A judge has shut down part of xAI’s case against OpenAI, rejecting the claim that the ChatGPT maker tried to steal trade secrets by recruiting a former Grok developer. The ruling gives OpenAI an early win in a legal fight that is as much about talent poaching as it is about intellectual property.
The dispute centers on Xuechen Li, a former senior xAI engineer who worked there from 2024 to 2025. xAI argued that Li was preparing to pass along confidential material tied to Grok’s development after being approached by OpenAI, but the court said that simply discussing prior experience in an interview is standard hiring behavior, not evidence of espionage.
Why the judge rejected xAI’s trade-secret claim
Judge Rita Lin found that xAI had not shown OpenAI encouraged Li to reveal trade secrets or that OpenAI staff knew he could do so. That distinction matters: in Silicon Valley, hiring across rivals is routine, and courts are usually reluctant to turn every résumé pitch into a secret-stealing conspiracy.
xAI had pointed to a presentation Li gave while OpenAI was trying to hire him, and to his claim of experience with reinforcement learning and post-training for AI models. But the argument appears to have run into a basic problem: the mere fact that a company wants a skilled engineer does not prove it wants the engineer’s former employer’s files.
OpenAI’s counterattack on talent poaching
OpenAI told the court that Li never worked for the company and that it never received xAI trade secrets. It also took the sharper route, saying xAI is losing the competition and cannot keep its staff from walking out the door. That’s corporate mudslinging, yes, but it also reflects a familiar AI-era reality: the real prize is often the people, not just the code.
- Li worked at xAI from 2024 to 2025.
- xAI said OpenAI tried to gain access to Grok-related secrets through him.
- The court said interview discussions about skills are normal.
- OpenAI said it never got xAI trade secrets.
The fight is not over for Xuechen Li
Li denied any role in trying to hand over xAI’s confidential material, and his former employer is still pursuing him in a separate case. That leaves the bigger question unresolved: how far can AI companies go in policing employee mobility before every hire turns into litigation? For now, at least, xAI has learned that suspicion alone is not enough to make a trade-secret claim stick.