OpenAI has escalated its legal confrontation with Apple by publicly rejecting allegations that it obtained or used confidential information from former Apple employees.
In a strongly worded response titled “Apple is getting this wrong,” OpenAI described the lawsuit as careless, aggressive and unusually personal. The company also published selected emails and private messages that it says contradict important parts of Apple’s account.
The dispute centres on OpenAI’s expansion into consumer hardware and the movement of experienced engineers between two of Silicon Valley’s most influential companies. Apple alleges that OpenAI, its hardware unit io Products and two former Apple employees improperly obtained trade secrets that could accelerate the development of competing devices. OpenAI denies possessing or wanting Apple’s confidential information.
No court has determined which company’s account is correct. The allegations remain contested, and the lawsuit could become an important test of how trade-secret law applies when technology companies recruit employees with years of highly specialised knowledge.
Apple Says Former Employees Took Valuable Hardware Information
Apple filed its lawsuit in July 2026 in the US District Court for the Northern District of California. The defendants include OpenAI, io Products and former Apple employees Chang Liu and Tang Yew Tan, commonly known as Tang Tan.
Tan previously held a senior hardware role at Apple and later became closely involved in OpenAI’s consumer-device operations. Liu is another former Apple engineer who joined OpenAI.
Apple alleges that the two men retained or accessed confidential material connected with hardware engineering, manufacturing systems, supplier relationships and unreleased products. The company argues that this information could provide OpenAI with an unfair advantage as it develops its own consumer hardware.
The complaint reportedly goes beyond accusing individual employees of mishandling files. Apple alleges that OpenAI followed a broader strategy of recruiting its workers and obtaining knowledge that Apple had spent years and substantial resources developing.
According to TechCrunch’s examination of Apple’s claims, Apple also accuses OpenAI of advising departing employees on how to avoid security procedures that might expose the removal of confidential information. OpenAI disputes that characterisation.
Apple Wants an Injunction Before the Case Is Decided
Apple is not waiting for the normal litigation process to reach a final judgment. It has asked the court for a preliminary injunction that would restrict OpenAI and the former employees from accessing, using or disclosing the information Apple identifies as its trade secrets.
The company has also requested expedited discovery. That process could require the defendants to produce documents and participate in depositions earlier than would normally occur in the case.
Apple argues that it could suffer irreparable damage if confidential information becomes incorporated into OpenAI’s products, development systems or supplier operations before the lawsuit reaches trial. It says that without immediate court intervention, it could eventually find itself competing against devices developed partly through its own protected technology.
A preliminary injunction would not amount to a final ruling that OpenAI stole trade secrets. It would instead be an early protective measure imposed while the court examines the evidence and the wider case continues.
The request could still have significant consequences. Restrictions affecting employees, documents or development processes could slow OpenAI’s hardware programme even if the company ultimately defeats Apple’s claims.
OpenAI Says It Neither Has Nor Wants Apple’s Secrets
OpenAI’s public response rejects the foundation of Apple’s injunction request. The company states that it does not possess Apple’s trade secrets and has no interest in using them.
It also says it offered to cooperate after learning about Apple’s concerns. According to OpenAI, Apple did not meaningfully raise the detailed allegations before filing the lawsuit, despite later suggesting that the companies had discussed the matter.
OpenAI’s response is unusually direct for an active corporate lawsuit. Rather than limiting its position to a short statement from legal counsel, the company published a detailed narrative accompanied by emails, text messages and contextual explanations.
The strategy appears designed to challenge Apple in both the courtroom and the court of public opinion. OpenAI is not merely arguing that Apple will struggle to prove its case. It is portraying the lawsuit as the product of misunderstanding, poor internal controls and an increasingly hostile effort to obstruct a new competitor. That interpretation remains OpenAI’s position rather than an established judicial finding.
A Mistaken Email Has Become Part of OpenAI’s Defence
One of OpenAI’s most striking arguments concerns communications involving Apple’s outside counsel and OpenAI general counsel Che Chang.
OpenAI published emails showing that Apple’s lawyer contacted Chang and referred to a previous conversation that Chang said had never happened. Apple’s representatives subsequently acknowledged that the message had been sent to the wrong person and that the mistake was being addressed.
