OpenAI and its co-defendants have filed a motion urging a federal court in California to strike two expert declarations and disregard newly submitted evidence in Apple’s high-stakes hardware trade secrets lawsuit. The filing, submitted on September 25, 2026, in the U.S. District Court for the Northern District of California, accuses Apple of improperly attempting to shore up a deficient preliminary injunction request by sandbagging the defense with voluminous new materials at the reply stage.
At the center of the dispute is Apple’s bid to stop OpenAI, former Apple design chief Jony Ive's hardware venture io Products, and two former senior Apple hardware engineers from allegedly leveraging Cupertino’s proprietary intellectual property. The aggressive pushback from OpenAI marks the latest escalation in a legal war that exposes intense Silicon Valley competition as the generative AI pioneer prepares its initial push into consumer hardware.
The Motion: OpenAI Strikes Back Against Apple's Reply Brief
In the new procedural maneuver, OpenAI, io Products, and former Apple engineers Chang Liu and Tang Tan asked U.S. District Judge Edward Davila to strike significant portions of Apple’s latest submission. The dispute concerns a reply brief Apple filed to support its request for a preliminary injunction, which aims to bar OpenAI from accessing, using, disclosing, or soliciting confidential Apple technology.
Under federal court procedure, a plaintiff must typically present its substantive arguments and core supporting evidence in its opening motion, using reply briefs only to rebut specific counterarguments raised by the defense. OpenAI contends that Apple violated this principle by withholding foundational arguments and introducing entirely new technical analyses only after reading the defense’s response.
By unloading fresh material at the final stage of briefing, the defense argues, Apple created an unfair tactical imbalance that denied the defendants an adequate opportunity to cross-examine or formally answer the assertions before the court. The defendants are asking the court to strike two specific expert declarations—along with five accompanying exhibits—and disregard additional expert opinions that Apple allegedly failed to present earlier.
What Apple Submitted: 126 Pages Across Five Expert Declarations
Photo: lawcommentary.com (source)
Apple's disputed submission was far from routine. To defend its injunction motion, the company filed a reply accompanied by 126 pages of testimony spread across five expert declarations, paired with 13 evidentiary exhibits.
The filings were designed to establish two main points: first, that Apple maintains rigorous internal security controls; and second, that its former engineers systematically transferred actionable trade secrets related to hardware components and manufacturing methods.
Apple's lineup of expert witnesses targeted key pillars of its intellectual property claims:
- Daniel Roffman: A forensic expert who analyzed digital artifacts, cloud access logs, and internal communications recovered from hardware used by the defendants.
- James Pooley: A trade secret and corporate security specialist evaluating Apple's internal protection policies and employee confidentiality safeguards.
- Dr. Ayman Fayed: An engineering authority on power-management systems who assessed the alleged misuse of confidential circuit schematics.
- Dr. Paul Hatch: An industry expert offering technical evaluations supporting the novelty and competitive value of Apple’s disputed designs.
- Dr. Jude Runge: A metallurgy and surface-finishing specialist testifying to the proprietary nature of Apple’s confidential metal-finishing processes.
Breakdown of Expert Filings and Defense Objections
The defendants scrutinized each of Apple’s expert submissions, arguing that the iPhone maker systematically exceeded the permissible scope of a reply filing. The table below outlines each expert's focus and OpenAI's formal objection.
| Expert Witness | Area of Focus in Lawsuit | OpenAI's Core Procedural Objection |
|---|---|---|
| Dr. Jude Runge | Proprietary metal-finishing and anodization processes | Entirely new expert witness introduced on reply to improperly reinforce previously deficient testimony. |
| Dr. Ayman Fayed | Power-converter circuits and power-management secrets | Introduced novel technical opinions that could and should have been included in Apple’s opening motion. |
| James Pooley | Corporate security standards and trade-secret policies | Relied on an internal investigation conducted only after Apple filed its original preliminary injunction motion. |
| Daniel Roffman | Forensic examination of devices and cloud storage activity | Introduced data available to Apple before filing; selectively presented Liu's messages out of context. |
| Dr. Paul Hatch | Hardware design architecture and component novelty | Addressed arguments that go outside the proper boundary of responsive reply briefs. |
Should Judge Davila decide not to strike Daniel Roffman’s declaration and exhibits, OpenAI and Chang Liu have requested that the court accept an alternative declaration from Liu himself. In this counter-declaration, Liu provides his own technical context regarding the device operations, simulation files, and internal communications highlighted by Apple.
