Two days before the Musk v. Altman trial opened in Oakland, California, Elon Musk texted Greg Brockman to float a settlement. Brockman replied by suggesting both sides drop their claims. Musk's response: "By the end of this week, you and Sam will be the most hated men in America. If you insist, so it will be."

It did not go that way. Instead, Brockman spent two days on the witness stand in early May 2026 — referencing a personal journal, describing a scene in which he feared physical assault, and delivering what may be the most consequential eyewitness account yet of how Elon Musk left OpenAI. For enterprise leaders and founders watching this trial, the details matter far beyond the legal outcome — they reveal how one of the defining power struggles in AI history actually played out, and what it signals about the industry being built on its wreckage.

The August 2017 Meeting That Broke Everything

Late August 2017. OpenAI was still a nonprofit research lab. Its co-founders had recognized they needed a for-profit structure to fund the compute required to pursue AGI at scale. Key figures converged to decide the shape of that future. Elon Musk, who had co-founded the organization in December 2015 alongside Sam Altman, Greg Brockman, Ilya Sutskever, and others, arrived with an agenda and, notably, gifts: he had given each of his co-founders a Tesla Model 3. Brockman testified he saw this as an attempt to butter them up — a precursor to Musk's demand for full control.

OpenAI's head of research, Ilya Sutskever, had commissioned a painting of a Tesla as a friendly gesture for Musk. The gesture didn't hold.

When the others refused to cede control to Musk, Brockman testified the room changed completely. "Something just shifted in him. You could sense it. He was angry, he was upset." Musk sat quietly for several minutes. Then: "I decline." He stood up and, in Brockman's words, "stormed around the table." Brockman testified: "I thought he was going to hit me. He grabbed the painting and started to storm out of the room. And then he turned around and said, 'When will you be departing OpenAI?”

Brockman and Sutskever didn't leave. Musk stopped his regular donations. Within six months, he was off the board.

The painting of the Tesla — begun as a peace offering — left the room in Musk's hands as a symbol of something more complicated: a co-founder who wanted to control what he funded, and walked when he couldn't.

What Brockman's Testimony Actually Establishes

Musk's central legal argument in this trial — that OpenAI betrayed its founding mission when it transitioned to a for-profit structure — depends on a narrative in which he donated in good faith to a charity and was deceived. His lawyers spent the trial's first week on that frame, accusing Altman and Brockman of "stealing a charity," noting Musk's total donations amounted to approximately $38 million against an organization now valued at over $852 billion.

Brockman's testimony punctured that frame in several places. He testified that Musk never made open-sourcing a formal condition of his support — directly contradicting one of Musk's core claims on the stand. "Honestly, it was not a topic of conversation," Brockman said flatly. He testified that no commitments were ever made to Musk about corporate structure, and that he never heard anyone else make them either.

More damaging: Brockman revealed that Musk had secretly recruited OpenAI researcher Andrej Karpathy — a key figure in the work that would become ChatGPT — to lead Tesla's Autopilot division. Musk approached Brockman afterward with what Brockman described as "an apology and a confession." A contemporaneous email from Musk to a Tesla VP read: "The OpenAI guys are gonna want to kill me. But it had to be done." This was 2017. Musk was still on OpenAI's board. He was using a nonprofit he co-founded to talent-scout for his electric vehicle company — and doing it without telling the organization's leadership.

"He did not and does not know AI. We did not think he was going to spend the time required to actually get good at it."

— Greg Brockman, testifying about Elon Musk, Musk v. Altman trial, Oakland, May 2026

This isn't just biographical color. It is the central counternarrative OpenAI needs the jury to believe: that Musk's role was donor and figurehead, not technical visionary; that he left because he couldn't be in charge, not because the mission changed; and that his subsequent creation of xAI — now folded into SpaceX at a $250 billion valuation — is the real story of what Musk wanted all along.

The jury won't decide who built the better AI company. But it may decide whether Musk's $38 million donation entitled him to veto power over the direction of an organization that now employs thousands and shapes AI policy in Brussels, Beijing, and Washington simultaneously.

The Journal That Became Exhibit A Against Its Author

The irony of Brockman's predicament is almost novelistic.

He kept a private journal — detailed, candid, the record of a 30-year-old executive navigating one of the most consequential relationships in tech history. That journal was unsealed earlier in 2026 and became a weapon in Musk's lawyers' hands. One September 2017 entry: "Financially, what will take me to $1B?" Musk's attorney Marc Toberoff hammered that line repeatedly, framing a nonprofit's president as someone secretly mapping his path to personal fortune while publicly invoking the mission of benefiting humanity.

