Key Takeaways:
Apple's forensic analysis of a former employee's MacBook shows OpenAI used Apple's confidential circuit schematics to train AI agents, escalating the trade secrets lawsuit filed in July.
Apple submitted new forensic evidence Monday showing a former employee used a confidential circuit schematic to train an AI agent at OpenAI, escalating the July trade secrets lawsuit against employees Chang Liu and Tang Tan.
"Mr. Liu not only downloaded a confidential Apple circuit schematic but also used it in his work at OpenAI," Apple said in the filing, which also alleges Liu sent instructions for destroying evidence to an OpenAI colleague who confirmed she would comply.
Liu downloaded the schematic in March, two months after leaving Apple, and used it to run power conversion simulations with LTspice, an electrical engineering tool. He described training an AI agent to run the software, writing that "my AI agent learned how to run LTspice, look at result, tune compensation parameter." The schematic was synced from a Mac mini to the Apple-issued MacBook Liu kept after departing.
The case, set for argument before Judge Edward J. Davila on October 1, could set precedent for how AI companies source training data and proprietary information. OpenAI has called the lawsuit "careless, aggressive and oddly personal" and asked the court to dismiss it, while Apple says the "shocking evidence" proves its trade secrets are in active use.
Apple's filing details four key findings from the forensic analysis. Beyond the schematic download, Apple alleges that OpenAI employees were aware of Liu's unauthorized access to Apple's third-party cloud storage. Liu also used a tool at OpenAI with the same name as an internal Apple engineering application used for development work.
The filing describes Liu messaging colleagues about his work, writing "[f]eeling AI all day long" and that "in the past hour" his AI agent "learned how to run LTspice, look at result, tune compensation parameter." When a colleague replied "Why do they even need you then?" the exchange showed how Liu's AI training work relied on Apple's proprietary data. Liu's messages suggest he used the Apple schematic to train an AI agent for power conversion development work, a task he said previously took a day but now took two hours with the AI agent.
The lawsuit sits within a larger context: OpenAI has poached more than 400 employees from Apple to work on AI projects, according to Apple's filings. The case touches on the intensifying competition between the two companies as Apple builds out its own AI capabilities while OpenAI expands into consumer-facing products. Apple initially requested expedited discovery earlier this month, also seeking an injunction preventing OpenAI from using its trade secrets. OpenAI responded by asking the court to dismiss the lawsuit, claiming the accusations were "meritless" and stating it does not "have, nor want, any of their trade secrets" because it is "much more interested in building innovative products and technologies that push the frontier."
The outcome could influence how AI companies source training data and proprietary information going forward. For investors, the case adds legal risk to OpenAI's operations at a time when the company's advertising business just hit $1 billion in annualized revenue run rate, according to Reuters. Microsoft, OpenAI's largest backer, could face indirect exposure if the court grants Apple's injunction request.
Apple now wants to investigate the Mac mini and other devices Liu still has in his possession, arguing that evidence is being destroyed. Judge Davila will hear arguments on October 1. If Apple prevails, the case could reshape how AI companies handle proprietary data from former employers, with implications for the broader AI training sector and for companies like Google, Meta, and Anthropic that compete in the same talent pool.
This article is for informational purposes only and does not constitute investment advice.
