X.AI Researcher Stole Trade Secrets to Take To OpenAI, Suit Says

Aug. 29, 2025, 4:22 PM UTC

A former researcher at Elon Musk’s X.AI admitted he stole trade secrets regarding its AI platform Grok to bring with him to rival OpenAI Inc., according to a new lawsuit.

Accomplished AI researcher Xuechen Li suddenly quit X.AI three days after selling employee stock shares totaling roughly $7 million, the complaint filed Thursday in the US District Court for the Northern District of California said. Once confronted, Li—one of the company’s first engineers—admitted in writing with a criminal attorney present that he stole secrets and tried to cover his tracks, X.AI said.

The complaint underscored the stakes in the “race to dominate the AI landscape,” as it said the technology will be worth hundreds of billions of dollars this year and top $1 trillion by the end of the decade. The confidential information allegedly misappropriated by Li could save a competitor billions and years in research and development, the complaint said.

Li, who earned a Ph.D. in computer science in 2024, joined X.AI in February of that year as part of an initial group of roughly 20 engineers, according to the complaint. He began developing and training Grok in a bid to catch and surpass OpenAI’s ChatGPT, which launched in 2022 after seven years of development, X.AI said.

The company provided stock equity as part of his compensation, and because it “valued his contributions, and wanted to retain him as a productive and successful employee,” allowed him to liquidate millions in stocks. He received the cash from sales on July 23 and July 25, the complaint said.

The same day he sold the second tranche of stocks, he extensively copied trade secrets onto his own devices, the complaint said, and three days later he resigned, having accepted a position with OpenAI with an Aug. 19 start date.

The complaint doesn’t name OpenAI as a defendant or allege that Li ultimately worked there.

On Aug. 11, X.AI discovered the theft during a routine review of security software logs and demanded Li return and destroy confidential data stolen in violation of his contractual obligations, the complaint said. Li instead retained criminal defense counsel and confessed, during Aug. 14 and 15 in-person meetings and in a handwritten note, “in an attempt to talk his way out of facing consequences for his theft.”

After further negotiations Li allegedly turned over devices and authorized X.AI to pore over them on Aug, 18, but he had changed a password Aug. 11 for a critical account with downloaded secrets, “thwarting the very purpose of the authorization” agreement. X.AI also said it discovered accounts Li didn’t disclose.

Li and OpenAI didn’t immediately respond to written requests for comment.

Winston & Strawn LLP represents X.AI.

The case is X.AI Corp. v. Li, N.D. Cal., No. 3:25-cv-07292, complaint filed 8/28/25.

To contact the reporter on this story: Kyle Jahner in Raleigh, N.C. at kjahner@bloomberglaw.com

To contact the editor responsible for this story: Adam M. Taylor at ataylor@bloombergindustry.com

Learn more about Bloomberg Law or Log In to keep reading:

Learn About Bloomberg Law

AI-powered legal analytics, workflow tools and premium legal & business news.

Already a subscriber?

Log in to keep reading or access research tools.