Altman Confronted on Stand with Accusations of Dishonesty at OpenAI Trial

On May 12, 2026, Sam Altman sat in the witness box at the U.S. District Court for the Northern District of California in Oakland and answered a question that has trailed him for years. "Are you completely trustworthy?" asked Steven Molo, the attorney representing Elon Musk [1]. Altman's response — measured, lawyerly, and carefully hedged — set the tone for two days of testimony that laid bare the internal fault lines of the most valuable AI company on Earth.

The question was not rhetorical. It was the centerpiece of a legal strategy built on years of internal communications, board memos, and sworn testimony from people who once worked alongside Altman and now say, under oath, that he cannot be trusted.

The Trial: What's Actually at Stake

Musk v. Altman, now in its third week before Judge Yvonne Gonzalez Rogers, is not a straightforward defamation case or a contract dispute. Of the 26 claims Musk originally filed in 2024, only two survived pre-trial motions: breach of charitable trust and unjust enrichment [2]. The core question is whether Altman and OpenAI president Greg Brockman made enforceable promises to Musk — who donated approximately $44 million to OpenAI's nonprofit — that the organization would remain a nonprofit dedicated to open-source AI for the benefit of humanity, and whether they violated those promises when OpenAI restructured into a capped-profit entity in 2019 and a public benefit corporation in October 2025 [3].

Musk is seeking damages exceeding $130 billion [2]. He has stated publicly that any award would be directed to OpenAI's nonprofit foundation rather than kept personally. Beyond money, Musk wants structural remedies: the removal of Altman and Brockman from their leadership positions, the unwinding of the for-profit conversion, and disgorgement of all equity and compensation the executives received during the restructuring [3].

The nine-person jury seated on April 27 is advisory — Judge Gonzalez Rogers will issue the final binding ruling, expected by mid-May 2026, though she has indicated she will weigh the jury's recommendations [4].

The Accusations: A "Consistent Pattern of Lying"

The most damaging testimony against Altman has come not from Musk, but from the people who built OpenAI alongside him.

Ilya Sutskever, co-founder and former chief scientist of OpenAI, testified that he spent approximately one year compiling a 52-page document for the board detailing what he described as Altman's "consistent pattern of lying" [5]. According to Sutskever, Altman's behavior included "undermining and pitting executives against one another," conduct he argued was "not conducive to any grand goal," including OpenAI's stated mission of safely developing artificial general intelligence [5]. Sutskever played a central role in Altman's removal from the CEO position in November 2023, though he later voted to reinstate Altman and testified that he regretted the firing [6].

Helen Toner, a former OpenAI board member who was among the four directors who voted to fire Altman in November 2023, has stated that the decision stemmed from "a pattern of behavior" related to his honesty and candor, as well as his resistance to board oversight [7]. Toner said the board learned about the launch of ChatGPT in November 2022 — one of the most consequential product launches in the history of technology — through Twitter, not from Altman [7]. She also testified that Altman did not disclose his involvement with the company's startup fund and gave the board inaccurate information about OpenAI's formal safety processes on multiple occasions [7].

Dario Amodei, now CEO of rival Anthropic and a former OpenAI executive, was cited during cross-examination as having accused Altman of misrepresenting the terms of an investment to him [8]. When confronted with this during testimony, Altman responded: "Dario has accused me of many things" [8]. Amodei has also separately stated publicly that Altman inserted undisclosed clauses into the deal with Microsoft that contradicted OpenAI's original commitments, and that when confronted, Altman denied their existence until shown the text [9].

Musk's attorney Molo also referenced testimony from former employees at Altman's previous startup, Loopt, and former OpenAI board members who characterized Altman as dishonest [1].

Altman's Defense: "An Honest and Trustworthy Businessperson"

Altman did not deny that people have called him a liar. When asked directly whether he always tells the truth, he replied: "I'm sure there are some times in my life when I did not" [1]. When asked whether business associates had called him a liar, he said: "I have heard people say that" [1].

