Apple is escalating its legal clash with OpenAI, arguing that potentially crucial digital evidence could disappear before the court has a chance to examine it. At the center of the dispute is a former Apple employee, an Apple-owned MacBook, and allegations that confidential hardware information may have followed him to OpenAI.

Apple has asked the court for expedited discovery, a faster-than-normal evidence-gathering process, because it says the risk is not theoretical. In a recent filing, Apple alleged that OpenAI delayed handing over a MacBook used by former employee Chang Liu, even though OpenAI had reportedly possessed the laptop since the lawsuit began in July. Apple says it did not receive the device until August 21.

That timing matters because modern corporate investigations often turn on details that do not sit neatly in a single folder. System logs, metadata, access records, cloud-sync information, deleted-file traces, and device usage history can all help establish what happened, when it happened, and who had access. Apple argues that these forensic artifacts can be transient – overwritten, lost, or deliberately erased – making a delayed inspection especially significant.

According to Apple’s account, an inspection of the laptop found that Liu had downloaded a confidential Apple circuit schematic and later used it while working at OpenAI. Apple also claims that Liu and others at OpenAI knew he retained access to an Apple third-party cloud-storage system after he left the company. Those are serious allegations, though they remain allegations at this stage rather than findings by a court.

The company says the MacBook also contained messages that raised fresh concerns about preservation. Apple alleges that Liu discussed the need to “restore” and begin using Apple-owned devices after learning of an internal investigation in June. Apple’s argument is essentially that waiting for the usual pace of litigation could leave it trying to reconstruct a digital trail after parts of that trail have already gone cold.

A broader fight over AI hardware

This evidence battle is part of Apple’s larger lawsuit against OpenAI, which accuses the ChatGPT maker of misusing Apple trade secrets in connection with an AI device effort. The dispute reportedly involves three former Apple employees who joined OpenAI, with Liu becoming a central figure because of the company-owned laptop and Apple’s claims surrounding confidential technical material.

The case lands at an awkward moment for the broader AI industry. OpenAI has become one of the most aggressive companies pursuing consumer-facing AI hardware, including its high-profile collaboration with former Apple design chief Jony Ive. Apple, meanwhile, has immense experience protecting tightly integrated hardware, software, component, and design work – much of which depends on information that is valuable precisely because it is not public.

Trade secret cases are often less about one dramatic document than about patterns of conduct. A company may try to show that an employee had access to sensitive materials, retained them after leaving, and then performed related work at a rival. The other side can challenge each part of that chain: whether the information was actually secret, whether it was accessed improperly, whether it was used at all, and whether the former employee had a legitimate reason to possess it.

That is why the fight over preservation is so important. Before either side can fully argue what the evidence means, they have to establish that the evidence still exists in a usable form.

OpenAI pushes back

OpenAI disputes Apple’s characterization of the situation. In its own filing, OpenAI argued that the case is a “mess of Apple’s own making,” contending that Apple’s employee exit process was disorganized and that Apple is attempting to shift responsibility for its own shortcomings.

OpenAI also maintains there is no evidence that Liu accessed Apple information through improper means or for an improper purpose. The company’s position is that Liu had no use for Apple’s confidential information, directly rejecting the suggestion that such material was needed for his work at OpenAI.

That response points to a central tension in the case. Apple sees the delayed laptop inspection and the alleged messages as warning signs that evidence could be compromised. OpenAI sees Apple’s claims as an effort to turn a messy departure process into a much bigger accusation against a rival company. The court will eventually have to sort through not only the technical evidence but also the competing stories about how the devices, files, and access permissions were handled.

For now, the most consequential question may be relatively narrow: should Apple be allowed to move quickly to preserve and inspect evidence before digital records change or vanish? If the judge agrees, the litigation could accelerate sharply, potentially putting more internal communications, device records, and former employee activity under scrutiny.

The stakes extend beyond Apple and OpenAI. As AI companies race to build new devices and recruit talent from established hardware firms, disputes over employee departures, confidential design work, and digital evidence are likely to become more common. In an industry where a schematic, prototype plan, or internal hardware roadmap can be worth far more than the laptop storing it, the integrity of the evidence may become almost as important as the alleged secrets themselves.