Musk’s Oakland Testimony: Betrayal Claim Bolstered by Big Damages — and Undercut by His Own Company’s Use of OpenAI Outputs
Elon Musk spent his first week on the stand in Musk v. Altman accusing OpenAI leaders of “stealing a charity” after converting the organization he helped found into a for‑profit engine — but his testimony also confirmed xAI used distillation from OpenAI outputs, a fact that complicates his legal posture. The suit seeks up to $134 billion, reversal of OpenAI’s for‑profit restructuring, and removal of Sam Altman and Greg Brockman from leadership.
Musk’s accusation, the financial stakes, and a rejected takeover
Musk testified that OpenAI’s pivot after ChatGPT’s 2022 launch and Microsoft’s $10 billion investment turned a nonprofit into a commercial enterprise that unjustly enriched its leaders; he framed that shift as a breach of the founding promise from 2015, when he donated roughly $38 million. His complaint asks the court to unwind the corporate changes and demands as much as $134 billion in damages, an amount that signals both legal ambition and the potential market value at stake.
Records filed around the case show Musk tried to regain control before suing: he submitted a $97.4 billion acquisition offer for OpenAI in early 2025, which Sam Altman rejected, and reached out to Greg Brockman days before trial to propose a settlement that was also rebuffed. Those bargaining moves underline the control battle behind the courtroom rhetoric and help explain why the trial has intense industry attention in Oakland federal court.
Admissions about xAI training and the narrow technical contradiction
Under cross‑examination Musk conceded that xAI — which merged with SpaceX in early 2026 and is publicly reported to carry a multibillion-dollar valuation — used a distillation process that relied in part on OpenAI outputs to train its models, a technique OpenAI’s terms of service generally prohibit. Musk described the validation step as “standard practice,” but that admission directly touches on OpenAI’s claim that competitors should not be able to bootstrap from its models while denying access to their own training pipelines.
OpenAI’s defense points to that admission and to Musk’s earlier behavior: internal evidence and filings show Musk pushed for for‑profit mechanisms in OpenAI’s evolution and then, after leaving the board in 2018, returned as a litigant only after founding xAI in 2023. The company sought to introduce earlier messages as evidence of competitive motive; the judge rejected that late filing, but the sequence of events — past advocacy, later rivalry, and now litigation — is central to OpenAI’s portrayal of Musk’s incentives.
How the judge is keeping the case about governance, not AI doom scenarios
Judge Yvonne Gonzalez Rogers has repeatedly curtailed wide‑ranging safety arguments, instructing both sides to avoid framing the trial as a debate over existential AI risk; she specifically barred testimony and argument that stray into cinematic doomsday analogies after Musk invoked “The Terminator” style warnings. The court’s focus is procedural and statutory: whether nonprofit governance promises were broken, not whether AI presents an existential threat to humanity.
That limitation changes what evidence is legally relevant — for example, internal intent documents about corporate structure or board actions are squarely in play, whereas speculative warnings about future risk are not. The ruling also concentrates the battleground on documentary trails and contracting choices made around the 2019–2023 period when OpenAI’s structure and external funding were actively debated.
Immediate checkpoints and practical governance implications
The trial’s near‑term pivot is Greg Brockman’s expected testimony and the possible admission of his journals and internal communications, which could provide contemporaneous intent on conversion to for‑profit status. Stakeholders in AI governance should watch three concrete signals from the next phase: whether Brockman’s records show deliberative intent to commercialize, whether Musk’s distillation evidence is corroborated by logs or contracts, and whether the court accepts new documentary evidence tied to control bids or settlement offers.
| Signal | What it could prove | Immediate consequence |
|---|---|---|
| Brockman’s journals | Contemporaneous intent on corporate structure and funding choices | May tip factual dispute over whether nonprofit promises were consistently honored |
| Musk’s distillation admission | Demonstrates technical dependence or competitive overlap | Weakens the narrative of pure principle; raises contract and damages questions |
| Control moves (2025 bid, settlement attempts) | Shows bargaining intent and timeline of dispute escalation | Frames remedies as corporate‑control disputes rather than abstract harms |
If Brockman’s records show explicit plans to monetize or if documentary evidence confirms systematic reliance by competitors, the case could produce enforceable checkpoints for how nonprofit‑origin AI labs document governance and donor expectations before pursuing commercial deals or IPOs. Conversely, if the evidence skews toward ordinary competitive behavior with contractual remedies, the court may limit remedies to damages rather than unwinding corporate structure — a distinction that will matter to founders, donors, and regulators monitoring how nonprofits convert or partner with venture capital and corporate investors.