But he pushed back against the characterization as a pattern. "I believe I am an honest and trustworthy businessperson," Altman testified [10]. He argued that the nonprofit structure was never viable for the scale of compute and talent required to pursue artificial general intelligence. Under direct examination, Altman said the nonprofit was "left for dead" after Musk departed in 2018, and that the for-profit restructuring was necessary to attract the capital required to compete [6].

Altman also testified that he never promised Musk the company would remain a nonprofit in perpetuity [11]. He described conversations in which Musk sought 90 percent equity in OpenAI and a board seat at Tesla for Altman — proposals Altman characterized as incompatible with OpenAI's mission [12]. Altman testified that there was a "morale boost" among OpenAI staff when Musk departed the organization [1].

The November 2023 Firing: Five Days That Exposed the Cracks

The events of November 2023 form a critical part of the trial record. On November 17, 2023, OpenAI's board voted to remove Altman as CEO, stating he was "not consistently candid in his communications" with the board [7]. The firing was immediate and set off a corporate crisis.

Within five days, Altman was reinstated as CEO on November 22, after overwhelming pressure from employees — more than 700 of OpenAI's roughly 770 employees threatened to resign — as well as investors and Microsoft, which had invested $13 billion and depended on OpenAI for a growing share of its Azure AI services [13]. The reinstatement came with a reconstituted board: Toner and fellow director Tasha McCauley were removed, and new members including Bret Taylor (now board chair) were installed [7].

The speed of the reversal raised questions about whether the nonprofit board ever had genuine authority over the organization it nominally controlled. The trial record shows that Microsoft, which was not formally on the board, exerted significant pressure during those five days. Testimony has revealed that Microsoft executives "feared being too dependent on OpenAI" and used the crisis to renegotiate elements of their commercial relationship [14].

The Money: $852 Billion and Counting

The financial stakes surrounding this trial are immense. OpenAI closed a $122 billion funding round in March 2026 at a post-money valuation of $852 billion, with Amazon investing up to $50 billion, Nvidia contributing $30 billion, and SoftBank adding another $30 billion [15]. Microsoft holds approximately 27 percent of OpenAI Group PBC on a fully diluted basis — a stake worth roughly $228 billion, representing a 17.6x return on its approximately $13 billion in cumulative investment [16].

Source: OpenAI / CNBC

Data as of Mar 31, 2026CSV

Source: CNBC / Bloomberg

Data as of Mar 31, 2026CSV

SoftBank's total committed capital to OpenAI exceeds $70 billion across multiple rounds, including a $30 billion contribution to the earlier $40 billion Series F round in March 2025 [15]. These investments were structured assuming the public benefit corporation would continue in its current form. A ruling that unwinds the conversion would force renegotiation of those tranches and create significant cash-flow complications [3].

Musk's claimed damages of $130 billion or more are pegged to the value he argues was diverted from the nonprofit mission. Several legal observers note that even if the court finds breach of charitable trust, the damages are unlikely to reach that figure. Most analysts expect "low tens of billions" in potential damages, with structural relief more likely to be addressed on appeal [3].

The Legal Standard: Can Musk Clear the Bar?

The remaining claims — breach of charitable trust and unjust enrichment — present distinct legal challenges. In California, establishing breach of charitable trust requires demonstrating that the trustees of a nonprofit diverted assets or departed from the charitable purpose in a manner that harms the public interest [3]. The question is whether Musk's early donations and communications with Altman and Brockman created a formal "charitable trust" with enforceable terms, or whether they were simply donations to a nonprofit with broad discretionary authority.

Legal scholars are divided. Some argue that the email record — in which Altman and Brockman discussed keeping AI research open and nonprofit-governed — establishes clear representations that Musk relied upon when donating [3]. Others counter that nonprofits routinely restructure and that donors do not generally acquire contractual rights to dictate an organization's future structure [3].

The unjust enrichment claim requires showing that Altman and Brockman personally benefited from the restructuring in a way that was inequitable given the nonprofit origins. The trial record includes testimony about equity grants and compensation packages that the executives received as part of the for-profit conversion [3].

Whistleblowers and the Culture of Silence

The trial has renewed attention on concerns about transparency at OpenAI that predate this lawsuit. In 2024, a group of more than a dozen current and former employees of OpenAI and Google DeepMind published an open letter calling for greater transparency and better protections for whistleblowers at AI companies [17].

