Apple’s lawsuit against OpenAI has quickly become one of the most closely watched trade-secrets cases in the tech industry—less because of the headline-grabbing premise (companies suing over alleged misappropriation is hardly new), and more because of the specific, unusually concrete behaviors Apple says occurred while its employees were interviewing for roles at OpenAI. In a complaint that reads like a blueprint of how competitive intelligence might be gathered in the hardware era of AI, Apple alleges OpenAI crossed boundaries involving confidential information, prototype-related activity, and even a kind of “bring your work” interview tactic that Apple characterizes as suspiciously tailored to product development.
The Verge’s reporting on the case highlights six of the most striking claims Apple makes in its complaint. While the lawsuit is still at the allegation stage—meaning none of these claims have been proven in court—the details Apple chose to emphasize offer a rare window into how companies believe trade secrets can be extracted not through hacking, but through relationships, recruiting pipelines, and carefully staged access.
Below is a deeper look at what Apple is alleging, why these claims matter beyond the courtroom, and what they suggest about the growing overlap between AI systems and the physical products—and supply chains—that increasingly define modern consumer technology.
1) The “real hardware” interview request: bringing components and unreleased samples
One of the most eyebrow-raising allegations is Apple’s claim that, during interviews, OpenAI’s hardware leadership asked candidates to show up with “real” materials—specifically components they were working on and unreleased product samples.
On its face, this could sound like an eccentric recruiting style: a hiring manager wants to see what you’ve built, how you think, and whether you can translate engineering intuition into tangible outcomes. But Apple frames it differently. In the complaint, the request is portrayed as unusual enough to be interpreted as a mechanism for extracting sensitive, work-in-progress information—information that would not normally be shared outside a company’s internal walls.
Why this matters is not just the content of what was allegedly requested, but the context. Hardware development is where trade secrets often live: in design choices, component selection, manufacturing constraints, test results, and the iterative decisions that never make it into public documentation. If Apple employees were asked to bring unreleased samples or components tied to ongoing projects, Apple argues that OpenAI wasn’t merely evaluating talent—it was potentially collecting proprietary artifacts and insights.
There’s also a subtle but important point here: Apple’s allegation implies that the “access” wasn’t limited to what candidates said in interviews. It was allegedly extended to what candidates physically brought. That shifts the risk profile. Verbal descriptions can be vague; physical samples can be analyzed, measured, reverse-engineered, and compared against other knowledge. Even if no one intends to copy a design, the mere act of seeing and handling unreleased hardware can create a trail of information that is difficult to unlearn.
In a world where AI companies are increasingly building their own devices, robotics systems, and edge hardware, Apple’s complaint suggests that OpenAI’s hardware ambitions may have made it more aggressive—or at least more opportunistic—in seeking direct exposure to product development realities.
2) Alleged theft of confidential Apple information
Apple’s second major thrust is that OpenAI misappropriated confidential documents it says it had access to through these relationships. This is the classic trade-secret allegation: that information obtained through employment transitions, recruiting networks, or partner interactions was used improperly.
What makes this claim significant is that Apple isn’t only saying “confidential information was taken.” It’s tying the alleged access to the relationships and circumstances described elsewhere in the complaint. In other words, Apple is attempting to connect the dots between who knew what, how they got it, and what OpenAI allegedly did with it afterward.
In many trade-secret disputes, the hardest part for plaintiffs is proving not just that someone had access, but that the defendant used that access in a way that caused harm. Apple’s strategy appears to be to establish a narrative of opportunity and intent: that the alleged interview behaviors and prototype-related conduct created a pathway for confidential information to move from Apple’s internal environment into OpenAI’s development orbit.
It’s also worth noting that “confidential documents” in a hardware-adjacent context can include far more than schematics. They can include internal specifications, testing methodologies, failure analyses, supplier communications, and performance targets. Even if a design is never copied verbatim, the underlying knowledge—what works, what doesn’t, and why—can be extremely valuable.
Apple’s complaint, as summarized by The Verge, positions this as more than ordinary competitive curiosity. It’s framed as misappropriation: the idea that OpenAI didn’t just learn from general experience, but allegedly took specific confidential material and used it in ways Apple believes violate legal protections.
3) Alleged spying or probing of Apple hardware prototypes
Another striking claim is Apple’s allegation that OpenAI engaged in conduct involving probing or observing hardware development work connected to Apple prototypes.
