Apple v. OpenAI: Inside the Trade Secret Lawsuit That Could Redefine the AI Hardware Race

On Friday, July 10, 2026, Apple filed a federal lawsuit against OpenAI, two former Apple employees, and Jony Ive's hardware studio, io Products, accusing them of a "coordinated pattern" of trade secret theft aimed at helping OpenAI build its own consumer hardware. It is one of the most dramatic corporate legal showdowns of the year: two of the most valuable and closely watched companies on the planet, suddenly on opposite sides of a courtroom.

Here is what is actually in the complaint, and what to make of the bigger questions swirling around it.

What the Case Is About

Apple filed suit in the U.S. District Court for the Northern District of California, alleging trade secret misappropriation and breach of contract. The complaint names two individual defendants:

• Tang Yew Tan: A nearly 24-year Apple veteran who rose to vice president of product design for iPhone and Apple Watch before leaving for io Products in 2024. He is now OpenAI's Chief Hardware Officer
  
  
• Chang Liu: A former senior system electrical engineer at Apple who joined OpenAI's technical staff in January 2026

Apple's core allegations, per multiple outlets that reviewed the 41-page filing, include the following:

• Liu allegedly retained an Apple-issued laptop after departing and later exploited an authentication bug to access confidential company systems, downloading dozens of files covering technical specifications, engineering presentations, and proprietary manufacturing data.
  
  
• Tan allegedly used confidential Apple codenames during OpenAI recruiting interviews, asked candidates still employed at Apple pointed questions about unreleased products, and — most strikingly — directed some candidates to physically bring "actual parts," CAD files, and prototypes from Apple to their OpenAI job interviews for "show and tell" sessions.
  
  
• Apple claims OpenAI used information gleaned this way to approach Apple's own manufacturing partners, including allegedly requesting a proprietary metal-finishing technique from a supplier.
  
  
• Apple also alleges a third former employee, Alyssa Peng, helped access confidential files using her own Apple credentials before leaving to join OpenAI's hardware team

The language in the filing is unusually blunt for corporate litigation. Apple's complaint states OpenAI's hardware business is "rotten to its core" from its "illegal reliance on misappropriated trade secrets," and that misconduct occurred "at every level, from members of its Technical Staff to its Chief Hardware Officer".

Apple is seeking injunctive relief, monetary damages, and declaratory judgments. Notably, Apple included a footnote clarifying that its existing ChatGPT-Siri integration deal with OpenAI is "not at issue here"; this is strictly about hardware.

Will OpenAI Fire the Named Employees?

Neither OpenAI nor Tan nor Liu has announced any personnel action, and there is no public indication a termination is imminent. 

Legally, firing them now would be a mixed bag for OpenAI. It could look like an admission that something was wrong, but keeping them in place (especially Tan, who runs OpenAI's entire hardware division) keeps the company fully invested in defending their conduct. Companies in this position often wait for outside counsel and internal investigations to run before making any personnel moves, partly to avoid the appearance of prejudging facts, and partly because a termination does not make an employee's prior conduct go away. 

Apple's claims are about what already happened, not about who currently holds the job.

The February Warning: Why an Internal Probe Is Likely Already Underway

One detail buried in the coverage is significant: Apple says it reached out to OpenAI back in February 2026 to raise concerns and request an investigation, and that OpenAI never responded.

That is five months of silence before Apple escalated to a lawsuit. From a legal-strategy standpoint, this is a common, almost textbook pattern: a company puts a rival on notice, gives them the opportunity to self-investigate and remediate, and only sues once that channel goes nowhere. It also matters for Apple's legal position. Having given clear notice back in February, Apple can argue OpenAI had every opportunity to investigate and correct course, which strengthens claims that any continued use of the information was willful. Whether OpenAI quietly ran its own internal review after that February contact and simply did not tell Apple or the public is not something we can confirm from public reporting. However, a responsible legal or compliance team receiving that kind of direct accusation would typically be expected to look into it internally, whether or not they responded externally.

Following Apple’s filing, OpenAI’s Director of Strategic Communications, Drew Pusateri, took to X to comment on the company’s behalf: We have no interest in other companies' trade secrets. We remain focused on building innovative technology that empowers people everywhere.

