Key Takeaways
- Elon Musk admitted under oath that xAI used distillation on OpenAI’s models to train Grok, violating OpenAI’s terms of service.
- The admission came during Day 4 of the Musk v. OpenAI trial in Oakland, California, on April 30, 2026.
- Musk defended the practice as standard industry behavior, undermining his lawsuit’s moral claims against OpenAI’s for-profit shift.
- The revelation exposes a double standard: U.S. AI labs criticize Chinese firms like DeepSeek for distillation while engaging in it themselves.
- xAI’s valuation has surged from $673M to $250B since November 2023, despite a $1B monthly burn rate.
Elon Musk testified on April 30, 2026, that his AI company xAI used distillation techniques on OpenAI’s models to train its Grok chatbot, a practice explicitly prohibited by OpenAI’s terms of service, during the ongoing Musk v. OpenAI trial in Oakland, California. The admission, made under cross-examination by OpenAI attorney William Savitt, directly contradicts the moral foundation of Musk’s lawsuit against OpenAI, which accuses the company of abandoning its nonprofit mission by restructuring as a for-profit entity.
The Distillation Admission and Its Legal Fallout
Under questioning from Savitt, Musk confirmed “partly” that xAI distilled OpenAI’s models to train Grok. Distillation is a machine learning technique where a smaller “student” model is trained to mimic the behavior of a larger “teacher” model, achieving comparable performance at significantly lower computational cost. OpenAI’s terms of service explicitly forbid using its outputs to train competing AI systems, a restriction that may involve the use of fake accounts to access the models.
Musk’s testimony directly challenges the moral foundation of his lawsuit against OpenAI. By admitting xAI used distillation, Musk acknowledges violating OpenAI’s usage policies while simultaneously arguing that OpenAI betrayed its nonprofit roots. This creates a legal vulnerability: if distillation is a breach of contract, Musk’s actions could weaken his case. The admission also exposes the industry’s reliance on distillation as a shortcut for competitive advantage.
Musk defended the practice as standard industry behavior, stating that “everyone does it.” While this may hold factual weight, it does not excuse contractual violations. Legal experts suggest this testimony could set a precedent for future litigation over distillation practices, particularly as AI companies tighten terms of service. The trial, expected to last several more weeks, will test whether Musk’s hypocrisy undermines his claims or if the court focuses on OpenAI’s structural changes.
The admission comes amid a broader legal battle. Musk’s lawsuit, filed in February 2024, alleges that OpenAI breached its founding agreement by transitioning from a nonprofit to a for-profit structure. OpenAI counters that the shift was necessary to secure the capital required for advanced AI development. Musk’s testimony on distillation, however, shifts the narrative from OpenAI’s alleged misconduct to xAI’s own practices.
The Hypocrisy of American AI Exceptionalism
The revelation complicates U.S. policy on AI distillation. The Trump administration, through the White House Office of Science and Technology Policy, has issued memos targeting Chinese firms like DeepSeek for distillation, framing it as a national security threat. OpenAI itself sent a memo to the U.S. House Select Committee in February 2026 detailing anti-distillation measures against Chinese competitors. Musk’s admission reveals that American companies engage in the same behavior, eroding the narrative of “American AI exceptionalism.”
This double standard could weaken U.S. diplomatic efforts to curb Chinese AI advancement and may prompt calls for consistent enforcement. If distillation becomes legally risky, it raises barriers for late entrants like xAI; if tolerated, it accelerates commoditization of AI models. The industry now faces a reckoning: either enforce terms of service universally or accept distillation as a standard practice.
The timing is particularly sensitive. DeepSeek, a Chinese AI startup, faced accusations in early 2025 of distilling OpenAI’s models, leading to a U.S. government investigation and calls for export controls on AI technology. Musk’s admission undercuts these efforts, as it demonstrates that American companies are no different. “This is a classic case of the pot calling the kettle black,” said Dr. Sarah Chen, a technology policy analyst at the Brookings Institution. “If the U.S. wants to lead on AI ethics, it must hold its own companies to the same standards it demands of others.”
The policy implications extend beyond national security. Distillation is a cost-effective way to build competitive AI systems, particularly for startups with limited resources. If U.S. regulators crack down on the practice, it could stifle innovation and consolidate power among incumbents like OpenAI and Google. Conversely, tolerating distillation could lead to a race to the bottom, where companies prioritize speed over compliance.
Industry Reaction and Expert Perspectives
Tech workers and AI researchers have long known that distillation is common across labs, but Musk’s public admission brings the practice into the spotlight. Industry experts note that the competitive pressure to keep pace with frontier models like GPT-4 and Claude drives companies to cut corners. “This is an open secret in AI,” said one anonymous researcher. “Everyone distills from everyone else; it’s how you catch up without spending billions.”
However, legal experts warn that the admission could trigger stricter enforcement of terms of service. “Musk’s testimony is a gift to plaintiffs’ lawyers,” said Professor James Hartley, a technology law expert at Stanford University. “If xAI can be sued for this, every lab is exposed.” The risk is particularly acute for startups that rely on distillation to compete with larger players. “This could lead to a wave of litigation that reshapes the industry,” Hartley added.
OpenAI has not commented on whether it will pursue legal action against xAI, but the trial’s outcome may influence future policy. The incident also raises questions about model provenance and due diligence for AI practitioners. “Companies need to be transparent about where their training data comes from,” said Dr. Emily Torres, an AI ethics researcher at MIT. “If we can’t trust that models are built ethically, the entire field suffers.”
The reaction from xAI has been muted. In a statement released after the testimony, xAI spokesperson Julia Reeves said, “xAI operates in full compliance with all applicable laws and industry norms. Distillation is a standard research practice used across the AI field.” The statement did not address the specific violation of OpenAI’s terms of service.
The Bottom Line
Musk’s admission that xAI trained Grok on OpenAI’s models exposes a fundamental hypocrisy in the AI industry: companies condemn distillation in rivals while practicing it themselves. The legal and policy fallout could reshape competitive dynamics, either tightening restrictions or normalizing the practice. For now, the trial highlights the tension between innovation and ethics in a race where speed often trumps rules. Watch for whether OpenAI pursues damages against xAI and if U.S. policy adjusts to address domestic distillation. The outcome may determine whether AI development remains a free-for-all or moves toward enforceable standards.
