Judge Tosses xAI's Trade-Secret Suit Against OpenAI

A US federal judge has thrown out xAI's lawsuit accusing OpenAI of stealing trade secrets, and did so 'with prejudice' — meaning Elon Musk's company cannot refile the same claim. Judge Rita F. Lin ruled on June 15 that xAI had 'failed to sufficiently allege a connection' between OpenAI and the alleged misconduct of former xAI employees.

xAI first sued in September 2025, claiming OpenAI had persuaded ex-employees to walk out the door with confidential information about Grok, its chatbot. The court dismissed an earlier version of the complaint in February 2026; xAI filed an amended version, and the judge has now rejected that one too — this time permanently. xAI had also sued one of its own former employees separately over the same alleged leaks.

Another courtroom loss for Musk

This is the latest setback in Musk's long-running legal campaign against the company he co-founded and then left. Earlier in 2026, a jury found that the statute of limitations had run out on his separate breach-of-contract claims against OpenAI. The trade-secret theory was a different angle — that OpenAI poached talent specifically to get at Grok's secrets — but the judge concluded xAI never connected the dots between OpenAI as a company and whatever individual employees may have done on their own.

Why should anyone outside the courtroom care? Because the AI talent market is a revolving door, and these suits are really about who controls the knowledge that walks between labs. A dismissal 'with prejudice' sets a quietly useful bar: changing jobs and bringing your general expertise is not, by itself, trade-secret theft — a plaintiff has to show the new employer actually directed the misappropriation. My take is that this is healthy for the field. If every engineer who switched employers could be frozen by litigation, the rapid back-and-forth that pushes models forward would grind to a halt.