Could OpenAI Just Ship the Product and Treat Any Penalty as a Cost of Doing Business?

This is the "efficient breach" question, and it is worth taking seriously, but hardware makes it a much riskier bet than it would be in a standard software contract dispute.

The classic efficient-breach logic (paying damages if they are cheaper than compliance) runs into a few hardware-specific obstacles here:

• Injunctions, not just damages, are on the table. Apple is not only asking for money; it is asking a court to block OpenAI's hardware efforts built on the disputed information. If a judge grants even a preliminary injunction, OpenAI could be barred from selling the product altogether, which defeats the "pay a fine and keep the profits" math entirely.

• Damages in trade secret cases can include disgorgement of all profits attributable to the stolen information rather than a fixed statutory penalty. That means the financial penalty is not a flat number that gets easily out-earned by sales; it can scale directly with how successful the product becomes.
  
  
• Willfulness can trigger enhanced damages. Under the Defend Trade Secrets Act (DTSA), if misappropriation is found to be willful or malicious, courts can award exemplary damages up to two times the compensatory award, on top of attorneys' fees. That is a very different calculus than a predictable, bounded penalty.
  
  
• Reputational and IPO timing risk. OpenAI is reportedly gearing up for a high-profile IPO. A protracted trade secret fight, let alone a jury verdict finding willful theft, is exactly the kind of overhang that spooks underwriters and public investors in a way a private software dispute would not.

So while "ship it and eat the fine" is a real strategy in some intellectual property contexts, it is a much shakier bet in trade secrets, precisely because the remedies are designed to strip away the benefit of the breach rather than cap it.

The Supply Chain Trap

This might be the most quietly dangerous part of the case for OpenAI. Apple's complaint alleges OpenAI did not just absorb information from ex-employees; it allegedly used that information to approach Apple's own manufacturing partners, including a request for a specific proprietary metal-finishing technique.

If true, this extends the exposure beyond merely hiring people who knew things. It suggests OpenAI's hardware supply chain—the vendors, tooling, and processes needed to actually manufacture a device at scale—may itself be built on information Apple considers proprietary. That is a much harder thing to unwind than reassigning an employee. You can put an engineer on leave, but you cannot as easily unpick a manufacturing relationship or a finishing process already qualified into a production line without real delay and cost.

Global Sales, U.S. Courts, and the DTSA's Extraterritorial Reach

Here is a legal wrinkle that matters a lot given how global consumer hardware sales are: the Defend Trade Secrets Act has been confirmed to apply well beyond U.S. borders.

In Motorola Solutions, Inc. v. Hytera Communications Corp., the Seventh Circuit ruled, and the Supreme Court declined to disturb, that the DTSA can reach all of a defendant's worldwide sales caused by the misappropriation, so long as an act in furtherance of the misappropriation was committed in the United States. In that case, Hytera was ordered to pay $407.4 million in damages calculated from worldwide sales of radios built using stolen Motorola trade secrets, even though most of the underlying conduct happened overseas.

The practical implication for Apple v. OpenAI: if a court finds that any part of the alleged misappropriation (recruiting conversations, file transfers, or supplier outreach) occurred on U.S. soil, Apple could potentially seek damages tied to every unit OpenAI sells anywhere in the world, not just U.S. sales. For a company hoping to sell an AI device globally, much like the iPhone, that turns a U.S. lawsuit into a massive global liability exposure.

How Big a Headache Is This, Really?

• For OpenAI: Potentially significant. Beyond the direct legal exposure, it lands right as the company is reportedly preparing for an IPO, adds to an already crowded legal docket, and threatens the timeline and market positioning of its first hardware product. Supply-chain analysts have suggested this could include a phone-like device as early as 2028, alongside earlier reports of an AI-powered speaker ecosystem.

• For Apple: Lower direct financial risk, but not zero cost. Litigation is a distraction, discovery cuts both ways (Apple's own internal practices and hiring processes could get scrutinized), and there is a narrative risk. Apple suing a company for hiring away its talent can look, to some observers, like sour grapes about competition rather than genuine IP theft, even if the specific facts alleged (physically bringing parts to interviews) go well beyond ordinary poaching.

