On October 13, 2026, a federal judge granted Apple expedited discovery in its trade secret lawsuit against OpenAI — ordering forensic images of engineers' devices, LINE chat messages, and job interview materials, after finding a valid risk that critical evidence could be destroyed. The order is two days old. The lessons inside it are permanent.
The Apple v. OpenAI trade secret case just moved at a speed most litigation never reaches. On October 13, 2026 — two days ago — U.S. District Judge Edward Davila partially granted Apple's motion for expedited discovery, ordering OpenAI and the named defendant engineers to produce forensic device images, LINE chat messages, and materials used to interview Apple job candidates within five days of the order. He found a valid risk that critical forensic evidence on the devices of engineers OpenAI had recruited could be destroyed.
The specific evidence that drove the judge's decision is striking. Apple's forensic expert concluded that defendant Chang Liu used an Apple trade-secret file in March 2026 and told another OpenAI colleague in June 2026 to have Apple-issued devices restored and to start using them more. That instruction — to restore devices and increase their use — is precisely the kind of conduct that courts treat as evidence of anticipatory evidence destruction. A device that has been restored has had its forensic history altered. A restored device from which a court is then trying to recover deleted files is dramatically harder to examine. The judge read that instruction the same way any experienced eDiscovery professional would: as a reason to move fast.
OpenAI has also filed its own counter-allegation, claiming a former employee who left for Apple presented details of an unreleased OpenAI product to senior Apple leaders, making this a two-sided trade secret battle that is accelerating rapidly. The court has not yet ruled on Apple's preliminary injunction motion. What it has done is give Apple forensic access to the most probative evidence in the case — and given OpenAI five days to produce it.
Granted — forensic device images. The court ordered production of forensic images of the individual defendants' devices, finding that the instruction to restore Apple-issued devices created a valid risk that critical forensic evidence could be destroyed. This is the most consequential part of the order — forensic images capture the full state of a device at the moment of imaging, including deleted files, metadata, access histories, and artifacts that ordinary document production would never surface.
Granted — LINE chat messages and interview materials. The court ordered production of private LINE messages between the defendant engineers and materials OpenAI used to interview Apple job candidates. LINE is a messaging platform popular in East Asia; its messages are not captured by standard enterprise email or Slack archiving. Their production requires device-level collection — exactly what the forensic imaging order enables.
Denied — OpenAI's broader R&D records. The judge declined to accelerate discovery into OpenAI's hardware research on metal finishes, power management, and battery functionality, ruling those requests are appropriate on the ordinary timeline. The court drew a clear line: urgency justified by evidence destruction risk gets expedited treatment; curiosity about a competitor's research does not.
The instruction that Liu allegedly gave his colleague — to restore Apple-issued devices and start using them more — is precisely the kind of message that transforms a trade secret case from a standard document review into an emergency forensic investigation. Device restoration, in the context of a departing employee who is also a named defendant in an anticipated lawsuit, is the digital equivalent of shredding documents. It overwrites the forensic record. It destroys file access histories. It eliminates the metadata that would show which files were opened, when, and by whom. And it does all of this in a way that, if done before a litigation hold is issued, can support an adverse inference instruction to the jury.
Judge Davila's order acknowledged the competing expert positions — Apple's expert found that Liu used the trade-secret file in March and gave the restoration instruction in June; OpenAI's expert found no evidence the file was used in connection with any AI platform — but found the risk of destruction sufficient to justify acting now. The judge warned that the alleged trade secrets could become so deeply entangled with OpenAI's hardware products that the harm cannot be undone. That warning is not just about this case. It is about what happens in every trade secret matter when evidence review is delayed long enough that the information has been integrated into the defendant's products and operations.
The LINE chat dimension adds a layer that many organizations miss entirely. LINE is a messaging platform whose logs do not appear in standard enterprise email archives, do not sync to corporate IT systems, and cannot be collected through the same tools used for Gmail or Outlook. Collecting LINE messages requires either device-level forensic access or a production request directly to LINE's legal team — a process that is not instantaneous and that can be complicated by the platform's terms of service and international data protection frameworks. The court's order requiring LINE chats within five days was a recognition that those messages, if they existed, were at risk.
Liu allegedly told a colleague to restore Apple-issued devices and use them more. That six-word instruction is what convinced a federal judge that waiting for ordinary discovery timelines was too dangerous. Five days. Forensic images. LINE chats. Done.
Expedited discovery — the process by which a court accelerates the normal timeline for evidence production — is one of the most powerful tools available in litigation. It is also one of the most misunderstood. Organizations often treat a motion for expedited discovery as an aggressive litigation tactic by opposing counsel. In practice, courts grant it on a specific and narrow basis: when there is a demonstrated risk that evidence will be destroyed, altered, or made unavailable if the normal timeline is followed.
The Apple case demonstrates the three elements courts look for when deciding whether to grant expedited discovery. First, a concrete and specific risk of evidence destruction — not a vague concern, but an identifiable action (the restoration instruction) that creates an identifiable threat to an identifiable category of evidence (the forensic history of specific devices). Second, a defined and limited scope — Apple got forensic images and LINE chats, not OpenAI's entire hardware research program. Third, urgency that cannot be addressed through the ordinary timeline — because device restoration, unlike most forms of document destruction, is a process that happens continuously as devices are used and cannot be reversed once it occurs.
For organizations that anticipate litigation — or that are already in it — the Apple order is a useful calibration of what expedited discovery actually requires and what it actually produces. The five-day forensic imaging order is not a sanction. It is a recognition that the evidence at risk is perishable and that the ordinary discovery timeline would allow it to disappear before anyone could preserve it. Every organization that has former employees with access to sensitive information on unreturned devices is, in some version, living in the pre-litigation phase of this story.
The Apple v. OpenAI October 13 order is the fastest-moving major eDiscovery development of 2026 — a case that went from filing in July to expedited forensic imaging in October, driven by a six-word instruction to restore devices that a judge read as a five-alarm preservation emergency. At Sovereign Discovery, we treat device return, access revocation, and exit interview completion as the three cornerstones of trade secret risk management at employee departure. The Apple case demonstrates what it costs when any of those three steps is skipped or manipulated.
We also note the LINE chat dimension specifically. Non-standard messaging platforms — LINE, WeChat, Telegram, WhatsApp, Signal — are among the most under-collected categories of ESI in litigation today. They exist. They contain relevant communications. They require device-level forensic access to collect properly. And they are increasingly the communications channel that matters most, precisely because they feel less formal and less monitored than corporate email. That feeling does not change their discoverability. The Apple order makes that unmistakably clear.
Two days ago, a federal judge ordered OpenAI to hand over forensic images of its engineers' devices and their LINE chat messages within five days, after finding that an instruction to restore company-issued devices created a valid risk that critical evidence could be destroyed. The order is not a finding that OpenAI stole Apple's trade secrets. It is a finding that the evidence needed to answer that question was at risk of disappearing before anyone could preserve it.
Every organization that has former employees with company-issued devices, unreturned hardware, or access to messaging platforms outside corporate archiving is living in some version of the pre-litigation phase of this story. The Apple v. OpenAI case is the most visible current demonstration of what it costs when those loose ends are not tied before the litigation clock starts running.