The core of this trial is the complete collapse of trust between an AI giant valued at over a trillion dollars and two former friends.

Written by: Jin Lu

Edited by: Bo Yang

On April 28, U.S. time, the case of Musk v. OpenAI opened in the federal court of Oakland, California.

Musk, as the first witness, attempted to explain to the nine-member jury why he originally wanted to establish OpenAI as a non-profit organization. OpenAI CEO Sam Altman and President Greg Brockman were present to listen.

Ten years ago, they were partners sharing visions for the future of AI. Now, Musk demands that Altman and Brockman leave OpenAI and return all "ill-gotten gains" to the OpenAI charity.

The core of this trial is the complete collapse of trust between an AI giant valued at over a trillion dollars and two former friends.

01 Three Core Disputes

On April 27, the jury selection for this case was completed. The presiding judge is Yvonne Gonzalez Rogers of the U.S. Northern District Court of California, who previously presided over the Epic Games v. Apple antitrust case in 2021.

The lawsuit Musk submitted in 2024 originally included 26 charges; after preliminary rulings, only two remain for trial: breach of charitable trust and unjust enrichment. The trial is divided into two phases—the liability phase, which involves the jury and whose verdict is advisory; if liability is found, the remedy phase will be adjudicated by the judge alone.

On the morning of April 28, opening statements were made sequentially. The three legal teams representing Musk, OpenAI (including Altman and Brockman), and Microsoft engaged in a battle over three core issues.

Dispute One: What is OpenAI's mission, and does Musk's $38 million donation come with conditions?

Musk's attorney, Steven Molo, displayed OpenAI's founding charter from 2015 in court. The document states that the organization "is not organized for the private benefit of any individual" and will create "open-source technology that serves the public interest." Molo claims that Musk's donation of approximately $38 million constitutes a charitable trust, requiring OpenAI to permanently maintain its non-profit status.

In the opening statement, Molo raised three questions for the jury to keep in mind:

• Does OpenAI have a charitable mission to operate as a non-profit organization, namely to develop safe, open-source AI for the benefit of humanity;
• Did Altman and Brockman violate this mission by establishing profit-oriented businesses;
• Did Microsoft know about this mission and assist Altman and Brockman in violating it?

He directly pointed out, "Without Musk, there would be no OpenAI," and claimed that OpenAI's shift to profit-oriented operations is tantamount to "violating every promise." "No one should be allowed to steal from charities," he said.

Musk echoed this stance on the witness stand. "It specifically refers to a charity that does not benefit any individual," he said, pointing to the founding charter, "I could have set it up for profit, but I chose not to. I chose to make it something that benefits all humanity."

Musk appeared as the first witness

Musk characterized the case as a fundamental issue regarding charitable donations: "Stealing from charities is wrong. If Altman and OpenAI win, it will open the door for the plunder of every charity in America." He further warned, "The consequences of this case go far beyond me or anyone here. The entire foundation of charitable giving in America will be destroyed."

OpenAI's lawyer, William Savitt, gave a completely opposite response. He told the jury, "The question is whether OpenAI made a specific commitment to Musk at the time of his donation. The answer is no." Savitt insisted that the donation did not come with any conditions. He also revealed that Musk had never fulfilled all of his donation commitments, causing the organization to seek additional support.

Savitt also tried to counter Musk's claims by referring to OpenAI's current structure. He told the jury that despite Musk's accusations, the company had not abandoned its non-profit mission. The non-profit foundation "still controls the organization" and "is doing cutting-edge work in treating diseases and promoting economic diversity." He pointed out that Altman has no shares in OpenAI but claimed that Altman profits through various companies that do business with OpenAI and has stated that he may acquire equity in OpenAI in the future.

Dispute Two: Did Musk ever endorse OpenAI's shift to profit?

Savitt submitted an email from former OpenAI board member Shivon Zilis, addressed to Sam Teller, who worked for Musk.

