A US District Judge has dismissed xAI’s lawsuit accusing OpenAI of stealing trade secrets by poaching employees” —ruling that hiring from a competitor is not the same as stealing trade secrets, even when some of those employees took files with them.

Judge Rita F. Lin found that xAI failed to show any evidence that OpenAI actually acquired, used, or benefited from the alleged trade secrets. The case highlights the high bar for proving trade secret theft in the AI industry, where talent moves quickly between companies.

What xAI Alleged

According to xAI’s complaint, several former employees engaged in suspicious conduct before leaving for OpenAI:

• Two employees admitted to downloading xAI source code before their departures
• One employee improperly recorded a Musk “All Hands” meeting
• One engineer uploaded xAI’s entire source code to a personal cloud connected to ChatGPT before interviewing at OpenAI
• A finance executive responded hostilely when xAI asked about confidentiality concerns

xAI argued this conduct, combined with OpenAI’s recruiting efforts, suggested a coordinated effort to steal trade secrets.

Why the Judge Dismissed It

Judge Lin’s ruling hinged on a fundamental principle: you have to prove the secrets were actually used, not just taken.

“Suspicious timing, aggressive recruiting, and even downloaded files are not enough on their own,” said Sarah Tishler, a commercial litigator who spoke to Ars Technica about the case.

The key problems with xAI’s case:

• The engineer who uploaded source code never worked at OpenAI. Xuechen Li had his job offer withdrawn after xAI secured a separate injunction blocking him from joining. Since he never started, he couldn’t have used xAI secrets there.
• Another engineer deleted files before joining. Jimmy Fraiture admitted to copying xAI data but said he deleted it before his first day. There’s no evidence he used it at OpenAI.
• No evidence OpenAI knew about or requested the files. xAI pointed to a recruiter texting “nw!” to Li after the upload, claiming it meant “no way!” (excitement). OpenAI said it meant “no worries.” Even accepting xAI’s interpretation, Lin found no proof the recruiter accessed or requested the files.
• Hostility isn’t evidence. The former executive’s vulgar response to xAI’s questions was rude, but “hostility toward one’s former employer during departure does not, without more, indicate use of trade secrets in a subsequent job,” Lin wrote.

What Happens Next

xAI can amend its complaint by March 17. Judge Lin allowed the company to refile if it can show that OpenAI actually received and used xAI trade secrets. But xAI cannot add new claims or parties.

Separate lawsuits continue. The case against engineer Xuechen Li proceeds separately, and Li faces additional pressure from an FBI criminal investigation into whether he stole trade secrets.

The broader feud continues. OpenAI characterized this lawsuit as part of Musk’s “ongoing campaign of harassment” against his former company.

Why This Matters for the AI Industry

The ruling is significant for an industry where talent moves quickly and competitive stakes are high:

• Hiring is still legal. Companies can recruit from competitors without fear of trade secret liability” —as long as they don’t ask for or use stolen information.
• The burden of proof remains high. Trade secret cases require showing actual use, not just opportunity or suspicious behavior.
• Employees bear individual risk. While OpenAI escaped liability, the employees who took files still face potential legal consequences” —Li is under FBI investigation.

The Bigger Picture

This ruling is one front in an ongoing conflict between Musk and OpenAI. Musk co-founded OpenAI in 2015 but left before it became the company behind ChatGPT. He has since sued OpenAI over its transition from nonprofit to for-profit, alleged the company abandoned its founding mission, and accused it of anticompetitive behavior.

Through xAI, Musk has launched Grok as a