Two weeks of trial testimony in Oakland have made one thing clear: Musk v. Altman is a story that is much larger than its plaintiff. The case has been framed in headlines as a personal feud — the world’s richest man suing his former co-founder over the AI company they built together — but the legal questions it raises are about something more consequential than ego or rivalry. They are about whether a charitable mission, established in informal emails and donor commitments in 2015, can legally bind a frontier AI company in 2026. Whether the corporate structures used by every major Western AI lab are genuinely accountable to the public-benefit principles they were founded on. Whether $850 billion of value created at the frontier of artificial intelligence belongs to the people who funded its creation as a charity, or to the people who built it into a commercial business.

What makes this trial unusually instructive — beyond the inevitable theater of Elon Musk under cross-examination — is that the discovery process has dragged the entire history of OpenAI’s corporate evolution into open court. Internal emails. Diary entries. Texts between co-founders. Board meeting records. Compensation structures. The kinds of documents that AI labs rarely have to disclose, presented to a federal jury and broadcast to a media following that has not been this attentive to AI governance since the November 2023 Altman firing. Even if Musk loses — and most legal analysts believe he will — what the trial has revealed about how OpenAI actually operates will shape how investors, regulators, and the rest of the industry think about AI lab governance for years to come. Let’s walk through it.

What is the Musk vs. Altman trial actually about?

The legal claims at the center of the trial are technically narrow but practically consequential. Musk’s lawyers argue that Altman, Brockman, and OpenAI’s nonprofit board breached fiduciary duties to devote the charitable assets of a nonprofit corporation solely to its mission — in this case, “developing artificial intelligence for the benefit of humanity.” They argue that by allowing OpenAI’s for-profit subsidiary to grow into the dominant operational entity (now valued at over $850 billion), and by allowing Microsoft to take a 27% stake and additional intellectual property rights in the 2023 transaction, Altman and Brockman effectively transferred charitable assets into private commercial control. Beyond breach of trust, Musk’s filing alleges unjust enrichment: that Altman and Brockman personally profited from a structure they were obligated to manage as charitable stewards.

OpenAI’s defense, led by attorney William Savitt, has built around three counter-arguments. First, no formal charitable trust was ever established — the early emails and donor commitments were not a legally binding instrument. Second, Musk himself was deeply involved in discussions about creating a for-profit structure in 2017, before he left the board, and at one point proposed merging OpenAI into Tesla under his personal control. Third, Musk’s lawsuit is filed in bad faith — he founded xAI in 2023 as a direct competitor to OpenAI, and in OpenAI’s framing he is “weaponizing the legal system” to handicap a rival. The defense has been unusually aggressive in painting Musk’s motives as competitive rather than ethical.

How did we get here? The eleven-year history

What’s been said in court so far?

The trial’s most consequential testimony has come from witnesses other than Musk and Altman themselves — and it has cut largely against the plaintiff’s narrative. Three sessions stand out as load-bearing for the case.

The cumulative effect of weeks one and two has been to undermine Musk’s central narrative without fully destroying it. OpenAI has effectively shown that Musk himself was deep in for-profit conversations in 2017, that his closest advisor approved the very transaction he now claims was a betrayal, and that his actions toward OpenAI personnel have been competitive rather than charitable. But Toner’s testimony — and the broader pattern of OpenAI governance issues that the trial has surfaced — gives Musk’s lawyers something real to work with in closing arguments. The question for the jury and the judge is not “did Musk get treated unfairly” but “was the corporate structure that emerged consistent with what was promised in 2015?”

The damages: $134 billion and what it means

The damages theory is unusual. Musk’s legal team, citing financial economist C. Paul Wazzan, frames the claim around what an “early backer” of a now-valuable company would be entitled to recover if defrauded. The argument is structured like an early-stage investor’s recovery on a successful startup: Musk’s $38 million in 2015–2018 donations, his lawyers argue, are equivalent to early venture capital that should have entitled him to a proportional share of the resulting $500+ billion valuation. The methodology is contested — donations to a nonprofit are not equity investments, and the entire premise of charitable giving is that the donor does not retain a financial claim on the recipient’s later commercial success. OpenAI’s defense has called the calculation “unserious” and the demand “part of an ongoing harassment campaign.”

Even legal observers sympathetic to Musk’s underlying argument view the full $134 billion as largely a negotiating posture. The realistic damages scenario, even on a Musk-favorable ruling, is more likely to involve specific disgorgement of identifiable gains — perhaps tens of millions to a few billion related to specific transactions — rather than a wholesale reattribution of OpenAI’s valuation. The judge has discretion in setting any award, and Judge Gonzalez Rogers has repeatedly signaled skepticism toward sweeping damage theories during pre-trial proceedings. The headline number is doing more work in shaping public perception than in setting realistic expectations of what the court might award.

Why most legal analysts think Musk will lose

Fortune’s analysis, published days before the trial opened, captured the consensus among legal observers: “Most legal analysts say Musk’s case is weak and that he’s likely to lose.” The reasoning is straightforward. To prove a breach of charitable trust, Musk’s lawyers must establish that the early OpenAI communications — emails between founders, donor commitment letters, internal discussions — created a legally enforceable trust instrument. This is a much higher bar than proving that the founders intended to operate as a nonprofit, which is undisputed. The jury and judge would need to find that those informal documents had the legal weight of a formal charitable trust agreement, which most charitable-trust precedent does not support.

