Brockman testifies in Musk v. OpenAI trial

The single most damaging exchange of the week was Brockman conceding, point by point under Steven Molo's cross-examination, the basic shape of the financial transformation Musk's team is challenging. Brockman agreed that he was part of the OpenAI founding group with Altman and Musk, that he was there when a for-profit subsidiary was created, that he was given an equity stake in that for-profit, that he did not contribute any money to get that stake, and that the stake is today worth close to $30 billion. He also acknowledged that the $100,000 he had pledged to the original nonprofit was never paid.

This is the through-line Musk's lawyers want the jury to walk away with: a nonprofit founded in 2015 to be a counterweight to Google produced, within a decade, a roughly $30 billion windfall for an executive who put in no cash. Brockman pushed back that the equity grant came in 2018 — years before ChatGPT, when 'financial and technical success was uncertain' — and that he did not vote on his own grant. But the ratio of inputs to outputs becomes the case's emotional anchor whether or not it maps cleanly to legal liability. A plaintiff doesn't have to win the doctrinal argument to win the narrative; the $0-to-$30B framing is the kind of fact a jury remembers when it weighs every other contested element of the testimony.

Eight years before this trial, Brockman was keeping a personal journal. One August 2017 entry asked, in plain language, 'Financially, what will take me to 1B?' Another acknowledged that converting the nonprofit could be characterized as morally bankrupt — a phrase the journal used about Brockman's own potential conduct. Those passages, subpoenaed and read aloud in court, do work that no piece of trial argument can match: they are contemporaneous, private, and pre-litigation. Musk's lawyers cannot have manufactured them, and Brockman cannot reframe them as performance for an audience.

This is why the TechBuzz trial analysis treats the journal as the legal weight of the case: 'It is another thing to defend that argument when the defendant's own private writings, from eight years before the lawsuit was filed, show him asking what it will take to get him personally to $1 billion and acknowledging in writing that the conversion would be morally bankrupt.' A breach-of-mission theory that rests on inferring intent from later corporate filings is hard. A breach-of-mission theory in which the defendant has, in his own handwriting, asked how to personally get to a billion dollars and described the contemplated conversion as morally bankrupt is qualitatively different. It moves the question from 'did the corporate form drift?' to 'what did the principals know about that drift, and when?' — territory where the plaintiff has the high ground.

If the journal is the intent evidence, Cerebras is the self-dealing evidence. Brockman acknowledged on the stand that he held a personal investment in AI chipmaker Cerebras during a period when OpenAI was discussing a potential acquisition of the company, and conceded he never disclosed that stake to Musk in writing. Molo's framing was unsubtle: 'Can you point to an email in which you told Elon you were an owner of Cerebras at the same time you were advocating that OpenAI do this transaction with Cerebras?' Coverage of the cross indicates Brockman could not.

The context multiplies the weight of that admission. OpenAI signed a $10 billion compute agreement with Cerebras in December 2025, and Cerebras's valuation roughly tripled to $23 billion by February 2026 — meaning whatever personal stake Brockman and Altman held appreciated alongside the deal flow they were directing from inside OpenAI. Reporting on the Altman side notes Brockman was also given an interest in Altman's personal investment fund or family office in 2017. None of this is, on its face, a finding of fraud — undisclosed adjacent investments are a fact pattern, not a verdict. But it gives Musk's team something more potent than the abstract for-profit-conversion argument: a concrete, dollar-traceable allegation that the same people running the nonprofit's procurement decisions were privately exposed to the upside of the counterparty.

The dollar architecture of this case spans six orders of magnitude, and once the figures are stacked the litigation logic becomes clearer. Musk's original nonprofit donation to OpenAI was $38 million. Brockman's stake in Altman's family office reportedly began around $10 million. Brockman's resulting OpenAI equity is worth roughly $30 billion. OpenAI's compute deal with Cerebras is $10 billion. The Mars colony Musk testified he wanted OpenAI's resources to help fund was $80 billion at the time. OpenAI's 2026 compute spend is reportedly around $50 billion. Musk is asking the court for up to $134 billion in damages. And the company at the center of it is now valued at roughly $852 billion.

What makes the chart legible as a story rather than a list is the gap between what Musk put in and what is now on the table. A $38 million donation, in nominal dollars, is roughly 0.004% of OpenAI's current valuation. The damages claim, if granted in full, would represent about 16% of the company. Whether or not the doctrine supports a full unwind, those proportions explain why both sides are litigating so hard: this is no longer a fight over a research lab's mission statement, it is a fight over the largest single transformation of nonprofit capital into private equity in modern tech history.

Two days before opening statements, Musk texted Brockman seeking a deal. When Brockman proposed both sides drop their suits, 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.' OpenAI's legal team has used those texts as evidence that Musk's litigation posture is closer to extortion than mission-defense. The most-engaged community thread on the entire trial week is precisely this exchange, and the dominant top-comment read flips the apparent meaning: a party threatening and pleading for a settlement two days before trial is, in plain street terms, a party that does not believe its own case.

That read carries through the rest of the public reception. The shape of community reaction is overwhelmingly hostile to both sides — a 'pox on both your houses' tenor in which Musk is treated less as a whistleblower and more as a billionaire unhappy he is no longer in on the deal, and Altman/Brockman are treated less as wronged executives and more as people whose own writings have indicted them. A small but visible contrarian thread among legally literate watchers is also pushing back on overinterpretation: the jury here is advisory, the judge is the actual factfinder, and Federal Rule of Evidence 408 governs whether parts of the settlement texts are even admissible for the inferences both PR camps want to draw. On the ground in mainstream coverage, the most-watched explanatory videos frame the central question as whether Altman betrayed the original nonprofit promise — and reserve the diary as the moment when that question stopped being a matter of opinion.