When Apple filed its lawsuit against OpenAI in July 2026, most observers wrote it off as corporate saber-rattling — the kind of legal posturing that tech giants deploy to slow down fast-moving competitors. Six weeks later, the case has turned into something far more serious, with allegations of circuit schematic theft, deliberate evidence destruction, and a troubling picture of how talent wars between AI companies are exposing deep vulnerabilities in Big Tech’s internal security.
The Engineer at the Center of It All
The case revolves around Chang Liu, a former Apple hardware engineer who left the company in January 2026 to join OpenAI as part of what has reportedly become a wave of roughly 400 Apple employees migrating to the AI startup. What started as a competitive talent concern for Apple became a full-blown legal crisis when internal investigators uncovered a critical forensic finding.
Approximately six weeks after Liu’s departure — on March 7, 2026 — Apple’s security team says Liu used his still-provisioned Apple-issued MacBook to access and download a highly confidential circuit schematic containing power-converter designs and simulation data. The schematic, according to Apple’s court filing submitted to the U.S. District Court for the Northern District of California, isn’t just proprietary in a general sense. It represents years of hardware engineering that Apple says is central to its next-generation device architecture.
What makes the filing especially striking is the follow-on allegation. When Liu apparently learned that Apple had launched an internal investigation in June 2026, he allegedly instructed an OpenAI colleague named Yu-Ting Peng to “restore” Apple-issued devices — a request that Apple’s legal team characterizes as a deliberate attempt to wipe forensic evidence. Peng, according to the filing, confirmed compliance with that request.
A Company Accused of Evidence Destruction
Apple’s brief accuses OpenAI not just of benefiting from stolen intellectual property but of actively aiding the destruction of evidence related to an ongoing legal proceeding. The evidence-destruction claim carries significant legal weight that goes beyond the original trade secret allegations. If proven, it would constitute obstruction that could dramatically shift the legal calculus in Apple’s favor, independent of whether the underlying schematic was actually used to develop OpenAI products.
OpenAI has pushed back forcefully. In a filing of its own, the company called the dispute “a mess of Apple’s own making,” arguing that Apple’s offboarding procedures were so flawed that it cannot credibly claim its former employees had access to materials that were properly controlled or restricted. OpenAI also argued that any files Liu accessed were irrelevant to the hardware work the company is actually pursuing.
That argument is notable for what it reveals: OpenAI did not specifically deny that Liu accessed the files, nor that Peng wiped the devices. It contested, instead, whether those files were material to OpenAI’s work — a narrower and more legalistic defense.
The Bigger Story: AI’s Talent War Has Consequences
Lost in the legal back-and-forth is a broader issue that this case has forced into the open. The AI industry is in the midst of one of the most intense talent migrations in technology history. The allure of frontier AI development — the salaries, the equity, the sense of being at the cutting edge of a transformational technology — has drawn engineers, researchers, and product managers away from established tech giants at an unprecedented pace.
Apple, historically one of the most secretive and operationally security-conscious companies in the world, has found itself unusually exposed. The company’s tight integration between hardware and software means that its engineers carry extraordinary amounts of proprietary knowledge — not just software architecture, but chip designs, power management circuits, manufacturing processes, and supply chain relationships that represent decades of refinement.
When 400 of those engineers walk out the door toward a competitor actively building consumer hardware, the risk isn’t hypothetical. It’s structural.
Apple’s lawsuit against OpenAI is partly about Chang Liu and one circuit schematic. But it’s also about the realization that standard offboarding procedures — the NDAs, the device returns, the exit interviews — were never designed for a world where the company hiring your best engineers is building the same kind of products you make, at enormous speed, with enormous capital.
What Happens Next
The immediate legal battleground is a preliminary injunction hearing scheduled for October 1, 2026, before Judge Edward Davila. Apple is seeking to halt OpenAI’s hardware development work while the case proceeds — a dramatic ask that, if granted, would significantly disrupt a major strategic initiative for the AI company.
Legal experts have noted that injunctions in trade secret cases are difficult to obtain precisely because courts require a clear showing that irreparable harm is ongoing and that the balance of equities favors the plaintiff. Apple will need to demonstrate not just that Liu had access to confidential materials, but that those materials are actively being used in OpenAI’s hardware pipeline.
OpenAI has argued for dismissal, and its lawyers will presumably make the case at the October 1 hearing that the schematic in question — a power-converter design — isn’t relevant to what OpenAI is building. That argument could succeed even if the underlying facts of access and device wiping are not seriously contested.
What’s unlikely to be resolved anytime soon is the broader question the case raises: in an era of unprecedented talent mobility and breakneck AI development, what meaningful protections do companies actually have for their most sensitive engineering work?
Apple’s Quiet Hardware Ambitions Suddenly Very Public
One underreported dimension of this case is what it reveals about OpenAI’s direction. Apple’s original July lawsuit alleged that OpenAI was misappropriating trade secrets related to “hardware design, manufacturing and supply-chain operations” — language that suggests OpenAI’s consumer device aspirations are more advanced and more specific than the company has publicly acknowledged.
OpenAI has long been rumored to be developing AI-native hardware to reduce its dependence on third-party devices and platforms. The company recruited Jony Ive’s design studio and has reportedly been working on a new category of AI hardware. Apple’s lawsuit suggests that some of the engineering talent fueling those efforts came directly — and potentially improperly — from the company that has defined consumer hardware for two decades.
The October 1 hearing will set the tone for what could become one of the most consequential intellectual property battles in the history of the technology industry. Regardless of the legal outcome, the case has already done something important: it has made viscerally clear that the age of friendly competition between AI labs and established tech giants is over.
Sources: Bloomberg, Gizmodo, 9to5Mac, Axios, AndroidHeadlines
