Elon vs. OpenAI: Leaked Brockman documents and the case for both sides

Summary

OpenAI will likely win this lawsuit on structural merits. Elon donated between $38 million and $45 million to OpenAI's nonprofit in its early days, a substantial sum but not essential to the company's survival. Without Elon's money, OpenAI could have raised capital elsewhere. Sam Altman had Stripe holdings to liquidate, a network that included Reid Hoffman, Peter Thiel, and Dustin Moskovitz, and the option to pursue a traditional Series B. Everyone at the table reached the same conclusion around the same time: scaling laws meant AI progress would eventually require tens of billions in capital, far beyond what philanthropy could sustain. That is not fraud. Elon even acknowledged in leaked emails that the nonprofit structure "may not be the right one now," which contradicts his claim that OpenAI deceived him about remaining nonprofit. Even if Elon wins $38 million in damages, that is xAI's operating budget for a week. What matters to Elon is disruption, not money—bogging down OpenAI's IPO with a high-profile trial and buying time for xAI to compete. The lawsuit is corporate lawfare, and the proper remedy is the market, not the courtroom.

The leaked Brockman documents, however, tell a different story. Judge Gonzalez Rogers rejected OpenAI's motion to dismiss, signaling the court sees enough circumstantial evidence to proceed. Elon's ultimatum was direct: either continue the nonprofit or go do something on your own. He even offered to fund the nonprofit independently if needed, an offer OpenAI never took up. Sam Altman assured Elon the nonprofit remained "the right structure," then immediately began restructuring into a for-profit without his consent or knowledge. Brockman's own notes state: "it'd be wrong to steal the nonprofit from him, to convert to a b corp without him. That'd be pretty morally bankrupt." Brockman went further: "our plan... it would be nice to be making the billions." This is not a principled pivot to address capital constraints but people who wanted the upside of a venture-backed exit. OpenAI's certificate of incorporation commits the entity to "exclusively charitable purposes with the technology intended to benefit the public." Raising venture capital to build a subscription app with ads is not charity. A donor gave $38 million based on explicit promises about structure and mission, and the promise was broken. The jury in Oakland will hear about elite tech founders who promised AI for humanity, took Elon's money, then built a $500 billion for-profit empire with Microsoft. Critically, Elon never got equity in the for-profit despite multiple opportunities to negotiate one. Once bridges burned, he walked away entirely.

One complication: OpenAI's defense argues Elon donated indirectly through a donor-advised fund via Y Combinator's fiscal sponsorship, not directly to OpenAI. If the donation was not direct, the charitable intent clause may not bind as tightly. This devolves into esoteric legal definitions of trust, the kind of detail that could decide the case on a technicality. Brockman's internal notes also contain a quote from Altman framing Reid Hoffman's decision to start rival lab Inflection AI as a con: "He supported us in a moment when no one else would... I think OpenAI would have been pretty effed if he hadn't stepped up." Yet Altman also wrote that Hoffman was "blinded enough by the startup of being able to call himself the cofounder of a company." Hoffman invested $10 million in Inflection while sitting on OpenAI's board. Nobody sued, and Inflection eventually sold to Microsoft. The precedent matters for OpenAI's negotiating power with future donors. Altman's notes suggest OpenAI felt they had leverage to demand non-compete promises from new investors, a tactic few companies ever succeed with. Brockman approached Patrick Collison about investing, conditioning participation on "not investing in AGI slash big model competitors." These conditional asks are standard pitch-meeting language but dangerous in discovery.

OpenAI is signaling to investors the max damage is roughly $38 million, the original donation amount. But it could be higher if the jury values Elon's pro rata equity stake in what becomes a multi-hundred-billion-dollar asset. The company is planning an IPO while this trial unfolds. Wall Street will likely underwrite it by factoring in the settlement as a line item, assuming the business fundamentals hold. The real risk is not the money but the narrative damage and the reopening of questions about how OpenAI's for-profit actually came into being. The leaked documents show internal conflict and ambiguity. Ilya Sutskever worried OpenAI was not taking open-source AI seriously enough. Mira Murati pushed back on rushing to compete with Stability AI, warning the move contradicted her motivations. These fragments suggest a leadership that was not unified in its strategic pivot. That division could matter to a jury deciding whether the transition was inevitable or opportunistic. Satya Nadella said in a deposition his role at Microsoft was to be "dissatisfied with the rate of progress" and ensure the company stayed competitive. He had skin in the game. Microsoft's AI credibility and shareholder returns depended on OpenAI succeeding. That motivation is now clear in the record.