Judge dismisses xAI trade secret suit against OpenAI

U.S. District Judge Rita Lin dismissed xAI's trade secret lawsuit against OpenAI Monday, June 15, in San Francisco, finding that Elon Musk's AI company did not show OpenAI induced former xAI engineer Xuechen Li to disclose confidential Grok information.

Lin dismissed the case with prejudice, according to Reuters, which bars xAI from filing another amended complaint in the case. She had dismissed an earlier version in February.

The suit targeted recruiting conversations around Li, a former senior engineer at xAI. xAI said OpenAI sought information tied to Grok 4, which xAI released in July 2025, and accused OpenAI of seeking an edge in reasoning, reinforcement learning and post-training work.

Lin rejected that theory. She said employers ask job candidates about prior work as part of routine hiring, and xAI did not show OpenAI pushed Li to leak confidential material or that OpenAI engineers knew he might have disclosed protected information.

OpenAI said Li did not work for the company and said it never acquired xAI trade secrets. Reuters reported that xAI and its lawyers did not respond to requests for comment.

The ruling gives OpenAI another legal win in its long conflict with Musk. On May 18, a federal jury rejected Musk's $150 billion lawsuit accusing OpenAI and CEO Sam Altman of betraying OpenAI's nonprofit mission.

For AI companies, the case points to a hiring risk that has grown as labs compete for engineers who understand frontier model training. Recruiting from rivals can give a company domain experience, but litigation can follow if a former employee worked on model architecture, source code, training data, alignment methods or launch plans.

The order also draws a line around trade secret claims in AI hiring. A company must show more than a rival's interest in a candidate's past work. It must connect the alleged secret, the disclosure and the recipient's conduct.

That standard matters because AI labs now treat talent as a strategic asset. OpenAI, xAI, Anthropic, Google DeepMind and Meta compete for researchers who have worked on large-scale training systems and post-training pipelines. Those workers carry experience in their heads. Courts still require plaintiffs to identify protected information and show that someone used or solicited it.

xAI remains a central piece of Musk's AI strategy alongside his broader companies. Grok serves as xAI's consumer chatbot and a visible rival to ChatGPT. OpenAI, meanwhile, keeps ChatGPT at the center of its consumer and enterprise push as competitors try to close gaps in reasoning, coding and multimodal tools.

Lin's dismissal reduces one legal overhang for OpenAI, though it does not end Musk's broader fight with the company. Reuters reported that xAI continues to sue Li in a separate case, and Li denies wrongdoing.

The business impact rests on talent movement. AI labs can keep recruiting from competitors, but legal teams will keep tightening interview rules, offer letters and onboarding checks. Companies that hire from rivals need clean records that show candidates did not bring code, internal documents or model details from prior employers.

The case also shows how hard trade secret suits can become in AI. Modern model development blends code, training choices, data handling, evaluation methods and research judgment. A plaintiff that claims theft must name the secret with enough precision for a judge to distinguish protected material from skill and experience.

Reuters reported the ruling Monday in a story by Jonathan Stempel. The dispute involves OpenAI, xAI and Grok, xAI's chatbot product.