The email discussed two restructuring options: integrating everything into a class B corporation (i.e., a public benefit corporation) or establishing separate class C corporations and non-profit organizations. Savitt claimed that Musk "never expressed the view that OpenAI must remain a purely non-profit organization," and stated that he "only supported profit-making organizations as long as he controlled everything."

Savitt also provided details about equity distribution. The evidence he showed the jury indicated that Musk's chief of staff had discussed giving Musk 55% of the profit shares while giving Altman 7.5% of the shares.

Molo preemptively addressed this issue in his opening statement. He admitted that Musk had indeed discussed the idea of creating a profit-oriented version of OpenAI but emphasized that Musk's condition was always that "the non-profit must control the for-profit entity." Molo claimed that the idea at that time was for Musk to control the for-profit subsidiary, and the importance of the subsidiary would "diminish over time," merely serving as a means to raise funds in the short term.

Musk himself directly responded to the structural dispute on the witness stand. He acknowledged that there had been discussions within OpenAI about creating a profit-oriented entity between 2017 and 2018. He was open to "a small profit-making institution that funds a non-profit organization," but on the condition that "as long as it does not become counterproductive"—the non-profit must maintain control.

He also explained the reasons for his eventual departure: when other founders demanded too much equity in the proposed profit-oriented department, he expressed in an email to the executive office in 2017 that this was "extremely frustrating" and seemed determined to leave. He officially left the OpenAI board in 2018.

Dispute Three: What is Musk's true motivation for the lawsuit?

Savitt's explanation was: "We are here because Musk was not satisfied with OpenAI."

He outlined a timeline: Musk left after a power struggle in 2018, claiming "they would surely fail," but Altman and others "had the courage to move forward and succeeded without him"; ChatGPT was released at the end of 2022, causing a global sensation; Musk founded his own for-profit AI company, xAI, in 2023, and subsequently filed the lawsuit in 2024.

Savitt characterized Musk's motivation as "sour grapes," stating, "Mr. Musk does not like it this way, but that does not justify a lawsuit." He bluntly stated, "Musk does not understand artificial intelligence very well."

Savitt further refined this timeline. He stated that Musk only expressed dissatisfaction with OpenAI's pursuit of profit after ChatGPT triggered a global AI arms race, by which time Musk had already founded xAI.

Microsoft's attorney, Russell Cohen, concurred with the OpenAI camp on this point and provided two specific arguments.

First, he cited a post Musk made in September 2020 on X: "OpenAI has essentially been captured by Microsoft," as evidence that Musk was aware but failed to sue in a timely manner, presenting a statute of limitations defense. Second, he revealed a more personal channel of communication between Musk and Microsoft CEO Satya Nadella: "Musk knows how to contact Nadella. In the five years since the partnership was announced, Mr. Musk never picked up the phone and said, 'You can't do this.'"

The Microsoft lawyer noted that Musk had direct contact with Nadella

He emphasized that Musk personally held Nadella's phone number. Cohen concluded that Musk launched xAI after the success of ChatGPT, "and then he suddenly made claims against Microsoft."

Molo sought to separate the case from Musk's personal interests. "This case is not about Elon Musk," he emphasized to the jury, but about the defendants "lining their pockets and violating the fundamental principles on which the organization was founded." Molo claimed this case is for "the benefit of all humanity," not for economic interests. He urged the jurors to set aside their preconceived notions about Musk: "Everyone seems to know Musk and has different views of him. Not everyone's view is good, and not everyone's view is bad."

Molo also conceded that the fact that Altman does not have shares in OpenAI might weaken Musk's assertion that Altman profited from the non-profit. But he contended that Altman profits through various companies that do business with OpenAI and has indicated that he may gain equity in OpenAI in the future.

02 Aside: Musk Talks About AI Doomsday

During the trial, numerous scenes appeared that drifted away from the three core disputes. These contents did not involve the nature of the donation, the legality of structural changes, or motives for the lawsuit but formed a dramatic narrative on the first day of the trial.

