Elon Musk texted Greg Brockman two days before trial to ask about settling. Brockman suggested Musk drop all claims against the individuals. Musk replied: “By the end of this week, you and Sam will be the most hated men in America. If you insist, so it will be.” The exchange, disclosed in a court filing on Sunday, was not shown to the jury. Judge Yvonne Gonzalez Rogers ruled that settlement communications are inadmissible. But the text captures the dynamic that has defined the first week of the most consequential AI trial in history: a dispute that both sides had the opportunity to resolve privately, and that both sides chose to fight publicly, for reasons that their own private writings make uncomfortably clear.

The testimony

Brockman is expected to take the stand on Monday in the federal courthouse in Oakland, California, where the trial over OpenAI’s nonprofit-to-profit conversion began last week with $150 billion at stake. Musk is suing OpenAI, Sam Altman, and Brockman, alleging that approximately $38 million he donated to the nonprofit was used for unauthorised commercial purposes. He is seeking tens of billions of dollars in damages, the removal of Altman and Brockman from their positions, and the unwinding of the for-profit conversion that OpenAI completed in October 2025, when it restructured into OpenAI Group PBC with an $852 billion valuation, the OpenAI Foundation retaining a 26 per cent stake, and Microsoft holding 27 per cent.

Musk dominated the first week on the stand, testifying over three days, calling himself “a fool” for funding OpenAI, accusing its leadership of “looting the nonprofit,” and repeatedly telling the jury: “You can’t just steal a charity.” Under cross-examination, he clashed with OpenAI’s lawyers, who pressed him on his competing AI company xAI, which he valued at $250 billion in its February merger with SpaceX while describing it in court as a fraction of OpenAI’s size. Musk acknowledged that xAI had “partly” used OpenAI’s technology to train its own models through distillation, a concession that complicates his positioning as a wronged benefactor rather than a competitive rival.

The journals

Brockman’s testimony is expected to be more damaging to his own side than to Musk’s, because the most cited evidence against OpenAI’s leadership comes from Brockman’s personal journals. In entries that Musk’s lawyers obtained through discovery, Brockman wrote: “Financially, what will take me to $1B?” In another entry, he described his public commitment to OpenAI’s nonprofit mission as “a lie.” Judge Gonzalez Rogers cited these journal entries in January when she denied OpenAI’s motion to dismiss the case, writing that the entries “suggest Brockman intended to deceive” about the organisation’s charitable purpose.

OpenAI’s lawyers have fought to contextualise the journals. They argue that the entries are “staged for maximum misrepresentation,” cherry-picked from hundreds of pages of reflective personal writing that includes self-doubt, aspiration, and the kind of unguarded introspection that looks incriminating when extracted from its full context. The defence will likely present Brockman’s journals as stream-of-consciousness reflection, not evidence of a plan to defraud. Whether the jury reads them as private ambition or premeditated deception will shape the outcome of the trial.

The tension is structural. Brockman has been OpenAI’s most prominent public advocate for the company’s technical ambitions, recently claiming that AI now writes 80 per cent of OpenAI’s code. He returned from an extended leave in November 2024 as the company’s “builder-in-chief,” the executive responsible for translating Sam Altman’s trillion-dollar vision into operational reality. His journal entries, written during the period when OpenAI was still formally a nonprofit, suggest that the commercial ambitions he now executes were present long before the corporate structure changed to accommodate them.

The settlement

The pre-trial settlement exchange adds a layer that neither side’s narrative fully accounts for. Musk reached out to Brockman on 25 April, two days before jury selection, to gauge interest in resolving the case. The fact that Musk initiated contact suggests he recognised the risks of trial, not least the cross-examination that would expose xAI’s own use of OpenAI’s work. Brockman’s response, proposing that Musk drop all claims against the individuals, was a non-starter: Musk’s entire case rests on the argument that Altman and Brockman personally betrayed the nonprofit mission. Dropping individual claims would have gutted the lawsuit.

Musk’s subsequent text, threatening to make Brockman and Altman “the most hated men in America,” is the kind of statement that would ordinarily be devastating if presented to a jury. The judge’s ruling that settlement communications are inadmissible means the jury will not see it. But the exchange has already shaped the public narrative of the trial, reinforcing Musk’s pattern of treating legal disputes as extensions of his social media persona: escalate, threaten, perform outrage, and frame the conflict as a moral crusade rather than a commercial rivalry.

The stakes

What the jury will decide is narrower than what the trial represents. The legal questions concern whether OpenAI’s conversion from a nonprofit to a for-profit public benefit corporation violated charitable trust obligations, whether Musk’s donations were obtained through misrepresentation, and whether the individuals who executed the conversion enriched themselves at the expense of the organisation’s stated mission. Musk’s lawyers have asked for up to $134 billion in damages from OpenAI and Microsoft, a figure that reflects not the value of Musk’s original donations but the scale of the commercial enterprise those donations helped create.

The broader question is whether a nonprofit AI research lab can legally become an $852 billion for-profit company and distribute equity to its founders. OpenAI completed that conversion with the approval of the attorneys general of California and Delaware in October 2025, structuring the new entity so that the Foundation appoints all board members of the PBC and can remove them at any time. The company then raised $122 billion from Amazon, Nvidia, SoftBank, and others, ended Microsoft’s exclusive licensing arrangement, and began laying groundwork for an IPO at a potential valuation of $1 trillion. The commercial expansion that Musk calls theft is, from OpenAI’s perspective, the fulfilment of its mission at a scale that the original nonprofit structure could never have supported.

The contradiction

Both sides are asking the jury to accept a version of events that their own private communications complicate. Musk presents himself as a selfless donor who funded a nonprofit to benefit humanity and was betrayed when its leaders converted that work into personal wealth. His own actions, founding a competing AI company, valuing it at $250 billion, acknowledging that it used OpenAI’s technology, and attempting to settle the case days before trial, suggest motivations more complex than altruism. His approach to corporate control at SpaceX, where the IPO filing confirms that Musk and insiders retain dominant voting power through a dual-class share structure, indicates that his objection to OpenAI’s governance is not to concentrated control itself but to concentrated control by people other than him.

Brockman and Altman present themselves as mission-driven founders who built the most important AI company in the world and structured its conversion to preserve nonprofit oversight. Brockman’s journals, in which he aspired to personal wealth of $1 billion and called the nonprofit commitment “a lie,” suggest that the commercial trajectory was not an unfortunate necessity but an anticipated outcome. OpenAI’s lawyers may successfully argue that journal entries are not corporate plans. But the jury will have to reconcile Brockman’s private words with his public role, and that reconciliation will be the central drama of the trial’s second week.

Musk tried to settle. Brockman suggested terms that would have let him keep his job. Musk threatened to destroy their reputations instead. Now Brockman takes the stand, and his own handwriting will be projected on the courtroom screen. The trial that both men could have avoided will be decided by nine jurors reading a journal that was never meant to be read by anyone. The verdict, expected by mid-May, will determine whether OpenAI’s leaders converted a charity into an empire, or whether they built something that outgrew the structure it was born in. Brockman’s journals suggest he knew which it was going to be. He wrote it down.