The ongoing legal battle between Elon Musk and Sam Altman over the future of OpenAI reached a pivotal moment in court on May 14, 2026, as attorney Eddy argued that firsthand documents contradict Musk's testimony. "The documents tell the truth here," Eddy stated, drawing a clear line between what records show and what Musk claims about the early structure of the artificial intelligence company.
At the center of the dispute is the notion of an "adjunct" for-profit entity, a term Musk used twice during his testimony. However, when opposing counsel Savitt attempted to use the phrase, Molo objected, accusing Savitt of fabricating terminology. Eddy emphasized that no such term appears in any of the brainstorming structure documents from OpenAI's founding. Instead, the documents reference parallel for-profit entities and the possibility of fully converting the nonprofit into a for-profit company while shutting down the nonprofit arm entirely.
This distinction is critical because Musk's lawsuit alleges that Altman and OpenAI's board violated the nonprofit's mission by allowing a for-profit subsidiary to exist. The court must determine whether the original founders ever intended for a hybrid structure that Musk now claims was always part of the plan. Eddy's argument suggests that Musk's recollection is not supported by written evidence, potentially weakening his case.
Jared Birchall, a former executive at OpenAI, also testified that he personally filed paperwork to register a company that would have taken over the nonprofit's operations. This aligns with Eddy's assertion that the documents show a clear plan for a complete conversion rather than an adjunct arrangement. Birchall's testimony adds weight to the idea that the board considered and even began executing a shift to a for-profit model, but the legal question remains whether such a shift violated the original charter.
Background of the Dispute
The court battle began in early 2025 when Elon Musk, a co-founder of OpenAI who left the company in 2018, filed a lawsuit against Sam Altman and the OpenAI board. Musk alleged that Altman had steered the organization away from its nonprofit mission of developing AI for the benefit of humanity. The lawsuit claims that the creation of a capped-profit subsidiary and subsequent moves toward a full for-profit conversion were illegal breaches of fiduciary duty and contract.
OpenAI was founded in 2015 as a nonprofit research organization with a pledge to develop AGI safely and openly. Musk contributed significant funding in the early years but departed after disagreements over the company's direction. Under Altman's leadership, OpenAI created a for-profit subsidiary in 2019 to attract investment, leading to partnerships with Microsoft and other tech giants. The company has since released powerful AI models like GPT-4 and DALL-E 3, generating substantial revenue but also raising concerns about commercialization.
Musk's lawsuit seeks to block OpenAI from completing a proposed full conversion to a for-profit entity, which would allow investors to take equity. The case has drawn immense attention because it could set legal precedents for the governance of AI companies and the enforceability of nonprofit missions.
Key Testimony and Document Evidence
During the hearing, both sides clashed over the interpretation of early documents. Eddy entered several exhibits from 2016 and 2017, including email chains and white papers that explored various corporate structures. One document proposed a "parallel for-profit" that would operate alongside the nonprofit but be separately governed. Another explored a "complete conversion" in which the nonprofit would dissolve and its assets would be transferred to a new for-profit entity.
Musk's lawyers argued that the concept of an adjunct for-profit was implicitly understood as a middle ground between these two extremes. But Eddy pointed out that none of the documents contain the word "adjunct" and that the term first appeared only in Musk's recent deposition. "If it was such a central idea, we would expect to see it somewhere in writing from the relevant period," Eddy told the court.
Birchall's testimony corroborated the document evidence. He stated that he filed incorporation papers for a new company in 2017, well before any for-profit subsidiary was officially created. Birchall said he understood the plan was to move all commercial operations into the new entity while the nonprofit would focus on research grants and public education. This suggests that the board was actively preparing for a full for-profit structure even while publicly maintaining that OpenAI remained a nonprofit.
The judge pressed both sides on whether such preparations violated the trust of donors who gave money to the nonprofit. Many early donors, including Musk himself, had contributed on the understanding that OpenAI would never become a profit-maximizing entity. If the court finds that the board misrepresented its intentions, it could order financial damages or halt the conversion.
Implications for the AI Industry
The outcome of this trial could reshape how AI research organizations are structured. Many AI companies have followed OpenAI's model of starting as a nonprofit and later creating for-profit arms. For example, Anthropic, founded by former OpenAI employees, has both a nonprofit and a public-benefit corporation. If the court rules that such dual structures are inherently deceptive, it could force many companies to restructure or face lawsuits from early donors.
Moreover, the case highlights the tension between open research and commercial secrecy. OpenAI originally promised to make its research publicly available, but its for-profit arm has increasingly withheld details to protect competitive advantages. Critics argue this violates the spirit of the original mission, while supporters point out that massive compute costs require revenue generation.
During cross-examination, Altman testified that the for-profit subsidiary was always intended as a vehicle to raise capital, not to supplant the nonprofit's goals. He claimed that the nonprofit board still has authority to veto any decisions that harm the public interest. Musk's attorneys countered that the board is stacked with Altman allies and lacks independence.
The legal argument over the "adjunct" term may seem semantic, but it cuts to the heart of the case. If the original founders never envisioned a hybrid model, then Musk's claim that Altman deviated from the plan is weakened. Conversely, if Musk can show that some form of adjunct was discussed even if not documented, he may salvage part of his case.
Forward-Looking Analysis
Legal experts following the trial note that document evidence is often more persuasive than oral testimony, especially when witnesses' memories are called into question. Eddy's focus on the absence of the word "adjunct" appears to be a strategic move to discredit Musk's entire narrative. However, Musk's legal team has argued that documents do not capture every informal conversation. The judge will likely weigh the credibility of each side's witnesses.
One piece of evidence that could tip the scales is a series of emails from 2017 in which Musk himself proposed a for-profit arm for OpenAI. Musk allegedly backtracked on that idea after leaving the company. If the court sees those emails, it could undermine his current position. The trial is expected to continue for several more weeks, with additional witnesses including former board members and investors.
Regardless of the verdict, the case is already influencing how AI companies communicate their missions. Some have begun adding explicit clauses in their bylaws that allow for-profit transitions under certain conditions. Others are strengthening the independence of their nonprofit boards to avoid accusations of mission creep.
Eddy's statement—"The documents tell the truth here"—may become a memorable line if it leads to a favorable ruling for OpenAI. But for now, the court must sift through hundreds of pages of evidence to decide who owns the future of one of the most transformative technologies in history.
Source: The Verge News