Musk's self-presentation took up a large part. His lawyer, Molo, depicted a complete portrait of an entrepreneur to the jury through a series of questions: Musk grew up in South Africa, worked as a lumberjack and waited tables in Canada before coming to the U.S., and carried "$100,000 in student debt" while in college. He also serves as CEO of Tesla and SpaceX, "working 80 to 100 hours a week," "with no vacation, no vacation homes."

When discussing SpaceX, Musk stated that the company's goal is "to make life multi-planetary," calling it "the life insurance for life as we know it." Regarding Neuralink, he said its long-term goal is "AI safety," and "if we can achieve coexistence between AI and humans, we can create an AI that is more beneficial to humanity."

On AI, Musk released strong risk warnings. He compared training AI to raising children: "Ultimately, when the child grows up, you cannot really control that child, but you can try to instill the right values." He predicted the pace of AI development: "It could be as intelligent as any human by next year at the earliest."

He then concluded, "It could make us more prosperous, but it might also kill us all." He used two movies to metaphorically describe the two possibilities: "We want 'Star Trek,' not 'Terminator.'"

A small deviation from the main line occurred during cross-examination. When Molo asked "Who is Shivon Zilis?" Musk paused for a moment, stumbled through a few words, then laughingly called her his "chief of staff." In fact, Zilis is a former OpenAI board member and the mother of several of Musk's children.

Altman maintained a very low profile on the first day of the trial. He did not take the opportunity during the opening statements to speak to the media or the public. The trial records described him only as: "arms crossed, looking concerned, talking to lawyers and team members"; during recess "typing on his phone"; when the judge warned both sides to stop online attacks, he remained silent, only agreeing with the judge's request.

03 Outside the Courtroom Struggle

Outside the trial, public confrontations between both sides continued.

Before the court session began, Judge Rogers summoned Musk and Altman to the bench, asking them to cease online attacks on social media and urging both sides to start with "clean hands," seeking to "speak less" on social media. Both agreed.

Regarding the industry impact of the trial, AI safety expert Vivian Dong predicted it would be "primarily limited to OpenAI."

"No specific AI safety policies or industry practices are being tried," she said, "it would be unprecedented for the court to order a structural change of the kind Musk is seeking regarding the private violation of charitable trust." She also added that the officials responsible for overseeing OpenAI's fulfillment of its charitable mission are the attorneys general of Delaware and California, not Musk himself.

Emarketer Chief Analyst Nate Elliott offered another perspective: "If Musk wins, it would represent a rare case of tech CEOs being held accountable. It might also signify the end of OpenAI's business and provide xAI and Grok a competitive opportunity that they currently lack."

After this trial session, Musk is scheduled to return to court on April 29, U.S. time, to continue facing direct questioning from attorney Molo, followed potentially by cross-examination from Savitt. The judge has expressly instructed Musk not to discuss matters with his lawyers at night.

Subsequent trial sessions will call more witnesses. Musk's legal team indicated that Musk would later have Jared Birchall testify, who manages his billions of dollars in assets at Musk's family office Excession LLC and also holds executive positions at xAI and Neuralink.

Musk's expert witnesses include AI researcher Stuart Russell from the University of California, Berkeley, and law professor David M. Schizer from Columbia University.

Key witnesses expected to testify include Altman, Nadella, Murati, Sutskever, and others

Additionally, witnesses expected to testify include Microsoft CEO Nadella, former OpenAI CTO Mira Murati, and early key OpenAI researcher Ilya Sutskever. Musk specifically mentioned the difficulty of poaching Sutskever in his testimony—calling it "a huge effort" and "an extremely difficult process," which led Page to "refuse to talk to me again." He described Sutskever as the "most important" researcher contributing to OpenAI's existence.

The trial is expected to last about four weeks. In the upcoming cross-examination phase, Musk's testimony regarding the founding intention, the nature of charitable donations, and structural changes will be tested by the opposing lawyer for the first time.