Court filings in the Oakland federal trial between Elon Musk and OpenAI allege that Musk sent messages to Greg Brockman and Sam Altman warning they would become "the most hated men in America" if they pursued the nonprofit-to-commercial conversion, threatened to fund competing AI safety organisations that would publicly attack them, and made settlement overtures to Brockman that OpenAI's legal team characterised as attempts to peel away a key witness while litigation was pending, a pattern of conduct OpenAI's lawyers described as systematic intimidation of current and former leadership designed to destabilise governance during a critical corporate restructuring period.
The evidentiary question is the one that matters before the rhetoric, and it requires separating what has been filed from what has been alleged in opening statements. The "most hated men in America" language appeared in Musk's own communications, which OpenAI's legal team presented as exhibits rather than characterised from memory. The settlement overture to Brockman is described in court filings referencing a Telegram message from Musk that Brockman received and preserved. Neither of these is contested as fabricated. The dispute is about what they mean: Musk's legal team has argued that his communications reflected genuine concern about what he saw as a betrayal of OpenAI's founding mission and that settlement discussions are standard litigation practice, not witness tampering. OpenAI's team argued that the combination of settlement overtures, public reputational threats, and a lawsuit filed against individuals by name rather than against the institution amounts to a pressure campaign intended to create uncertainty about the personal liability of anyone who participated in the governance decisions Musk is contesting. The jury will ultimately decide which characterisation is more consistent with the documents and testimony presented.
The legal discovery record this trial is producing is more broadly illuminating than either party's preferred framing. Documents introduced through the SC 13D and 8-K process, through deposition excerpts, and through the exhibits presented during the first week of testimony have provided the most detailed public accounting of OpenAI's internal governance decisions from 2015 through 2024 that has ever been available. Musk's original email pledging $1 billion to OpenAI and describing it as a counterweight to DeepMind, Sam Altman's early communications about the tension between the nonprofit structure and the capital requirements of frontier AI research, Brockman's journal entry about "flipping" OpenAI for profit, and the board communications around Sam Altman's November 2023 firing and reinstatement are all now part of the public record in ways that the companies involved would not have chosen. That record is not just a courtroom document. It is a map of how the frontier AI industry made its most consequential governance decisions under conditions of extreme uncertainty about both technology timelines and legal constraints.
The founder conflict dynamics this trial exposes are worth examining honestly because they are not unique to Musk and Altman. What the evidence describes is a company founded on a mission statement, funded by a small group of powerful individuals with divergent views about how that mission should be executed, and then forced to make structural choices about control, equity, and commercial viability that some of those founders never agreed to. That scenario repeats across thousands of venture-backed companies every year at smaller scale. The version at OpenAI is distinguished by the size of the stakes, the visibility of the personalities involved, and the strategic significance of the outcome to an industry worth several trillion dollars. But the underlying dynamics, a founding-era member who retained reputational investment without retaining governance rights finding themselves unable to influence decisions they believe are wrong and turning to litigation as the only available lever, are recognisable to anyone who has lived through a serious founder dispute at a company they helped build.
The IPO and public offering context that Greg Brockman disclosed under oath last week sits directly underneath the trial's most contested questions. Musk's lawsuit alleges that the nonprofit-to-commercial conversion is a breach of the charitable trust under which OpenAI was founded, that the conversion improperly enriches insiders including Altman and Brockman at the expense of the public benefit mission, and that the process was conducted without adequate independent board oversight. If the conversion was improper, Brockman's $30 billion stake in the resulting for-profit entity is the product of that impropriety, which is why Musk's lawyer spent significant examination time on Brockman's personal financial interests. The IPO disclosure adds a layer: if OpenAI proceeds toward public markets, the conversion must survive judicial scrutiny and California Attorney General approval, and any finding in the trial that the conversion process was procedurally or substantively flawed creates legal risk for the offering that investment banks and institutional investors will require to be resolved before underwriting the deal. The trial is not just about Musk and Altman. Its outcome has a direct bearing on whether OpenAI can reach public markets on its current timeline and at its current implied valuation.
What founders can take from this trial is not primarily about what to do when a former co-founder sues you, though that lesson is embedded in the record. The more useful lesson is about control documentation at inception. Every agreement Musk made with OpenAI was in writing, which is why the discovery process has been so productive. His $44.7 million in donations, his board service, his early influence over hiring decisions, and his departure terms are all documented in ways that allowed OpenAI's lawyers to reconstruct a timeline the jury can evaluate. What is not clearly documented is the nature of the binding commitments he was given about OpenAI's nonprofit permanence, if any such commitments were made. Musk's legal theory depends on showing that he donated money and time in reliance on representations that the nonprofit structure was permanent, and that the conversion violated those representations. OpenAI's defence has argued that no such binding commitment was ever made and that Musk's expectations were based on his own interpretation of the founding documents rather than contractual promises. The gap between what founders believe they agreed to and what is actually documented in binding form is where most serious founder disputes live. In this case, that gap is being argued in a federal courtroom, with a jury deciding which version of institutional memory is more credible.
The AGI competition context provides the final frame. Neither side disputes that OpenAI is pursuing artificial general intelligence or that the stakes of that pursuit are extraordinary. Musk founded xAI explicitly to compete in the same race. The trial's subtext is that two of the most consequential AI projects in the world are now led by people in active litigation with each other, and that the legal proceedings are generating public documentation about how frontier AI decisions get made that will influence regulation, governance standards, and investor expectations for years after the jury returns its verdict. Whatever outcome the trial produces, the documents it has introduced are now permanent features of the AI industry's institutional record.
","excerpt":"Court filings in the Oakland Musk versus OpenAI trial allege that Musk sent messages warning Altman and Brockman they would become "the most hated men in America" if they pursued the commercial conversion, and made settlement overtures to Brockman that OpenAI's lawyers characterised as witness intimidation.
Also read: OpenAI Just Published Its WebRTC Infrastructure Playbook for Voice AI and Founders Should Read It as a Competitive Signal, Not a Tutorial • Dairy Queen Is Pausing Middle East Expansion While Deploying AI at 50 Drive-Thrus and the Two Decisions Are More Connected Than They Appear • Developers Are Asking If Codex Is the Best Coding Agent Right Now and the Answer Depends Entirely on What You Mean by Best
