Musk vs Altman OpenAI trial

The single most striking number in this trial is the one almost no one believes. Musk's damages expert, Berkeley Research Group's Dr. C. Paul Wazzan, has built a model that values Musk's roughly $44M in early contributions at 50–75% of OpenAI nonprofit's enterprise value, producing a disgorgement range of $79B–$134B. The legal-analytics firm Darrow has separately concluded that under conventional fraud principles, restitution would cap Musk's recovery at his roughly $38M donation plus interest — a figure on the order of $20M to $38M. That is not a rounding error. It is a 3,500x gap between what Musk is asking for and what a sober estimator thinks he can actually win.

Judge Yvonne Gonzalez Rogers has not been subtle about which side of that gap she sits on. She admitted Wazzan's testimony but openly told the courtroom that 'a jury is going to understand that he is pulling these numbers out of the air.' That is an extraordinary line for a presiding judge to deliver before opening statements, and it tells you how the remedies phase — which she alone decides — is likely to go even if the advisory jury hands Musk a liability win. The $134B headline is doing real work in the press, but inside the courtroom it functions less as a damages estimate and more as an anchor: a number large enough to make any settlement or structural concession look like a discount.

On April 24, three days before jury selection, Musk's team voluntarily dropped fraud and constructive-fraud claims, narrowing 26 counts in the November 2024 amended complaint down to two: unjust enrichment and breach of charitable trust. That is not a retreat. It is a choice of weapon. Fraud requires proving a knowingly false promise to Musk personally — a high bar, and one that would have invited brutal cross-examination about Musk's own emails, his role in pushing OpenAI toward a more aggressive funding model, and his subsequent founding of xAI. Charitable-trust doctrine bypasses all of that by asking a different question entirely: did the assets contributed to a California nonprofit get used in a way that breached the trust under which they were given.

California has unusually strict charitable-trust law, and outside experts have flagged it as potentially the strongest theory in the case. If Musk wins on that count, the natural remedy is not a dollar payout to him — it is structural. Disgorgement returns to the OpenAI nonprofit; the judge can order removal of trustees who breached duty (Altman and Brockman are named); and she can unwind transactions that improperly converted charitable assets into private gain. By dropping the fraud claims, Musk shed his weakest legal theory and his most personal liabilities, and concentrated his fire on the doctrine most likely to produce the structural relief he actually wants. That is a litigator's move, not an injured donor's.

The most quotable evidence in this trial is not a financial model. It is a 2017 diary entry by Greg Brockman, OpenAI's president and co-founder, in which he wrote: 'I cannot believe that we committed to non-profit if three months later we're doing b-corp then it was a lie.' Plaintiff's counsel will read that line to the jury, and there is no clean rebuttal. It is contemporaneous, internal, and from one of the defendants. Trial coverage describes it as a key piece of evidence on whether OpenAI's leadership knew the nonprofit framing was being abandoned in real time.

Musk, Altman, Brockman and Microsoft's Satya Nadella are all expected to take the stand, and that is where reputational risk compounds. Trial discovery has already surfaced internal artifacts including Altman emails, and a four-week trial gives plaintiffs' lawyers room to walk witnesses through every governance pivot from the 2019 capped-profit conversion to the 2023 quietly-amended 20% profit-cap escalator to the 2024 PBC proposal. For executives whose strategic narrative depends on portraying OpenAI's structure as principled and continuous, an extended public examination of the moments where that narrative bent is a problem independent of the verdict. The depositions are already on the record; the trial puts them in front of cameras.

If Musk wins disgorgement, the money does not go to him. It goes back to the OpenAI nonprofit. That is the structural feature of charitable-trust remedies and it is the fact that most quickly resolves the puzzle of why the world's richest man is suing over a $44M donation he made nearly a decade ago. The financial recovery is essentially symbolic for him personally. What is not symbolic is the ask attached to the dollars: removal of Altman and Brockman, and an unwinding of OpenAI's for-profit conversion.

OpenAI's defense team has named the alternative motive directly, calling the suit 'a harassment campaign that's driven by ego, jealousy and a desire to slow down a competitor.' That competitor is xAI, which Musk founded in 2023 and recently merged with SpaceX at a reported $1.25 trillion combined valuation. OpenAI itself listed Musk and the xAI litigation as risk factors in its own disclosures. Read together, the structural relief Musk is seeking — leadership change, governance unwind, plus expert testimony putting Microsoft on the hook for up to $25B — would land most heavily not on OpenAI's balance sheet but on its ability to operate at its current speed and capital structure. Whether the jury sees that as a principled defense of charitable assets or as competitive litigation will likely turn on how convincingly each side characterizes Musk's own emails from the founding era.

Even if the dollar figure ends up far below Musk's $134B headline, a liability finding on breach of charitable trust would reverberate well past OpenAI. A growing number of mission-driven tech entities — research labs, infrastructure projects, AI safety nonprofits — have explored or executed conversions from nonprofit to capped-profit to public benefit corporation structures, often with the same argument OpenAI made: that scaling the mission requires commercial capital. A California court ruling that those conversions can constitute a breach of trust to original donors would force every general counsel advising such an entity to redo the math on what donor consent, mission lock, and asset transfer actually require.

The second-order effect is on capital formation. OpenAI's own May 2025 retreat — keeping the nonprofit in control as the for-profit became a PBC after public backlash — already showed how much pressure these structures absorb when contested in the press. A judicial precedent attached to that pressure would change which lawyers approve which conversions and at what speed, and could push frontier-AI capital toward entities founded as for-profits from day one rather than nonprofit-to-for-profit hybrids. That is a quieter outcome than $134 billion in disgorgement, but it is the more likely one, and it is the reason this trial deserves attention from people who do not particularly care who wins between Musk and Altman.