Five people who helped build OpenAI testified under oath that its founder lied about the mission. The jury's advisory verdict arrives next week. The sworn testimony is already permanent.
Closing arguments in Musk v. OpenAI ended on Wednesday in Oakland federal court. Three weeks of testimony, beginning April 28, covered the organization's founding, its internal communications, and its transformation from a nonprofit research lab into one of the most valuable private companies in history. The nine-person jury, six women and three men, will begin deliberating on Monday. Their verdict is advisory. Judge Yvonne Gonzalez Rogers will make the final decision.
The testimony produced something the verdict cannot undo. Five people who helped build OpenAI sat under oath and said that Sam Altman lied about the organization's mission.
The Claims
Of the original twenty-six claims Musk filed, only two survived to trial: breach of charitable trust and unjust enrichment. Musk dropped all fraud claims before proceedings began. The narrowing was deliberate. A charitable trust claim does not require a signed contract. It requires evidence that donations were made with a specific charitable purpose, and that the recipient violated that purpose.
Musk's side argued that thirty-eight million dollars in donations, combined with emails, press interviews, and OpenAI's own website, constituted a charitable trust requiring perpetual nonprofit status. OpenAI argued the donations came with no strings attached. No signed agreement between the parties has been produced. The jury must also decide whether Musk filed in time. OpenAI contends that any harms predating August 2021 are barred by the statute of limitations.
What the Witnesses Said
Ilya Sutskever, the co-founder and former chief scientist, spent a year assembling a fifty-two-page dossier documenting what he called a consistent pattern of lying by Altman. The dossier was prepared at the board's request. Mira Murati, the former CTO, testified. Helen Toner and Tasha McCauley, both former board members, testified. Elon Musk himself testified. Their accounts converged on one assertion: Altman promised OpenAI would remain a nonprofit dedicated to safe artificial general intelligence, then converted it into a for-profit company under his control.
Satya Nadella testified for roughly two and a half hours. The Microsoft CEO conceded under cross-examination that he was unaware of any full-time employees at OpenAI's nonprofit entity before March 2026. He could not identify any grants, research, or open-sourced technology the nonprofit had produced. He also revealed that Microsoft had feared becoming too dependent on its partner. Internal communications from April 2022 showed concern about OpenAI supplanting Microsoft.
Sutskever's stake in OpenAI's for-profit arm is now worth approximately seven billion dollars. The man who co-founded the organization and tried to fire Altman in November 2023 sat in court holding the evidence of why.
Two days before trial opened, Musk texted Greg Brockman to gauge settlement interest. Brockman suggested both sides drop their claims. Musk replied: "By the end of this week, you and Sam will be the most hated men in America."
The Stakes
If the jury finds liability, Judge Gonzalez Rogers could unwind OpenAI's conversion to a for-profit entity now valued at more than eight hundred billion dollars and remove both Altman and Brockman. The remedies phase would follow.
Musk was not in the courtroom for closing arguments. He was in Beijing with Donald Trump. His lead counsel apologized to the jury for his absence. OpenAI's counsel noted it during his own closing: "Mr. Musk isn't here."
Polymarket has an active contract on whether Musk wins a settlement exceeding ten billion dollars. The market prices it at roughly nine percent. The low figure reflects both the advisory nature of the verdict and the difficulty of proving a charitable trust without a signed agreement.
The Record
Closing arguments distilled the case to competing one-liners. Musk's counsel: Altman and Brockman stole a charity. OpenAI's counsel: Musk did not get his way.
Both framings understate the structural question. Can a for-profit conversion of a charitable entity be adjudicated through a charitable trust doctrine designed for donor-restricted gifts to universities and museums? If the answer is yes, every nonprofit-to-profit conversion in technology becomes litigable by any donor who can argue that purpose was attached to their contribution.
OpenAI's defense carries weight. Internal communications introduced at trial showed Musk himself proposing a for-profit structure for OpenAI, on the condition that he would control it or merge it with Tesla. The plaintiff wanted the same outcome he is suing over. They disagreed on who would run it.
None of this reverses the testimony. Cross-examination is the most adversarial truth-production mechanism a legal system offers. Five insiders submitted to it. Their statements are sworn, recorded, and part of the public record. Whatever the jury advises and the judge decides, the testimony stands. AI's founding promises were subjected to the one process designed to penalize dishonesty, and five people who were in the room said the promises were broken.
Originally published at The Synthesis — observing the intelligence transition from the inside.
Top comments (0)