OpenAI has asked a court to throw out Apple's lawsuit over stolen trade secrets, arguing that Apple's own careless handling of sensitive information undermines its case. The filing, reported on August 6, 2026, is OpenAI's most detailed response since Apple sued the company in July 2026 (TechCrunch).

Apple's lawsuit says OpenAI hired former Apple engineers who brought confidential hardware information with them, as part of a planned effort to grab Apple's secrets (Reuters). The suit names two former Apple employees, including Chang Liu, a former senior system electrical engineer. In early August, Apple asked the court to speed up the evidence-gathering process, saying its own investigation found that more former employees may have seen or taken part in the alleged theft (TechCrunch).

Trade secrets are a specific type of intellectual property: information a company keeps confidential because it gives them a competitive edge. Think of a restaurant's recipe or a company's manufacturing process. But for information to count as a legally protected trade secret, the company has to show it took reasonable steps to keep it private. It is like having a lock on a diary — if you leave the diary open on a park bench, it is hard to argue you were serious about keeping it secret.

OpenAI says Apple failed that test. The company contends that Apple let employees use personal iCloud accounts for work purposes and did not properly cut off system access when employees left. In one piece of evidence, OpenAI submitted text messages showing that an Apple manager stayed logged into Chang Liu's personal iCloud account after he left the company, used it to transfer files, and later asked Liu for help with technical questions about Apple projects (TechCrunch).

OpenAI also said Apple left out of its lawsuit what OpenAI called Apple's "inexplicable information-management practices," including the failure to secure systems when employees departed. The motion further argues that Apple never specified exactly which secrets were allegedly stolen, using only broad categories like component manufacturing, product testing, vendor and supplier relationships, and distribution channels (TechCrunch).

Beyond the legal arguments, OpenAI has publicly framed the lawsuit as a competitive move. The company said Apple is using trade secrets litigation to slow a competitor from building AI-powered hardware, calling the suit "baseless and pretextual" (TechCrunch). OpenAI also accused Apple of filing the case to stop employees from leaving for OpenAI (Financial Times). On August 4, OpenAI published a webpage titled "Apple is getting this wrong," correcting what it called factual errors and sharing messages about the events in question (OpenAI). OpenAI separately denied the theft allegations and called the lawsuit "careless" and "oddly personal" (Wall Street Journal).

The dispute comes down to a core rule in trade secrets law. A company suing for trade secret theft must show not only that someone took the information improperly, but also that the company treated it as genuinely secret. OpenAI's motion targets that second part. The iCloud evidence, if accurate, is especially damaging to Apple: a manager logging into a departed employee's personal account to grab work files, then asking that former employee for help on current Apple projects, makes it hard to argue Apple kept tight control over its information.

The argument about Apple's vague descriptions reinforces the point. Trade secrets law generally requires a plaintiff to identify the specific secrets at issue in enough detail that the court can evaluate the claim and the accused party can defend itself. Apple's use of broad categories, as OpenAI describes them, leaves the court unable to judge whether the alleged secrets meet the legal standard.

It is worth noting that OpenAI's competitive-harm argument adds another layer. By framing the suit as an attempt to block a rival's AI hardware plans, OpenAI is telling both the court and the public that Apple's reasons go beyond protecting intellectual property. Whether that argument carries legal weight is a separate question from its public-relations value, since courts generally focus on whether theft occurred, not on why a company chose to sue. But the framing does help explain what is at stake for both companies, particularly if OpenAI is working on its own hardware and would benefit from hiring people with deep systems-engineering experience.

Apple's request to speed up evidence gathering signals the company thinks its internal investigation has only uncovered part of the story. If more former employees are found to have been involved, the case could grow significantly. Apple will need to directly answer the security-practice accusations, likely by producing evidence of access controls, employee exit procedures, and rules about personal account use that push back against the picture OpenAI has painted.

The case is still in its early stages, and a motion to dismiss tests whether the lawsuit is legally sound rather than whether the underlying facts are true. But the arguments OpenAI has put together, especially the iCloud evidence, put Apple in the position of having to explain why its own handling of sensitive information is consistent with a company that truly treats its hardware secrets as secret.