Elon Musk testified Wednesday that he felt “deceived” into funding OpenAI’s early years, claiming he was a “fool” to believe the organization would remain a nonprofit dedicated to the public good.
The second day of Musk’s testimony in the landmark civil trial against OpenAI and its CEO, Sam Altman, highlighted the bitter collapse of a partnership that once aimed to safeguard humanity’s future.
It also underscores the enormous stakes in a closely-watched trial that could be a “preview of the next decade of AI,” Empromptu CEO Shanea Leven said in an email. “This trial isn’t just Musk versus Altman. This isn’t a model race. It’s a control battle over incentives, data, and who defines how intelligence operates at scale. While everyone debates which model to use, enterprises are still building on rented intelligence. The real risk isn’t choosing the wrong provider. It’s not owning the intelligence your business depends on.”
Musk, who contributed approximately $38 million between 2015 and 2017, accused Altman of “stealing the charity” to build a capitalistic powerhouse now valued at an estimated $852 billion.
The core tension of the trial rests on OpenAI’s evolution from a research-focused nonprofit to a massive commercial entity fueled by a $10 billion investment from Microsoft Corp. Musk testified that he felt a sense of betrayal as the company shifted gears, describing the transition as a “bait and switch.”
“It’s not OK to steal a charity,” Musk told the jury in an Oakland, Calif., court, wearing a black suit and occasionally glancing at the panel. He argued that the “moral high ground” of the nonprofit status allowed the company to recruit top talent and gain public trust, only to later pivot toward profit maximization.
To boldface his claims, Musk’s legal team presented 2019 text exchanges with Altman about the Microsoft deal. Musk testified that Altman’s offer to let him buy a stake in the new for-profit arm felt like a “bribe.”
OpenAI’s legal team, led by William Savitt, painted a sharply different picture. During a testy cross-examination, Savitt suggested that Musk’s lawsuit is born of “jealousy” over OpenAI’s success following his 2018 departure from the board.
Savitt introduced 2015 emails in which Musk himself appeared to brainstorm a for-profit structure for the company. Musk dismissed the evidence as “brainstorming,” asserting that “discussions are not a deal.”
The defense further alleged that Musk only soured on the venture when his own attempt to take full control of the company was rebuffed. They argued that Musk’s current legal challenge is an attempt to hobble a competitor and bolster his own AI firm, xAI.
“Musk v. Altman is a test of whether corporate form sets meaningful boundaries on AI governance, or whether it just slows the inevitable drift toward conventional capital structures,” Stavros Gadinis, faculty director at Berkeley Center of Law and Business, said in an email.
“The nonprofit rested on a single assumption, which was that frontier AI could be financed through charitable contributions,” he said. “By the late 2010s that assumption had collapsed. The compute bill had moved from the millions into the billions, and no philanthropic donor was prepared to commit at that scale to an organization that could offer neither equity nor the prospect of financial return.”
The atmosphere in the Oakland courtroom grew increasingly volatile as Musk and Savitt sparred over the billionaire’s contributions. When Savitt questioned the $38 million investment, Musk raised his voice, asserting that his reputation and naming of the company were worth far more. “Without me, OpenAI wouldn’t exist,” Musk claimed.
Seated in the audience, Sam Altman and co-founder Greg Brockman watched the proceedings closely, with Brockman seen taking frequent notes.
Musk is seeking $150 billion in damages — to be paid back to OpenAI’s charitable arm — along with the removal of Altman and Brockman from leadership. While a judge recently dismissed Musk’s fraud claims, the trial continues charges of breach of charitable trust and unjust enrichment.
“However this lands, it’ll quietly shape how the next generation of AI companies are structured, particularly those trying to balance a public-interest mission with the realities of raising capital and competing at scale,” says Joseph Hoefer, chief AI officer at Monument Advocacy. “What OpenAI attempted with its hybrid model was, in many ways, an experiment in aligning long-term safety goals with commercial incentives.”
“This case underscores how difficult that balance is to maintain over time, especially as competitive pressures intensify,” he said. “If those tensions prove legally or operationally unsustainable, it could push future AI labs toward more conventional corporate structures, or, at the very least, force a rethink of how governance, accountability, and fiduciary duties are designed from the outset. It’ll be theater though.”
The trial is expected to last three weeks, serving as a public reckoning for the founders of the technology that has come to define the modern AI era.
