The lawyer leading Sam Altman's defence in the Oakland federal courtroom this past week is the same one who, four years ago, made Elon Musk buy Twitter. William Savitt, a partner at Wachtell, Lipton, Rosen & Katz, represented Twitter in 2022 when the company sued to force Musk through with the forty-four billion dollar purchase he had spent the spring trying to wriggle out of. The case never got to a verdict; Musk capitulated shortly before trial, signed the cheque, and renamed the company. Savitt won by closing the exits.

He is now back across the courtroom from the same person, this time defending OpenAI and Altman against Musk's claim that the 2019 conversion to a for-profit structure betrayed the company's founding charter. According to Business Insider's profile this weekend, Altman picked Savitt specifically because of the prior result. The lore travels with him. The reasoning, told to a journalist by people who work with him, is that Musk responds poorly to opposing counsel who refuse to be impressed.

This is one of the small, structural facts about American high-stakes litigation that does not get talked about much. The top corporate trial bar is small. The same fifteen or twenty people show up across most of the cases that matter, and they remember each other. A defendant who has been across from a particular litigator before knows what that lawyer will do on cross, knows which exhibits they will dwell on, knows the register of voice they use when they are about to spring something. Savitt has had Musk's deposition strategy in his head since 2022. Most of the work was done before he walked into the Oakland building.

It also tells you something about how Altman thinks about risk. He could have hired any of the white-shoe firms in the country. The choice of the lawyer who had specifically beaten Musk in a high-profile commercial case was a tell. It said: this is going to be conducted as a contest of personalities, and we are bringing the personality who won last time. The trial coverage this week, which has spent as much time on Musk's demeanour as on the underlying corporate-governance question, suggests Altman read the room correctly. The judge, Yvonne Gonzalez Rogers, has already told Musk to stop making things worse outside the courtroom. Savitt has barely had to push.

What I find genuinely interesting is what this implies about the shape of the AI industry's coming decade in the courts. There will be many more of these cases. The IP questions around training data, the contractual questions around model distillation that came up on the stand last week, the corporate-governance questions about who actually owns a foundation-model lab when its mission and its valuation pull in opposite directions: all of these will be litigated by the same small bar, in front of the same handful of judges who handle complex commercial disputes in the relevant districts. The matchups will repeat. Reputations will compound. The lawyers who win one of these early cases will be the ones every defendant calls for the next one, and the lawyers on the other side will be the ones the plaintiffs' bar reaches for.

The trial is not over. Brockman and Altman himself are still to testify. The verdict could go either way. But one piece of the dynamic is already settled, which is that the most litigious billionaire in American technology has been put back across the table from the lawyer who got him to write a forty-four billion dollar cheque he did not want to write. That is a small, lawyerly victory in itself, and it happened before opening arguments.

Sources: