OpenAI, along with former Apple employees Chang Liu and Tang Tan and io Products, filed a new motion, asking the court to dismiss Apple‘s trade secret misappropriation complaint with prejudice, a ruling that would permanently prevent Apple from bringing the same claims again.
According to 9to5Mac, OpenAI’s latest filing argues that Apple has failed to meet the legal threshold for a trade secret claim under California law. Under the Defend Trade Secrets Act (DTSA), a plaintiff must connect three distinct propositions: what the particular trade secret is, why it qualifies for legal protection, and how the defendant improperly acquired, disclosed, or used it. OpenAI contends that Apple’s complaint and its opposition to OpenAI’s earlier motion to dismiss never complete that chain.
OpenAI claims Apple’s allegations remain fatally vague, stating in its brief that “Apple’s supposed ‘trade secrets’ are and remain vague categories and catch-alls, and Apple cannot substitute attorney characterization for missing facts.” The company argues that Apple failed to identify protectable trade secrets with enough specificity, failed to plausibly allege that any defendant misappropriated them, and failed to show resulting injury or ongoing harm.
Apple’s counterargument: Liu downloaded confidential files, Tan used codenames
Apple filed its response on August 19, urging the judge to reject OpenAI’s motion and accusing OpenAI’s defense of relying on “distortion, speculation, and improper extrinsic evidence.” Apple argues that OpenAI included material, text messages, a website listing Apple codenames, and a copy of an interview guide, that should not be considered at the motion-to-dismiss stage, which is based solely on the complaint.
Apple’s filing alleges that Chang Liu exploited a “rare authentication bug” to access Apple’s network storage weeks after joining OpenAI in January 2026 and downloaded dozens of confidential engineering files, including a presentation on how to manufacture and test main logic boards. Tang Tan, who left Apple in February 2024 to work with Jony Ive, allegedly used internal Apple project codenames to question interview candidates about unreleased products and asked one Apple employee to bring components (batteries, system-on-chips, logic boards, and shields) to OpenAI for a “show and tell” session.
OpenAI’s rebuttal: Apple’s security failures, not coordinated theft
OpenAI’s defense shifts focus to Apple’s own security practices as the root cause. The company argues that Apple’s security protocols, allowing staff to use personal iCloud accounts for work and later reviewing personal messages left on company devices, gave former employees leftover system access, not evidence of a coordinated scheme to steal hardware trade secrets.
This framing represents a strategic pivot by OpenAI: rather than disputing that information left Apple, the company is arguing that Apple’s controls were inadequate and that any access was opportunistic rather than orchestrated.
October 1 hearing will determine whether the lawsuit survives
With OpenAI’s motion to dismiss, Apple’s opposition, and OpenAI’s reply all now filed, briefing on the motion is effectively complete ahead of oral arguments scheduled for October 1, 2026. Judge Edward J. Davila will hear the arguments that day.
A ruling in OpenAI’s favor would not only dismiss the current lawsuit but prevent Apple from filing the same claims again. A ruling in Apple’s favor would allow the case to proceed toward discovery and potential trial, where Apple would need to present evidence of the specific trade secrets and how they were taken.
The lawsuit sits against a backdrop of significant brain drain from Apple to OpenAI. Over 400 former Apple employees now work at OpenAI, and the acquisition of Jony Ive’s io Products brought more than 50 additional engineers and developers into the fold as part of a $6.5 billion deal. Apple and OpenAI had spent roughly two years partnering on ChatGPT integration into Siri before the relationship soured as OpenAI moved into hardware under Ive’s leadership.
Despite the legal battle, OpenAI has stated it still plans to unveil its first hardware product in 2026 and release it in 2027. The outcome of the October 1 hearing could significantly shape OpenAI’s hardware development trajectory if the lawsuit proceeds beyond the motion phase.
