Apple escalates legal fight over alleged trade-secret theft tied to OpenAI

Apple has expanded its lawsuit against OpenAI with new allegations that a former employee used confidential company material in work for the AI lab and may have helped destroy evidence after learning he was under investigation. The filing raises the stakes in an already tense dispute over whether Apple trade secrets were used to advance OpenAI’s hardware ambitions.

The dispute centers on former Apple engineer Chang Liu, who now works at OpenAI, and comes as Apple seeks a court order that would limit OpenAI’s hardware efforts while the case moves forward. The new claims matter because they suggest the company sees this as more than a personnel dispute: Apple is trying to persuade a judge that sensitive technical information may have crossed company lines and that other former employees could be involved.

What Apple says happened

Apple’s latest court filing alleges that Liu brought confidential Apple material into his work at OpenAI, including a circuit schematic tied to Apple engineering and a tool with a name similar to an internal Apple application. Apple also says OpenAI knew Liu had access to Apple information and did not take adequate steps to prevent misuse.

The company further claims Liu asked a colleague at OpenAI, Yu-Ting Peng, to delete evidence in June after he became aware that Apple was investigating him. Apple describes those claims as part of a broader pattern showing that the company is not conducting a speculative search but pursuing evidence it believes is directly relevant to the misuse of trade secrets.

Apple’s filing argues that the MacBook it received from Liu’s legal team provides only a narrow look at the available evidence, but that even this limited material supports its claim that trade secrets were used and that evidence may have been destroyed.

Why the laptop matters

The laptop in question became important only after Liu’s legal counsel turned it over for review earlier this month. Apple says the device yielded “shocking evidence,” though the specific material is not visible publicly because the filing was heavily redacted.

That lack of public detail is common in trade-secret litigation, where parties often ask courts to seal the most sensitive technical information. In this case, Apple appears to be using the sealed material to strengthen its case for urgent discovery and a temporary restriction on OpenAI’s hardware work.

How OpenAI has responded so far

OpenAI has defended Liu by saying he only accessed Apple files after leaving the company, and only to help former co-workers who had asked him for assistance. The company has also sought to shift the focus away from intentional misconduct and toward what it describes as a broader systems problem at Apple.

In a recent blog post, OpenAI argued that Apple’s explanation of “residual access” was incomplete and claimed the issue stemmed from Apple’s failure to manage access rights properly when employees depart. Apple now disputes that account, saying Liu retained access because he took advantage of a rare authentication flaw that had not previously been known.

OpenAI has said Apple is trying to blame Liu for a “residual access” problem and has suggested that Apple’s internal offboarding controls are to blame for the continued access to files.

TechCrunch has asked OpenAI for comment on the newest allegations. As of the filing, OpenAI had not publicly addressed Apple’s latest claims in court.

Why is Apple asking for an injunction?

Apple is asking for a preliminary injunction because it wants the court to halt any OpenAI hardware work that allegedly relies on Apple technology while the lawsuit proceeds. A preliminary injunction is an emergency-style remedy designed to preserve the status quo when a company believes it may suffer irreparable harm if the work continues.

Apple is also seeking expedited discovery, which would speed up evidence gathering and allow lawyers to move faster through records, depositions and device examinations. The company says that broader inquiry is necessary because it believes additional former Apple employees may be implicated, not just Liu.

What expedited discovery would change

If granted, expedited discovery could force a faster look into what information was accessed, who had it and whether any OpenAI projects were influenced by Apple material. In practical terms, that could mean a quicker and more intrusive examination of records than would normally happen early in a case.

For Apple, speed matters. The company wants to prevent what it sees as a continuing risk that proprietary hardware ideas or engineering know-how could shape products still under development. For OpenAI, the same request could widen the dispute and draw more former Apple staff into the litigation.

Who is Chang Liu, and why is he central to the case?

Chang Liu is a former Apple employee now working at OpenAI, and Apple says his conduct sits at the center of the dispute. According to the company, he had continued access to Apple resources even after leaving and used that access in ways that benefited OpenAI.

