A Condensed Fact-Finding Report
Important note : Everything below attributed to Apple is an allegation in a civil complaint, not a proven fact. OpenAI disputes the claims.
The Facts
On July 10, 2026, Apple sued OpenAI, its hardware unit io Products, and two former Apple employees — Tang Yew Tan (now OpenAI's Chief Hardware Officer, a ~25-year Apple veteran who led iPhone/Watch product design) and Chang Liu (an eight-year Apple engineer who left in January 2026) — in federal court in San Francisco, under the Defend Trade Secrets Act.
Apple alleges Liu kept an Apple laptop after leaving and used a security bug to download confidential hardware files, and that Tan directed job candidates still at Apple to bring parts and prototypes to OpenAI interviews for "show and tell," and coached departing staff on evading exit security. Apple says it wrote to OpenAI in February 2026 with no response before filing suit, and later sent evidence-preservation letters to roughly 40 more former Apple employees now at OpenAI. It's seeking an injunction, damages, and attorneys' fees.
OpenAI says it has "no interest in other companies' trade secrets." Sam Altman responded on X that he's "not afraid of Apple" but has "tremendous respect" for it. Over 400 former Apple employees now work at OpenAI — itself unremarkable, since hiring from a competitor isn't illegal; Apple's case depends on showing specific confidential material moved with them, not just skill and experience.
The backdrop: OpenAI acquired Jony Ive's hardware startup io in 2025 for ~$6.4B, and is reportedly targeting consumer hardware and a public listing later this year — both of which this suit now complicates regardless of its eventual outcome.
Why It's Hard to Predict the Outcome
Trade secret law protects specific confidential information, not general know-how an employee is entitled to carry to a new job — and that line is historically hard to prove. Waymo v. Uber (2017) is the classic precedent: strong initial allegations, but the case settled for a fraction of what Waymo sought after the judge signaled skepticism it could show the technology actually reached Uber's product. More tellingly, a similar suit — xAI v. OpenAI — was dismissed with prejudice just last month for lack of proof. Courts in this space have consistently required hard evidence of actual use, not just opportunity or motive.
Conclusion
This case is best read less as a bellwether for how it will personally resolve, and more as an early, unusually detailed example of a dynamic that is going to keep recurring across the AI industry. Three things follow from that, in order of how confident they can be stated:
First, and most certain: the immediate business damage happens regardless of the verdict. Litigation like this chills hiring, clouds product timelines, and complicates IPO diligence the moment it's filed — Apple doesn't need to win to impose real cost on OpenAI's hardware push and public-listing plans. That's already happening.
Second, fairly confident: proof, not the law itself, is the real bottleneck, and history favors defendants more often than plaintiffs expect. Waymo settled for far less than sought; xAI's nearly identical theory against OpenAI just failed outright. Apple's complaint is unusually granular — specific messages, specific files, specific interview conduct — which may put it on firmer footing than those cases, but the core evidentiary burden (that secrets, not just talent, crossed over and were used) hasn't gotten any easier to meet, and won't for Apple either.
Third, more speculative but worth naming: the stakes of this pattern are rising as AI companies push into hardware and adjacent physical-product markets where they're recruiting directly from incumbents with deep institutional know-how. The legal question of what counts as "contamination" is also getting harder as confidential information moves not just into a rival's product but into an AI company's training and development pipelines, where courts and regulators (per Waymo and separately the FTC's Everalbum case) have shown willingness to treat integration itself, not just disclosure, as a harm that's difficult to reverse. Expect more suits shaped exactly like this one — mixed success at trial, but real and immediate leverage the moment they're filed — as the same small pool of hardware and research talent keeps moving between a handful of well-capitalized companies.
Sources: Reuters, CNBC, TechCrunch, Fortune, 9to5Mac, Forbes, NBC News, Bloomberg, Bloomberg Law, Computerworld, The Hill, and Lowenstein Sandler's client alert on Apple Inc. v. Chang Liu et al., No. 5:26-cv-07078 (N.D. Cal.