Human
Molo asks Brockman if Musk was “being mean” to him.
Live updates from Elon Musk and Sam Altman’s court battle over the future of OpenAI
2 days ago
OpenAI’s boardroom drama has finally gone on the record: in sworn testimony, company president Greg Brockman says there was a moment he genuinely thought Elon Musk might hit him — and that fear came at the exact point when Musk was denied majority control of the lab he helped found.
The setup: a partnership turns into a power struggle
The Musk v. Altman trial, nominally about contracts and control, has turned into an airing of the early, chaotic years of OpenAI — and, crucially, of Musk’s role in them. In court, lawyers have been walking witnesses through the pivotal period when OpenAI shifted from idealistic nonprofit to power‑player AI company.
At the center of this particular clash: a meeting where Musk wanted more than influence. He wanted the keys.
According to Brockman’s testimony, tensions spiked when Musk pushed for majority equity in OpenAI and the rest of the leadership refused. Musk “was angry that no one wanted to agree for him to have majority equity,” Brockman recounted, describing the breakdown of negotiations inside the company’s upper ranks.
The meeting that went off the rails
The crucial confrontation unfolded behind closed doors, but it came into focus in court.
Brockman testified that as the equity talks stalled, Musk’s demeanor shifted from frustration to something more menacing. Describing the moment on the stand, Brockman said: “I thought he was going to hit me… I truly thought he was going to physically attack me.”
This was not, in his telling, a passing flare‑up but a genuine fear of imminent violence — a remarkable claim given the stakes: control of one of the most consequential AI labs on the planet.
What pushed the scene over the edge, Brockman said, was the rejection of Musk’s demand for majority equity and the implication that OpenAI would remain, in some form, beyond the unilateral control of its most famous co‑founder.
As Musk “was storming out of the meeting,” Brockman recalled, the Tesla and SpaceX CEO turned to him and OpenAI chief scientist Ilya Sutskever with a pointed question: when were they planning to leave OpenAI? The two, Brockman said, were confused — this wasn’t part of any negotiated transition or governance plan. It sounded more like a threat.
Musk then allegedly laid down his leverage in plain language: “I will withhold funding until you decide what you are going to do.” Soon after, according to Brockman, Musk followed through and “stopped his promised quarterly donations to OpenAI.”
In an instant, what had been a debate over cap tables and mission statements became a funding crisis.
In the courtroom: parsing “being mean” vs. being menacing
Fast‑forward to the present trial. In cross‑examination, Musk’s side has tried to reshape that explosive episode as hard‑nosed, not violent.
Attorney Molo pressed Brockman on the characterization of Musk’s behavior, pointedly asking whether Musk was simply “being mean” to him. The line of questioning suggested an attempt to minimize Brockman’s description of fear and to frame the whole thing as a heated but standard‑issue Silicon Valley negotiation.
The reporting from the courtroom notes that Brockman “is being really squishy about this” under questioning, even as he sticks to his earlier claim that he believed Musk “was going to hit him at one point.” That hesitation matters: Musk’s camp appears to be drawing a distinction between aggressive deal‑making — allowable, even expected — and behavior that crosses into intimidation or potential assault.
Molo’s framing went further. He painted Musk as “hardheaded” in negotiations, not dangerous, and pivoted to criticize Brockman instead, suggesting that the OpenAI president “wasn’t familiar with corporate governance.” In other words: this wasn’t a tyrant raging at subordinates; it was a sophisticated founder exasperated with executives who didn’t fully understand the structure they were steering.
The volume in the room matched the stakes. Molo was “notably, raising his voice at Brockman,” according to the report — a meta‑commentary on power dynamics in the very act of interrogating them.
Two narratives of the same moment
What emerges from this testimony is a split‑screen view of the same confrontation.
From Brockman’s side, the timeline looks like this:
- Musk demands majority equity in OpenAI.
- Leadership resists, declining to hand him control.
- Musk becomes visibly furious.
- Brockman believes Musk might hit him and describes fearing a “physical attack.”
- As Musk storms out, he questions when Brockman and Sutskever will leave.
- Musk threatens to withhold funding “until you decide what you are going to do,” then halts his quarterly donations.
Threaded together, it’s a story of a founder using not just money, but presence and anger, to try to force a governance outcome — and crossing a line into what at least one executive perceived as physical intimidation.
From Musk’s legal camp, laid out through Molo’s questioning, the story is recast:
- Musk is indeed forceful and “hardheaded” in negotiations.
- The dispute is fundamentally about structure and control, not violence.
- Brockman is inexperienced in “corporate governance,” so his interpretation of standard founder behavior is unreliable.
- Any suggestion of looming physical assault is either an exaggeration or a misreading.
In that telling, this was the ugly but familiar part of tech‑company adolescence: a powerful founder trying to claw back control from a mission‑driven leadership team that didn’t understand the game.
Why this matters beyond courtroom theatrics
The clash over that meeting is more than gossip about a volatile billionaire. It cuts to the long‑running argument over what OpenAI is — and who it is for.
Musk has repeatedly cast OpenAI’s later pivot to a capped‑profit structure and close partnership with Microsoft as a betrayal of the lab’s original nonprofit, open‑science mission. The meeting described by Brockman marks the moment when those values collided head‑on with an old‑fashioned control grab: a demand for majority equity by the very person now suing to “restore” the purity of that mission.
If Brockman’s version holds up, it portrays Musk as willing to jeopardize OpenAI’s funding and stability when he could not secure unilateral control — and to do it in a way that made at least one top executive fear for his physical safety.
If Musk’s narrative wins the day, the lesson is different: that OpenAI’s leadership was naive, sloppy on governance, and perhaps weaponizing their discomfort with Musk’s personality to justify their own consolidation of power and their eventual embrace of big‑tech money.
The optics of power in AI’s founding myths
In 2026, with AI systems reshaping economies and politics, the industry likes to talk about alignment, safety, and responsible governance. The picture emerging from this trial is less polished: shouting matches, threatened funding pullouts, and an executive who says he braced for a swing from a co‑founder.
Even the courtroom exchange over whether Musk was just “being mean” lands awkwardly in that context. These are the people arguing over how to control systems that could one day outstrip human capability — and they are, by the eyewitness account, replaying the same egos‑and‑equity drama that has defined startups for decades.
What happens next in the trial will determine more than reputations. It will shape the legal and moral story Silicon Valley tells about the birth of the modern AI era: whether OpenAI’s trajectory was a principled break from a domineering founder, or a power shift rationalized after the fact by painting that founder as unstable.
Brockman has already put his marker down: “I truly thought he was going to physically attack me.” Musk’s lawyers, by contrast, are asking the court — and the public — to see it as something simpler: a billionaire “being mean” in a fight over who really owns the future of AI.