Separately, OpenAI whistleblowers filed a complaint with the Securities and Exchange Commission alleging that the company used employment agreements with non-disparagement clauses that failed to exempt disclosures to the SEC, required prior consent for sharing confidential information with federal authorities, and included provisions that could penalize workers who raised concerns with regulators [18]. These agreements also contained clauses requiring employees to waive whistleblower compensation incentives established by Congress [18].

Senator Chuck Grassley introduced the bipartisan AI Whistleblower Protection Act in May 2025, partly in response to reports about OpenAI's practices [19]. OpenAI employees have also described internal pressure to rush safety evaluations — staff reported feeling pressured to accelerate reviews for GPT-4 Omni in spring 2024, with the company planning release celebrations before the preparedness team could determine if the model was safe [20].

The professional consequences for those who raised concerns have been notable. Sutskever left OpenAI after the failed board coup to found Safe Superintelligence Inc. Toner was removed from the board. Several employees who signed the open letter have since departed the company [17].

The Nonprofit Question: Was the Structure Ever Workable?

One of the trial's broader questions is whether OpenAI's nonprofit structure was ever adequate for the task it set out to accomplish. Altman's defense has leaned heavily on this argument: that building frontier AI requires billions of dollars in compute, that nonprofits cannot attract that capital, and that the restructuring was pragmatic rather than duplicitous [6].

There are partial precedents. Mozilla operates the Firefox browser through a for-profit subsidiary (Mozilla Corporation) controlled by the Mozilla Foundation, a structure that has drawn criticism for executive compensation but has not faced equivalent legal challenges [21]. The key difference with OpenAI is scale: Mozilla Corporation's annual revenue is roughly $500 million, while OpenAI's valuation now approaches $1 trillion.

OpenAI's own restructuring has gone through multiple iterations. In October 2025, OpenAI announced that OpenAI Inc. would become the OpenAI Foundation, while the for-profit entities would be reorganized as a public benefit corporation called OpenAI Group [22]. After backlash, the company reversed course in part: the nonprofit now retains a 26 percent financial stake in OpenAI Group and the power to appoint and remove board members [22]. Microsoft holds 27 percent, with the remaining 47 percent distributed among other investors and current and former employees [22].

Legal scholars have questioned whether this arrangement genuinely preserves nonprofit control. Bloomberg Law analysis argued that the conversion "sheds nonprofit purpose without justification" and that applicable law requires the nonprofit to retain fair market value and independence from OpenAI [23]. Others note that the public benefit corporation structure legally requires the board to balance profit with public interest — a meaningful constraint, even if untested at this scale [22].

What Happens Next

Closing arguments are scheduled for this week, with the advisory jury expected to begin deliberating shortly after [4]. Judge Gonzalez Rogers has indicated she will issue her ruling by mid-May 2026.

If Musk prevails, even partially, the consequences ripple outward. OpenAI's planned IPO, rumored for 2027, would likely slip by at least a year due to appellate uncertainty [3]. SoftBank, Amazon, and Nvidia would face renegotiation of investment terms. Congress would likely accelerate legislation requiring enhanced disclosures for nonprofit-to-for-profit conversions [3]. A full unwinding of the PBC conversion — the most extreme remedy Musk seeks — would throw the company's entire capital structure into question, though most legal observers consider this outcome unlikely [3].

If OpenAI prevails, the ruling would validate the nonprofit-to-PBC conversion model and establish a precedent that donors cannot enforce contractual rights after nonprofit transformation [3]. It would strengthen OpenAI's negotiating position for future capital raises and clear the path toward an IPO. Musk's rival AI company, xAI, would lose a narrative advantage in its own fundraising [3].

Whichever side loses will almost certainly appeal to the Ninth Circuit, extending uncertainty into 2027 [3].

The trial has already accomplished one thing neither party may have intended: it has produced the most detailed public record ever assembled of how a nonprofit AI lab became an $852 billion corporation, who made the decisions, and what they told — or didn't tell — the people who were supposed to be watching.