process reviewed and approved by the California attorney general.

The Brockman diary cuts directly against this defense. It does not capture Musk being explicitly deceived in a conversation. It captures Brockman himself, in a private document written in real time, describing the gap between what the company committed to and what it was doing as a "lie." Judges and juries react differently to private documents than to legal arguments crafted years later. Judge Gonzalez Rogers referenced that entry specifically in her summary judgment ruling.

Intent. For fraud, the false representation has to be made knowingly, not by mistake. OpenAI will argue its transition was a legitimate response to competitive and financial realities, carried out in good faith. The problem is the timeline: capped-profit discussions were underway within a few years of OpenAI's founding. Musk's attorneys will argue that the speed of the pivot, from "nonprofit for humanity" to "we might be doing a b-corp", is circumstantial evidence that the original framing was instrumental rather than sincere.

Reasonable reliance. A jury has to find it plausible that Musk donated $38 million because of the nonprofit framing specifically, not out of general interest in AI or a desire for influence. This is where OpenAI's counter-narrative about Musk's motivations becomes relevant. OpenAI has previously suggested that Musk left the board after being denied operational control, and that the lawsuit reflects competitive frustration, particularly given that Musk subsequently founded xAI, a direct competitor. The jury will have to assess those competing accounts of what Musk believed and why he acted as he did.

The amendment's strategic effect. Musk's decision to redirect any winnings to OpenAI's nonprofit arm is the most legally sophisticated move in the case. It dismantles the obvious counterargument, that a billionaire is suing for money out of spite or greed, and forces the dispute onto the merits. It also creates a coherent narrative: Musk is not asking to be compensated for a failed investment. He is asking for the organization to be restored to the purpose it said it had when he funded it. Whether a jury finds that framing credible will shape the verdict more than any single document.

OpenAI's strongest argument is procedural. Even if individuals made promises, those promises were not contractual commitments that survive a board decision and regulatory approval. California's attorney general reviewed parts of OpenAI's transition. The law generally permits nonprofit organizations to restructure under appropriate oversight. Whether attorney general approval forecloses a fraud claim against individual founders is precisely the question the trial will answer.

What Other AI Companies Have Done Differently

The clearest contrast is Anthropic, founded in 2021 by former OpenAI researchers. Anthropic operates as a public benefit corporation, a legal structure that writes mission obligations directly into the company charter, making them legally binding on the organization even through external investment. Anthropic's founding documents include safety commitments that are embedded in the company's legal structure, not just articulated in press releases. It has still raised over $30 billion in funding. But it did so through a container designed specifically to prevent governance drift of the kind now in dispute.

That structure wasn't accidental. Several of Anthropic's founders had worked at OpenAI and observed its evolution. The public benefit corporation structure was chosen, in part, to ensure that mission language would carry legal weight rather than aspirational weight.

Google DeepMind has never faced this problem because it has never operated as a nonprofit. It is a division of Alphabet, a publicly traded company. No charitable promise was ever made to funders. xAI, Musk's own company, was incorporated as a commercial entity from day one, accepted investor funding on explicitly commercial terms, and made no representations about operating for charitable purposes.

The industry implication is forward-looking. Organizations that want to accept mission-motivated funding, from donors, from researchers who work below market rates because they believe in the cause, from governments that fund AI safety research, now face a clearer choice: either be explicit that mission statements are aspirational and non-binding, or embed those commitments in legally durable structures. The informal middle ground, where companies describe charitable purposes while building commercial organizations, is now facing direct judicial scrutiny in a federal courtroom for the first time.

Four Weeks of Testimony, and What Comes After

Jury selection begins April 27. The trial is expected to run roughly four weeks. Musk, Altman, Brockman, and Nadella are each expected to testify. Each day of testimony is likely to produce new disclosures, specific communications, internal strategy documents, board minutes, that will simultaneously shape the legal outcome and the public understanding of how OpenAI was built.

Three outcomes are realistically possible.

The jury sides with Musk on the core fraud claims, and the court issues relief including a requirement that OpenAI return to nonprofit operation. This scenario freezes the IPO timeline, creates legal uncertainty for Microsoft and Amazon's existing stakes, and forces a restructuring of the most highly valued AI company in the world. It is the outcome OpenAI has organized its entire legal strategy around preventing.

The jury sides with OpenAI, finding that oral mission commitments don't create enforceable contracts and that the governance transition was legally executed under appropriate oversight. OpenAI proceeds toward IPO, Musk absorbs a legal defeat, and the industry takes note that mission language alone does not bind. This outcome is likely the one OpenAI's legal team considers most consistent with California charitable law.

A settlement reached during or after trial. Both parties have reasons to prefer a negotiated resolution, OpenAI avoids reputational and structural uncertainty, Musk achieves some public acknowledgment of the underlying issue. Settlements can happen at any point in a proceeding, even after testimony begins.

What won't change, regardless of verdict, is the industry's awareness that this question has been asked in federal court. Future AI organizations will face pressure to be more precise about the legal status of their mission commitments. Investors, employees, and regulators have been reminded, in the most public way possible, that "we're building AI for the benefit of humanity" is a statement that may eventually be held to account. For professionals following four weeks of daily disclosures, having a personal knowledge base to capture and search filings, testimony excerpts, and analysis as they accumulate makes the difference between following a story and understanding it.

The openai lawsuit that begins Monday is being called the trial of the century by some observers. That framing may be overconfident about the verdict's reach. What is not overconfident: for four weeks starting Monday, a federal jury will read Greg Brockman's private diary, listen to Sam Altman explain what he told Elon Musk in 2015, and decide whether the founding of the most valuable AI company in the world was built on a commitment that was later broken.

That question is worth watching, not because of who's involved, but because of what it asks about the industry they built.