Musk v. OpenAI trial: Nadella and Sutskever testify as case nears conclusion

The most striking exhibit from Nadella's testimony was an internal April 2022 email surfaced at trial, in which the Microsoft CEO drew an explicit historical parallel to Microsoft's 1980s PC partnership with IBM — worrying that, as Microsoft prepared to invest another $10 billion in OpenAI, it could end up being eclipsed the way IBM was after the original PC deal ^[1]. It is a remarkably candid admission from a sitting Big Tech CEO, captured in writing and now part of the public record. On the witness stand, however, Nadella's framing pulled the other direction: he testified that 'OpenAI had all the rights and resources they always had' and that Musk — who has his direct phone number — never once called to flag concerns about the deepening partnership ^[1]. The juxtaposition matters because Musk's case hinges on the claim that Microsoft's escalating investments effectively privatized a charitable mission. Nadella's testimony attempts to thread a needle: yes, Microsoft was strategically anxious about being subordinated by OpenAI, but no, the partnership did not strip OpenAI of independence or breach any commitment to its nonprofit roots.

Ilya Sutskever's appearance was arguably the trial's most psychologically loaded moment. He testified that he spent roughly a year gathering evidence of what he described as a 'consistent pattern of lying' and Altman 'undermining and pitting executives against one another,' culminating in a 52-page document delivered to the OpenAI board ^[2]. That dossier is exactly the kind of contemporaneous, insider-built record Musk's lawyers want the judge to weigh against Altman's character. Yet in the same proceedings, Sutskever expressed regret over his 2023 vote to fire Altman, saying simply, 'I simply cared for it, and I didn't want it to be destroyed' ^[2]. That softening — combined with his eventual vote to reinstate Altman — gives OpenAI's defense the human-scale narrative it needs: a brilliant scientist who acted on incomplete information, then corrected course. Developer-leaning YouTube discussion around the deposition has fixated on reporting that the firing was planned roughly a year in advance, framing it as evidence of a slow-building governance crisis rather than a panic move.

What the headlines don't always make clear is how much narrower this trial is than the case Musk originally filed. Of the 26 claims Musk asserted in 2024, only two remain at trial: breach of charitable trust and unjust enrichment ^[3]. The original lawsuit sought up to $134 billion in damages and included fraud and racketeering theories that did not survive procedural challenges ^[3]. That winnowing is itself a story. It means the legal question the jury and judge are actually being asked is not 'did Altman and Microsoft commit fraud?' but the narrower charitable-trust question that expert witness David Schizer framed bluntly: a for-profit subsidiary of a nonprofit is appropriate only when it advances the charitable mission, 'not when used as a vehicle to create wealth and profit for insiders' ^[4]. That framing is consequential because California's October 2025 regulatory approval of OpenAI's for-profit conversion already happened — the trial is now litigating whether the path to that approval was lawful, not whether the structure can exist.

The dollar comparisons surfaced at trial reframe the dispute as a fight over scale, not just principle. Musk's roughly $38 million in early donations helped seed an entity now valued in the hundreds of billions — Musk himself cited an $800 billion figure on the stand and called the dynamic a 'bait and switch' ^[5]. Microsoft has since deployed more than $13 billion across 2019, 2021, and 2023 tranches ^[1]. And critically, OpenAI raised approximately $122 billion after Musk filed suit — meaning the litigation has functioned as background noise rather than a chilling event for investors ^[6]. Fortune's Jeremy Kahn argues the trial generates 'more heat than light on the bigger concerns about who controls AI' ^[6]. The contrarian read circulating in finance-skeptical online communities is the inverse: that a Musk win could block an OpenAI IPO and trigger investor clawbacks. Both can be true — fundraising momentum during the trial does not mean OpenAI is safe from an adverse final ruling on remedies.

Local News Matters analysts assessed at the trial's midpoint that OpenAI's defense strategy is to 'keep the focus on Musk, his credibility and his motives — both those announced and unannounced' ^[4]. The trial record gives them ammunition. Under cross-examination, Musk admitted that his rival xAI uses OpenAI's models for validation — partially distilling them — telling the court, 'It is standard practice to use other AIs to validate your AI' ^[5]. Greg Brockman testified that Musk himself pushed for OpenAI to become a for-profit with Musk as CEO and walked away when offered only equal equity ^[3]. Shivon Zilis, a former OpenAI board member, added that Musk had previously tried to poach Altman to lead a Tesla AI lab ^[3]. Community reaction online has split along predictable lines: a 'both sides lose' contingent on AI-skeptic forums treating the trial as mutual reputational damage, and a smaller legally-curious cohort arguing the charitable-trust theory is stronger than the personality drama suggests. Procedurally, the judge — not the nine-person advisory jury — issues the final ruling, which means the credibility narrative shapes a one-person decision.