Jony Ive's Sidelining: Does the Optics Angle Matter Here?

This is genuinely interesting because of what the complaint does not say. Jony Ive, Evans Hankey, and Scott Cannon, the co-founders of io Products alongside Tan, are not named as individual defendants anywhere in the filing. io Products itself is named as a corporate defendant, and Apple describes it as complicit in "a coordinated pattern of misconduct at an institutional level," but Ive personally is not accused of anything in the complaint's text.

That is a deliberate choice, and it cuts a few ways. On one hand, it may reflect a genuine evidentiary reality; Apple may simply not have specific facts tying Ive personally to the alleged conduct, as opposed to Tan and Liu, where the complaint cites specific actions. On the other hand, leaving out arguably the most famous name attached to OpenAI's hardware effort is also strategically useful for Apple. It keeps the story focused on theft, not on "our former design chief left to compete with us," which is a much less sympathetic frame for Apple to push in public.

For Ive personally, the optics are still complicated even without being named. He built his post-Apple career on the idea of a clean break: a fresh, unencumbered reinvention of consumer hardware for the AI era. A lawsuit alleging the products under his design leadership rest on stolen Apple blueprints undercuts that "created from scratch" narrative regardless of whether his name appears on the complaint.

OpenAI acquired io Products for roughly $6.5 billion in 2025 — a price that reflected Ive's design pedigree and the promise of an original, ground-up hardware vision, not an assessment of Apple's internal IP. If a court finds that vision was built in part on misappropriated Apple trade secrets, it doesn't just complicate Ive's personal narrative — it retroactively calls into question what OpenAI actually bought at that price, and whether shareholders or future investors were sold a "from scratch" story that wasn't accurate.

Bigger Than Apple v. Samsung?

Almost certainly not, at least not in the way Apple v. Samsung became a massive cultural touchstone. The Samsung fight played out over years, went through multiple jury trials and a Supreme Court appearance, and became shorthand for an entire era of smartphone patent wars, with billions of dollars and years of appeals attached.

Apple v. OpenAI is fundamentally different in kind:

• It is a trade secret case, not a design or utility patent case. The legal theories, evidence types (internal emails, recruiting notes, download logs), and remedies are quite different from standard patent infringement analysis.

• It is unlikely to produce the same kind of viral, easily explainable cultural artifacts that the Samsung case did. Jurors debating whether one phone looks too similar to another is inherently more visual and public than analyzing whether an engineer emailed himself internal supplier documents.

• It could still be legally consequential, potentially more so given the DTSA's extraterritorial reach and the stakes for a major forthcoming product line, but it is less likely to become a mainstream pop-culture reference point.

Where it could rival or exceed Samsung is in raw dollar exposure, if OpenAI's hardware product turns out to be a hit and a court applies worldwide-sales damages under the DTSA framework described above.

Echoes of Google v. Uber (Waymo v. Uber)

The parallels to the 2017–2018 Waymo v. Uber case are hard to miss, and they were likely on Apple's mind when drafting this complaint:

• The pattern: A senior engineer (Anthony Levandowski) with deep access to unreleased proprietary technology leaves for a competitor, allegedly taking confidential material along the way.

• The scale of alleged document exfiltration: Waymo alleged its departing engineer downloaded over 14,000 files before leaving. Apple's complaint similarly centers on allegations of massive confidential file downloads and retained hardware devices.

• The corporate acquisition angle: Just as Uber acquired the engineer's startup and effectively absorbed the disputed IP along with it, OpenAI acquired Ive's io Products in a multi-billion dollar deal, which Apple's suit frames as having absorbed a pattern of trade-secret-tainted practices along with the acquisition.

• The key difference: Waymo v. Uber was about a genuinely singular piece of hardware (a LiDAR sensor design) at the center of an industry-defining technology race. Apple's complaint against OpenAI alleges a broader, more diffuse pattern of recruiting practices, supplier relationships, and cultural norms around confidential information spread across over 400 former Apple employees now at OpenAI. That breadth could make Apple's case harder to prove with the same laser focus Waymo had, but if proven, it paints a picture of something more systemic than one rogue engineer

This is a fast-developing story, and details will likely be updated as OpenAI files its formal response and the case proceeds through discovery.