The second week of the Musk v. Altman trial has shifted from the opening phase—where the courtroom was introduced to the personalities, the origin story, and the broad stakes—into something more granular: how OpenAI’s early governance decisions were made, what those decisions meant in practice, and whether the nonprofit structure that exists today is being treated as a mission-protecting mechanism or as a legal wrapper around a rapidly commercializing reality.
At the center of the dispute is a question that sounds simple until you try to litigate it: what did OpenAI’s founders actually promise, and what obligations followed from that promise? Elon Musk argues that OpenAI abandoned its founding mission of developing AI for the benefit of humanity and pivoted toward profit-driven priorities. OpenAI, through its representatives, counters that the lawsuit is baseless—an attempt to derail a competitor—and that Musk’s claims are motivated by rivalry rather than by any genuine effort to enforce a mission.
But the trial is not only about intent. It is also about governance mechanics: who had authority, what constraints were supposed to exist, and how the nonprofit’s role should be interpreted as AI capabilities accelerated. In other words, the case is increasingly about incentives and institutional design—about how an organization built to pursue a public-good objective can end up behaving like a high-stakes market actor, even if it still uses nonprofit language.
This week’s testimony and courtroom activity have reflected that shift. Major OpenAI leadership witnesses have continued taking the stand, including Greg Brockman, with additional testimony scheduled after him. The proceedings have also included a live audio stream at points during the trial, which has amplified public attention far beyond what most corporate or governance disputes typically receive. That streaming element matters in a subtle way: it changes how quickly narratives form, how easily clips circulate, and how much pressure the parties feel to frame their positions in real time. In a case where credibility and interpretation are everything, visibility becomes its own kind of evidence—at least in the court of public opinion.
The trial’s structure also makes it clear that the lawyers are not simply trying to prove “who was right.” They are trying to prove which version of events is legally relevant. Musk’s side is seeking remedies that go beyond damages. According to the case posture described in the reporting, Musk is asking for removal of Sam Altman and Greg Brockman, for OpenAI to stop operating as a public benefit corporation, and for the nonprofit to receive up to $150 billion in damages if he wins. OpenAI’s response frames the suit as a baseless bid to derail a competitor, and it also ties the broader context to Musk’s own ecosystem—SpaceX, xAI, and X—where Grok has been positioned as a competing product to ChatGPT.
That framing is not just rhetorical. It influences how the jury may interpret motive. In many mission-and-governance disputes, the same factual record can support different conclusions depending on whether the plaintiff appears to be enforcing a principle or pursuing leverage. Musk’s testimony earlier in the trial emphasized his interest in founding OpenAI as a humanitarian project—portraying his involvement as “save humanity” rather than as a business play. OpenAI’s counter-narrative, by contrast, suggests that the lawsuit is jealous and strategic, not principled.
Week two has leaned into the details that make those competing narratives testable. Brockman’s testimony, for example, has been used to explore how OpenAI operated and how key relationships and decisions unfolded. The reporting indicates that Brockman’s direct examination included discussion of early days and relationships, including references to Tesla—an unusual detail that underscores how intertwined the early AI ecosystem was. When a company’s origin story includes multiple overlapping networks, the question becomes: were those networks part of a coherent mission-building plan, or were they simply the natural byproduct of building a technology company in a world where capital and talent flow through existing channels?
The courtroom has also been dealing with the practicalities of witness safety and availability. One of the most consequential procedural issues this week is that testimony from Shivon Zilis—described as a former OpenAI board member who shares four children with Musk—may not have audio available on the stream. Lawyers cited threats against her and her children. That decision highlights a grim reality: when high-profile AI governance disputes become public spectacles, the risk does not stay inside the courtroom. It spills outward into the lives of witnesses and their families. For the trial itself, it creates a tension between transparency and protection. For the public, it creates another tension: people want to watch, but the system sometimes has to limit what can be broadcast to keep people safe.
If you zoom out, the trial is also functioning as a referendum on how society should treat “mission” when the organization is simultaneously dependent on markets. OpenAI’s nonprofit structure is often described as a safeguard—an attempt to ensure that the lab’s pursuit of advanced AI remains aligned with public benefit rather than private extraction. Musk’s argument challenges that safeguard’s effectiveness. OpenAI’s argument challenges Musk’s standing to claim that safeguard was violated, and it suggests that the lawsuit is not about mission enforcement but about competitive disruption.