OpenAI presents this exchange as evidence that Apple’s version of the pre-lawsuit discussions is unreliable. It argues that Apple did not clearly explain the specific accusations and then filed a sweeping complaint months later.
The email error does not automatically disprove Apple’s substantive trade-secret allegations. A mistaken recipient and disputed description of earlier communications are separate from the question of whether confidential files were removed or used.
However, the mistake could affect how the court evaluates the credibility and urgency of Apple’s story, particularly because Apple is asking for expedited discovery and immediate injunctive relief.
Messages From Chang Liu Challenge Apple’s Timeline
OpenAI also published messages involving Chang Liu that it says provide missing context about his departure from Apple.
Apple alleges that Liu improperly accessed confidential files after leaving the company. OpenAI responds that any continuing access existed because Apple had not properly removed it during the offboarding process. It further says Apple employees continued contacting Liu for assistance with work-related matters after his departure.
The published messages appear to show former colleagues asking Liu for help locating or understanding files. OpenAI argues that this undermines the suggestion that every instance of post-departure access was secretive or intended to benefit his new employer.
Apple may respond that accidental access does not grant permission to retain, transfer or use protected information. Trade-secret litigation usually depends not only on whether a person could open a file, but on what information was accessed, why it was accessed, where it went and whether another company benefited from it.
Those questions will require fuller evidence than the selected communications either side has made public.
The Lawsuit Threatens OpenAI’s Hardware Ambitions
The dispute matters because OpenAI is attempting to extend its influence beyond software and cloud-based artificial intelligence.
Its hardware operation includes former Apple design and engineering talent, including people connected with io Products. Apple argues that the recruitment pattern is relevant because more than 400 former Apple employees reportedly work at OpenAI.
Hiring experienced employees from a competitor is not itself unlawful. Workers generally retain their skills, professional judgment and general industry knowledge when they move to another company. They are not, however, permitted to take protected documents or disclose specific trade secrets covered by legal obligations.
The difficult question is where personal expertise ends and proprietary corporate information begins. That boundary can become especially complicated in hardware development, where manufacturing methods, component tolerances, supplier capabilities and unreleased design decisions may result from years of confidential work.
The outcome could influence how quickly OpenAI can build its planned hardware and how aggressively Apple protects employees, suppliers and technical information from emerging AI competitors. Reuters’ report on the injunction request notes that Apple views OpenAI’s consumer-device push as a direct competitive threat.
The Conflict Is More Complicated Because Apple and OpenAI Are Partners
The lawsuit is particularly notable because Apple and OpenAI are not conventional corporate enemies.
In 2024, the companies announced an agreement to integrate ChatGPT into Apple experiences, including Siri and Writing Tools across iOS, iPadOS and macOS. The original OpenAI partnership announcement presented the collaboration as a way to make ChatGPT available within Apple’s operating systems while maintaining user privacy protections.
That software relationship can continue even as the companies become competitors in hardware. Apple controls one of the world’s most important consumer-device ecosystems, while OpenAI increasingly wants to create physical products built specifically around artificial intelligence.
The lawsuit shows how quickly a strategic partner can become a threat when it moves closer to a company’s core business. Apple may value access to OpenAI’s models while simultaneously fearing that OpenAI could eventually produce devices that reduce consumers’ dependence on the iPhone.
Publishing Private Communications Is a High-Risk Strategy
OpenAI’s decision to release emails and private messages gives the public more context, but it also increases the intensity of the confrontation.
Apple can argue that selected communications do not represent the complete evidentiary record. OpenAI can argue that Apple’s court filings omit context that fundamentally changes the meaning of the alleged conduct.
The judge will ultimately evaluate authenticated documents, testimony, technical records and legal arguments rather than corporate blog posts. Public messaging can shape opinion, but it cannot replace discovery or determine whether information qualifies as a legally protected trade secret.
OpenAI’s response nevertheless makes one point clear: the company does not intend to treat Apple’s case as a routine employment dispute. It is framing the lawsuit as an effort by an established hardware leader to slow an ambitious new rival.
Apple, meanwhile, is presenting the conflict as a defence of legitimate intellectual property rather than resistance to competition.
The court will have to decide whether the evidence supports either narrative. Until then, OpenAI’s published “receipts” strengthen its public defence, but they do not settle the central question of what information left Apple and whether it was ever used.
