Apple is taking its trade secrets fight with OpenAI from background noise to courtroom focus, and the alleged mechanics are unusually specific. Filed Friday, Apple’s lawsuit accuses OpenAI of “a coordinated pattern of misconduct at an institutional level,” centered on two claims: a security failure that allowed continued access to Apple systems after a former employee left, and recruitment practices that Apple says functioned like intelligence-gathering for confidential hardware.

The complaint also names Apple’s former hardware executive, Tang Tan, alleging he played a role in interview “show and tell” sessions. Apple frames this as more than talent poaching. It argues OpenAI’s nascent hardware business is “rotten to its core,” built on misappropriated trade secrets, and says it has asked OpenAI earlier this year to investigate whether Apple’s confidential information entered OpenAI’s business without response.

To understand why this matters, zoom out for a second. Apple and OpenAI were one of Silicon Valley’s higher-profile AI partnerships, and the relationship deepened after Apple integrated ChatGPT into Apple Intelligence in 2024. But the source describes how the partnership frayed as OpenAI ramped up consumer hardware ambitions, in part by hiring former Apple executives and engineers. When hardware and AI collide, corporate incentives are straightforward: faster product iteration, tighter supply chain execution, and differentiating on technical execution. Those are also the exact areas that trade secret lawsuits try to police when the argument is “you didn’t just learn from the market, you took what you weren’t supposed to take.”

Apple’s first major allegation is about a former employee, Chang Liu, who allegedly left for OpenAI in January. Apple says Liu failed to return a company laptop and that it later discovered Liu could still access Apple’s internal systems due to what the complaint describes as an authentication bug. The key claim is that, rather than reporting the issue, Liu downloaded dozens of confidential engineering files while working at OpenAI. Apple lists categories of what was downloaded: documents related to unreleased products, technical specifications, presentations, and manufacturing processes.

Apple goes further than “access happened.” It alleges that Liu encouraged another Apple engineer who was interviewing at OpenAI to study confidential Apple materials ahead of her interviews. Apple’s complaint also claims Liu advised that engineer on how to avoid attracting attention from Apple’s security team while copying files, and that the two moved conversations to a private messaging app to avoid detection. This is the kind of detail trade secret cases lean on, because it attempts to show intent and coordination, not just accidental exposure.

The second major allegation is about recruitment. The lawsuit claims OpenAI used its job recruitment process to obtain trade secrets, including hardware-adjacent information. The complaint points to OpenAI’s status as a widely sought-after employer, noting the increase in hardware engineering talent leaving Apple. Apple then alleges that Tang Tan, described as Chief Hardware Officer and a former Apple vice president with 24 years at Apple, asked candidates to bring physical components for “ show and tell ” sessions during interviews. The lawsuit also alleges OpenAI employees asked candidates for prototypes and information about vendors.

Apple’s narrative includes a specific example. It alleges that one then-Apple employee screenshotted and downloaded files from a “highly confidential Apple project,” and that Tan later asked about that project during the interview. Apple uses that to make its larger argument about causality: it says OpenAI’s recruiting practices suggest it hires these individuals at least in part because of confidential Apple-specific knowledge and expertise they have and could improperly obtain.

Apple also claims it raised concerns earlier this year and asked OpenAI to investigate whether Apple confidential information had entered OpenAI’s business. According to the complaint, OpenAI never responded. Apple then proceeded to file suit on Friday, seeking damages in an amount to be determined at trial, and court orders preventing OpenAI and the individual defendants from possessing or using Apple’s trade secrets.

OpenAI’s response, as captured by the Business Insider report, is a direct denial of motive. An OpenAI spokesperson told Business Insider the company has “no interest in other companies’ trade secrets,” and that it remains focused on building innovative technology that empowers people everywhere.

There’s also a broader compliance and legal backdrop here. The report notes that the ChatGPT creator is involved in a separate, high-profile legal battle with The New York Times over claims that it allowed its chatbot to use copyrighted material to train its systems. And it reminds readers that Elon Musk previously sued OpenAI, alleging CEO Sam Altman and president Greg Brockman “stole” the original OpenAI nonprofit and unjustly enriched themselves by changing it to a for-profit model. A jury handed OpenAI a legal victory in May, concluding that Musk missed the three-year statute of limitations to sue.

Finally, for executives at AI and consumer hardware companies, this is the strategic takeaway that keeps you up at night. Lawsuits like this often look like internal disputes, but they quickly become board-level risk management questions: How do you structure employee offboarding? What are your controls around device return, authentication, and post-departure access? How do your recruiting and technical evaluation processes avoid crossing the line from “learning” into “acquiring”? And when partnerships are fraying and hardware ambitions accelerate, does your legal posture keep pace with the incentives your teams face? Apple’s complaint is already trying to connect talent movement, technical interviews, and alleged access into one chain. Whether a court agrees with Apple’s claims, the direction is clear: the world is treating AI-era competition as a trade secrets problem, not just a product problem.