The most interesting part of week two is that the case is increasingly about the meaning of governance documents and the interpretation of early decisions. The reporting indicates that the trial has involved discussions around early OpenAI decisions and relationships, including matters involving Microsoft’s investment timeline. That matters because Microsoft’s involvement is not merely financial; it is structural. When a major platform company invests in an AI lab, it can influence distribution, compute access, product integration, and strategic direction. Even if the nonprofit retains formal constraints, the operational reality can shift. The jury is effectively being asked to decide whether those shifts represent mission drift—or whether they were always part of the path to building frontier models responsibly.
In that sense, the trial is not only about whether OpenAI “changed.” Almost every organization changes as it grows. The legal question is whether the change violates a binding obligation tied to the original mission. That is why the courtroom keeps returning to early governance and ownership questions. The reporting includes references to arguments over ownership and discussions about who would own OpenAI. Those topics sound abstract, but they are the backbone of mission enforcement. If the nonprofit is supposed to protect the mission, then ownership and control determine whether the nonprofit can actually do that job.
Musk’s side has repeatedly emphasized control and the idea that he was promised a certain kind of governance. In earlier testimony, he portrayed himself as demanding the ability to make decisions and framed his concerns as existential—AI risks, extinction risks, and the danger of letting powerful actors control digital superintelligence. Those themes are not just ideological; they are designed to make the jury see governance as a matter of survival, not bureaucracy. If the jury accepts that premise, then governance deviations become more than technical violations—they become moral failures.
OpenAI’s side, meanwhile, has tried to reframe the lawsuit as opportunistic. The reporting includes OpenAI’s statement that the lawsuit has always been a baseless and jealous bid to derail a competitor. That statement is paired with the broader context of Musk’s companies launching Grok as a competitor to ChatGPT. In a trial where motive can color credibility, that pairing is significant. It suggests that even if Musk’s story about mission drift resonates emotionally, the jury may still ask: why now, and why through these specific legal demands?
Week two also shows how the trial’s narrative is being built through small moments of courtroom friction. The reporting includes sharp exchanges and pushing-and-pulling during questioning, along with sidebars and courtroom management as arguments arise. These moments may look like theater, but they often signal deeper disputes about relevance, admissibility, and how the lawyers want the jury to interpret the record. When a lawyer fights hard over a line of questioning, it usually means the line threatens the other side’s framing—either by introducing damaging facts or by undermining the credibility of a witness’s account.
There is also a recurring theme of “what was said” versus “what was done.” In mission disputes, plaintiffs often argue that promises were made and later ignored. Defendants often argue that promises were never binding in the way the plaintiff claims, or that the organization adapted responsibly as circumstances changed. The trial’s focus on early days and early decisions suggests that the jury will be asked to evaluate whether adaptation was legitimate or whether it was a cover for a pivot away from mission.
Another layer that emerges from the reporting is the trial’s relationship to AI safety discourse. While the case is fundamentally about governance and mission, the courtroom has included expert testimony and discussions about AI risks. The reporting mentions that there was a very boring expert witness testifying to AI risks—an aside that signals the presence of technical or risk-focused material. Even when the expert testimony is not dramatic, it can matter legally. It can provide context for why governance constraints exist in the first place. If the jury believes that AI risks are severe and that governance is therefore crucial, then the nonprofit’s role becomes more central. If the jury believes the risk framing is overstated or irrelevant to the legal claims, then the mission argument may lose force.
The trial’s public visibility also affects how people interpret those safety discussions. When AI governance becomes a livestream event, viewers may treat it like a debate show rather than a legal process. But the jury is not deciding which side sounds more persuasive on social media. It is deciding which legal claims are supported by evidence and which remedies are appropriate under the law.
That brings us back to the relief Musk is seeking. Removal of Altman and Brockman is an extraordinary remedy. It implies that the alleged mission breach is not merely a past wrong but an ongoing governance failure requiring immediate correction. Asking OpenAI to stop operating as a public benefit corporation is similarly sweeping. And demanding up to $150 billion in damages for the nonprofit—if Musk wins—turns the case into a high-stakes financial and institutional restructuring event.
OpenAI’s response, as described in the reporting, is that the lawsuit is baseless and competitive