Apple previously pointed to text messages from Liu as evidence that he knew he still had access to Apple files. In those messages, which Apple says were preserved in prior filings, Liu reportedly used laughing emojis while discussing the issue, suggesting to Apple that he did not treat the access problem as accidental or trivial.

OpenAI has taken a different view, saying that Liu’s post-employment access was incidental and arose while he was assisting former colleagues. The legal fight now appears to be about whether the access was harmless and administrative, or deliberate and tied to confidential technical work.

What kind of evidence is Apple relying on?

Apple’s court papers refer to a confidential circuit schematic and a software tool name that allegedly mirrors an internal Apple engineering application. Those references are important because they suggest the company is not merely concerned about broad ideas or general expertise, but about specific technical artifacts and potentially copied workflows.

Trade-secret cases often hinge on exactly this kind of detail. Courts typically look at whether the information at issue is truly secret, whether the defendant had access to it, and whether the plaintiff can show that the information was used or disclosed without permission.

Apple has not made the underlying evidence public, but the company says it has enough to justify court intervention. Its filings suggest that the dispute has moved beyond a narrow employment issue and into questions about whether former employees carried material from one of the world’s most tightly controlled hardware companies into a rival AI organization.

How big is the broader Apple-to-OpenAI employee pipeline?

Apple says the matter may be larger than one former engineer. In its initial filing, the company said more than 400 former Apple employees now work at OpenAI, underscoring the scale of talent flow between two of the most powerful brands in technology.

That number does not itself prove wrongdoing, but it helps explain why Apple is pressing for broader discovery. When one or two alleged incidents arise inside a large hiring pipeline, companies often worry that they may reveal a wider pattern involving knowledge transfer, design overlap or even intentional misuse of sensitive material.

Why this fight matters beyond one engineer

The lawsuit is important because it sits at the intersection of AI competition, hardware development and trade-secret law. OpenAI is increasingly associated with products that reach beyond software, and Apple appears worried that its hardware expertise could be part of that expansion.

If Apple can show that sensitive engineering knowledge was carried across employer lines, the case could influence how aggressively other tech companies police employee departures. It may also shape how AI firms document the provenance of design work, especially when hiring engineers who previously worked on tightly guarded consumer devices.

The case also highlights the unusual friction created by the modern AI talent market. Companies compete for the same engineers, often from the same elite hardware and software teams, while trying to guard confidential information in a field where the boundary between general experience and protected know-how can be difficult to draw.

What courts usually look for in trade-secret disputes

Courts generally ask three core questions: whether the material was actually secret, whether the defendant had lawful access to it, and whether the plaintiff can show misuse or likely misuse. Apple’s latest filing is designed to strengthen all three parts of that test.

• Was the information confidential and technically valuable?
• Did the employee still have access after leaving?
• Is there evidence the information was used in later work?
• Was any record destroyed or concealed after the investigation began?

Timeline of the dispute

The conflict has unfolded over several months and intensified as Apple says more evidence has come to light.

What happens next?

The next phase will likely focus on whether the court grants Apple’s request for an injunction and whether discovery is accelerated. If the judge allows those steps, Apple could gain access to more records and potentially build a stronger case around how the information moved from one company to another.

OpenAI, meanwhile, may try to keep the case narrow and challenge Apple’s reading of the evidence. The company’s response will likely hinge on whether it can persuade the court that any access was inadvertent, limited in scope and not tied to product development.

For now, the lawsuit remains one of the clearest signs yet that the race to build AI hardware and AI-powered devices is colliding with the tech industry’s strictest secrecy rules. Apple is making a pointed argument: that the evidence is not only enough to justify its lawsuit, but serious enough to demand immediate court action.

Key facts at a glance

Here is a concise overview of the main points in the case.

As the case develops, the court will determine whether Apple’s newest material is enough to justify fast-tracked discovery and restrictions on OpenAI’s hardware work. Until then, the legal battle is likely to remain a closely watched test of how far companies can go to protect trade secrets when top talent moves between rivals.