“Spying” is a loaded word, and lawsuits often use strong language to describe behavior that may later be debated in court. Still, the core allegation matters: Apple is asserting that OpenAI’s actions went beyond passive interest and into active observation or investigation of prototype-related work.
This is where the case becomes especially relevant to the broader tech industry. Prototype work is often treated as a fortress of secrecy. Companies don’t just protect finished products; they protect the process—what’s being tested, what’s failing, what’s being iterated, and what’s being considered for future releases. If a competitor can observe that process, it can compress its own learning cycle. It can also anticipate product directions, identify likely technical bottlenecks, and adjust strategy accordingly.
Apple’s complaint suggests that OpenAI’s alleged conduct may have been aimed at understanding Apple’s hardware trajectory rather than simply hiring people who once worked there. That distinction is crucial. Talent acquisition is legal; using trade secrets is not. Apple is essentially arguing that the line was crossed.
Even without proof of direct copying, the ability to observe prototype development can provide a competitive advantage. In hardware, timing is everything. Knowing what’s coming—and when—can influence partnerships, procurement decisions, and product roadmaps. Apple’s allegations imply that OpenAI sought that advantage through improper means.
4) Alleged involvement in proprietary product design work via a trusted partner
Perhaps the most consequential claim—because it involves third parties—is Apple’s allegation that OpenAI didn’t just request information, but allegedly pushed a trusted partner to carry out a proprietary product design technique.
This is a different category of allegation than “someone stole documents.” It suggests a more operational approach: using intermediaries to obtain or execute proprietary methods. If true, that would mean the alleged misuse wasn’t confined to internal discussions or document handling. It extended into action—work performed by someone else, potentially under the influence of OpenAI’s requests.
The phrase “proprietary product design technique” is important. It implies a method, not just a component. Techniques can be deeply embedded in a company’s engineering culture and process. They can involve specialized workflows, manufacturing steps, calibration procedures, or design heuristics that are hard to replicate without knowing the underlying rationale.
If a trusted partner was allegedly encouraged to perform such a technique, Apple’s argument would be that OpenAI effectively outsourced the extraction or implementation of trade-secret knowledge. That raises the stakes for both the defendant and the partner involved, because it suggests a deliberate attempt to obtain value while potentially reducing direct evidence of wrongdoing inside OpenAI itself.
This is also where the lawsuit’s “wildest claims” framing becomes understandable. The allegation isn’t merely that OpenAI wanted information; it’s that OpenAI allegedly tried to convert that information into real-world design execution through another entity. That’s the kind of behavior that plaintiffs typically reserve for cases they believe are more than speculative.
5) The case centers on specific people and specific roles
Apple’s complaint, according to The Verge’s summary, focuses heavily on the actions of three individuals, including Tang Tan, a 24-year Apple veteran who recently served as vice president of the Apple Watch. Apple’s narrative ties alleged events to particular roles and timelines, which is a common legal tactic: specificity helps establish credibility and makes it easier to argue causation.
When lawsuits are vague, defendants can dismiss them as rumor or conjecture. When lawsuits name individuals and describe what they allegedly did, plaintiffs are trying to show that the alleged misconduct wasn’t abstract. It was personal, procedural, and traceable.
The inclusion of Tang Tan is particularly notable because it signals that Apple is not only targeting OpenAI as an institution. It’s also pointing to the human movement of knowledge—how expertise travels when executives and engineers switch companies. In the tech industry, that movement is normal. People leave; they bring experience. But Apple’s complaint is essentially arguing that some of that experience crossed legal lines.
By centering the case on specific people, Apple is also attempting to show that the alleged conduct wasn’t accidental. It wasn’t just “someone talked too much.” It was allegedly structured around certain behaviors—like the interview request for hardware materials—and around interactions that Apple believes enabled misappropriation.
6) A rivalry framed around hardware and strategy—not just hiring
At a high level, Apple is framing this as more than typical talent-switching or competitive curiosity. The lawsuit positions the dispute as a hardware-and-strategy rivalry playing out in court.
That framing matters because it influences how the case will be perceived by judges, juries, and the public. If Apple can persuade the court that the alleged actions were part of a deliberate effort to gain strategic advantage—especially in hardware—then the case becomes more than a dispute over employment transitions. It becomes a story about how AI companies may seek to accelerate their physical product capabilities by leveraging access to a rival’s confidential development work.
This is also where the industry context comes in. AI is no longer purely software. The push toward on-device inference, robotics, sensors, and specialized compute hardware means that AI companies increasingly need hardware expertise and