The discovery record has also been less helpful to Musk than his pre-trial messaging suggested. Internal emails from 2017 — surfaced during depositions and during Brockman’s and Zilis’s testimony — show Musk himself proposing for-profit structures, suggesting OpenAI merge into Tesla, and engaging in extensive structural discussions that included multiple commercial pathways. The “weird halfway breakup” Zilis described undercuts the narrative of a clean ideological break in 2018. And Zilis’s vote in favor of the 2023 Microsoft transaction is, in OpenAI’s framing, the strongest single piece of evidence against Musk’s claim that the transaction “stole the charity” — because his closest confidant, with full information, voted yes.

What it means for the AI industry

The most immediate industry consequence is that nonprofit-to-for-profit AI lab conversions are no longer routine corporate maneuvers. Until Musk’s lawsuit, the assumption was that an AI lab founded as a nonprofit could pivot to a for-profit structure as long as the original charitable mission was preserved in some governance form. The trial — regardless of outcome — has demonstrated that this pivot is now subject to extensive legal scrutiny, multi-year litigation risk, and significant reputational exposure. The October 2025 attorneys general agreements that OpenAI struck with California and Delaware show how much governance compromise these conversions now require. Future labs considering similar pivots will look at OpenAI’s experience and either lock in tighter mission-protection structures upfront or stay nonprofit longer.

Anthropic’s Public Benefit Corporation structure, by comparison, looks better positioned. Anthropic was founded as a PBC — its commercial mission is baked into its corporate form rather than added on later via conversion. The Long-Term Benefit Trust structure adds another layer of mission-aligned governance. None of this immunizes Anthropic from future challenges, but it does mean the kind of “you converted from nonprofit to for-profit” claim Musk is making against OpenAI doesn’t apply to Anthropic in the same way. Investors evaluating AI lab governance risk are likely to view Anthropic’s structure more favorably as a result of the trial, regardless of who wins.

For xAI, the trial is creating an awkward parallel. Musk’s company is structured as a private for-profit entity with no charitable mission framework. While that is legally cleaner — there is no nonprofit conversion to challenge — it also undercuts the moral framing of Musk’s lawsuit. If charitable mission alignment is so important that Altman and Brockman should face billions in damages for diluting it, why is xAI not held to the same standard? Several legal observers have pointed out this asymmetry, and OpenAI’s defense has used it effectively. For other AI lab founders watching the trial, the lesson is clear: pick a corporate structure that aligns with your actual mission and operating reality, and stay with it.

What it means for tech leaders, investors, and the market

How does the trial affect AI development itself?

The most underappreciated effect of the trial is what it has revealed about how AI labs actually manage the tension between mission and commercial pressure. The discovery process has surfaced that, in the case of OpenAI, mission-protection was an ongoing negotiation rather than a fixed framework — board members debated for-profit structures, voted on commercial transactions while holding fiduciary nonprofit duties, and managed external pressures from major investors with limited governance scaffolding. That picture is not unique to OpenAI; it is how most early-stage research-mission organizations have always operated. What the trial has changed is the visibility of these dynamics, and the legal exposure that comes with operating them informally.

For the AI development trajectory specifically, three concrete effects are likely. First, more labs will adopt formal mission-protection mechanisms upfront, modeled on Anthropic’s PBC + Long-Term Benefit Trust structure rather than OpenAI’s nonprofit-with-for-profit-subsidiary structure. Second, AG-level oversight of AI lab governance will likely expand. California and Delaware are already involved; other states with significant AI presence (New York, Massachusetts, Texas) may follow. Third, internal AI lab governance will become more legible — documented decision-making, formal mission-impact assessments for major commercial transactions, more independent board oversight. These are quiet, structural changes, but they are likely to shape AI development for the rest of the decade.

What happens next

The immediate near-term schedule is well-defined. The remaining two weeks of testimony are expected to feature Ilya Sutskever — OpenAI’s co-founder and former chief scientist whose 2023 vote helped trigger the Altman firing — and Microsoft CEO Satya Nadella, whose testimony will address Microsoft’s strategic interest in the OpenAI partnership. Sam Altman, who was present in court but has not yet testified extensively, may be recalled. Closing arguments will likely take two to three days. The advisory jury will then deliberate, likely producing a verdict within the second half of May.

The advisory verdict is not the end. Judge Gonzalez Rogers will issue the binding ruling, which can take weeks or months — particularly if she chooses to address remedies as well as liability. Both sides have signaled that appeals are virtually certain. The Ninth Circuit Court of Appeals will likely become the next venue, extending the legal proceedings into 2027 or beyond. In the meantime, OpenAI’s restructuring remains in place, the company continues to ship products, and Microsoft’s 27% stake remains intact. The day-to-day operation of OpenAI is unlikely to be materially affected by the verdict for at least 6–12 months.

Final take

For all its theater — the Karpathy text messages, the Tesla board seat offer, the diary entries, the storming around the conference table — the trial has revealed less about the legal merits of charitable trust law than about the governance reality of frontier AI labs. The picture that has emerged is not one of villains stealing a charity, nor one of crusaders defending a mission. It is one of co-founders making consequential decisions about a fast-growing technology under significant uncertainty, with informal arrangements that proved inadequate to the scale of value the technology eventually created. That is a much more recognizable story than the one Musk’s lawsuit set out to tell — and it is the story that will shape AI governance for the next several years.

The simplest summary I can offer: Musk will likely lose the case but is succeeding at something his lawsuit may not have intended — establishing a public record of how AI labs are actually governed, and creating durable pressure for that governance to become more formal, more transparent, and more constrained. Whether that is good for AI development depends on whether you think the constraint is calibrated correctly. Reasonable people will disagree. But the era of casual, informal AI lab governance is functionally over — and that may be the most important thing this trial accomplishes, regardless of who walks out of the Oakland courthouse with a verdict in their favor.

Frequently asked questions