Brockman acknowledged the journal's existence, called it "deeply personal writings that were never meant for the world to see," and added: "There's nothing in there I'm ashamed of." His stake in OpenAI is now worth approximately $30 billion. The financial ambitions of 2017 were not unfulfilled.

What the journal entry actually reflects — and what Brockman's attorneys argued — is less damning than Toberoff framed it: a technologist in his early thirties, working at a nonprofit, wondering whether the enterprise he was helping build would ever produce personal financial return. Any honest founder, reading that entry, recognizes the question.

Musk's own lawyers introduced it as evidence of bad faith. The more accurate reading is that it's evidence of human ambition coexisting with genuine mission focus — not contradicting it. OpenAI spending $50 billion annually on compute, as Brockman testified, requires the kind of capitalization that no purely altruistic structure can provide. The for-profit transition was not a betrayal. It was arithmetic.

The Global Stakes of an Oakland Courtroom

For enterprise decision makers outside the United States, this trial is not merely a spectacle. Its outcome shapes the governance architecture of the AI tools they depend on.

In Europe, regulators at the EU AI Office — established under the EU AI Act that began applying to general-purpose AI systems in August 2025 — are watching closely. OpenAI's for-profit conversion, its Microsoft relationship (Microsoft is a named defendant), and the question of whether a charitable mission can survive commercial scale are precisely the governance questions EU regulators must assess as they determine compliance obligations for frontier AI models deployed across member states.

In India, where OpenAI has been expanding enterprise relationships and where the government's IndiaAI Mission has committed ₹10,372 crore to domestic AI infrastructure, the governance model for foundation AI labs is an active policy question. Whether OpenAI's nonprofit oversight structure remains meaningful — or whether Musk's lawsuit validates concerns about mission drift — affects how regulators in New Delhi, Seoul, and Singapore approach licensing and partnership frameworks for foreign AI platforms.

In the Gulf region, sovereign wealth vehicles including Saudi Arabia's Public Investment Fund and Abu Dhabi's MGX have invested in AI infrastructure across both sides of this dispute: MGX participated in OpenAI's March 2026 funding round that closed at an $852 billion valuation. These investors need clarity about what they're actually owning a piece of.

Key Takeaways

$852 billion — OpenAI's valuation as of its March 2026 funding round, which raised $122 billion in committed capital. Musk donated approximately $38 million across OpenAI's early years. The gap between those two numbers is the financial backdrop of this lawsuit.

August 2017 — The month Brockman testified that Musk demanded full control of a for-profit OpenAI, was refused, grabbed a painting of a Tesla, and asked the co-founders when they would be leaving. Six months later, Musk was off the board.

$250 billion — xAI's valuation upon its merger with SpaceX in February 2026. The AI company Musk built after leaving OpenAI is now embedded in a $1.25 trillion conglomerate preparing for a public listing.

$50 billion — OpenAI's estimated annual compute spend, per Brockman's testimony. This figure is the practical argument for why a nonprofit structure was always unsustainable at frontier AI scale.

The Karpathy maneuver — Musk's secret recruitment of Andrej Karpathy to Tesla Autopilot while Karpathy was at OpenAI in 2017. Brockman testified Musk admitted it, apologized, and confessed. Karpathy went on to become one of the most recognized AI researchers in the world and later returned to OpenAI before leaving again in 2023.

What This Tells Founders and Enterprise Leaders

Strip away the legal arguments and the tabloid texture — the haunted mansion, the Tesla gifts, Amber Heard serving whiskey — and what Brockman's testimony reveals is a governance failure with a lesson that transcends this case.

When a co-founder's financial contribution is treated as purchasing control rather than enabling a shared mission, the organization is operating without a governing framework resilient enough to survive disagreement. OpenAI's early structure gave Musk enormous informal leverage — through donations, through recruitment assistance, through name-recognition in talent pitches — without formalizing what that leverage entitled him to. That ambiguity is what this lawsuit is, nine years later, trying to resolve retroactively in a federal courtroom.

For founders building companies today, particularly those structured around a social or technical mission with substantial outside financial backing, the lesson is uncomfortable: the governance documents that feel like administrative overhead are the only thing that protects a company from exactly this kind of dispute when the money gets large enough to matter.

Sam Altman is expected to testify next. Whatever he adds, the underlying dynamic that Brockman established is now on the record. Musk didn't leave OpenAI because OpenAI changed. How Elon Musk left OpenAI, according to Greg Brockman, is that he stood up, grabbed a painting, and walked out because he couldn't be in charge.