The Core Accusations: Hardware Schematics, AI Agents, and Lost Laptops
Photo: neoteo.com (source)
The procedural fight is anchored in Apple’s original complaint, Apple Inc. v. Chang Liu, Tang Yew Tan, OpenAI Foundation, OpenAI Group PBC, and io Products, LLC, filed on July 10, 2026. The lawsuit accuses OpenAI of systematically raiding Apple's hardware engineering talent to fast-track its own unannounced device ecosystem.
Apple focused heavily on Chang Liu, a senior system electrical engineer who left Apple after 8 years to join OpenAI in January 2026. Apple claims that upon departing, Liu retained an active, company-issued MacBook for months. Forensic data allegedly revealed that on March 7, Liu downloaded a confidential power-converter circuit schematic. Days later, on March 18, he allegedly ran simulations using LTspice—an electronic circuit design tool—on a Mac Mini, synchronizing data to the MacBook via iCloud.
Apple also pointed to chat records indicating Liu explored building an AI agent capable of operating LTspice autonomously to parse circuit outputs, purportedly compressing a full day of engineering optimization into moments. Furthermore, Apple accused OpenAI Chief Hardware Officer Tang Tan—a 24-year Apple veteran—of soliciting trade secrets by querying interviewing Apple staff using proprietary internal codenames.
'A Mess of Apple's Own Making': OpenAI's Defense Strategy
OpenAI and its legal representatives have mounted an aggressive defense, calling Apple’s lawsuit "a mess of Apple’s own making". OpenAI maintains that it has neither requested nor retained Apple trade secrets, arguing that Apple's claims are an attempt to stifle employee mobility and intimidate departing staff.
In previous counter-filings, OpenAI detailed what it described as sloppy offboarding and infrastructure practices inside Apple. The defense noted that Apple routinely permitted engineers to connect personal iCloud accounts to corporate devices while failing to cleanly terminate lingering cloud access upon resignation.
Regarding Liu's post-employment document access, the defense asserted that former Apple teammates actively reached out to Liu after his exit, asking for his help locating legacy files and troubleshooting unresolved engineering projects. Additionally, OpenAI argued that Apple has consistently failed to pinpoint distinct, protectable trade secrets, relying instead on overbroad generalities that cover standard electrical engineering methodologies.
The High Stakes of the OpenAI and Jony Ive Hardware Push
Photo: neoteo.com (source)
This legal clash is inextricably linked to OpenAI’s secretive push into consumer devices. In 2025, OpenAI acquired io Products, the hardware startup founded by legendary former Apple design chief Jony Ive. The acquisition set OpenAI on a direct collision course with Cupertino.
To build standalone consumer AI products, OpenAI has recruited aggressively across Silicon Valley, reportedly hiring more than 400 former Apple employees over recent years. While the exact nature of OpenAI's first device remains closely guarded—court filings suggest only that it is not an in-ear audio product—developing specialized consumer electronics requires deep institutional expertise in power efficiency, thermal envelope management, custom battery architecture, and compact enclosure materials.
Because these technical disciplines represent core Apple strengths honed across decades of iPhone and Mac iterations, Apple views OpenAI’s hardware division as an existential IP risk. An injunction barring OpenAI from utilizing disputed power-management techniques or metal-finishing processes could significantly stall OpenAI’s product roadmap.
Industry Impact: What This Means for Tech Workers and AI Hardware
The procedural battle in San Jose holds wide implications for engineers, tech employers, and consumers:
- Employee Mobility and Offboarding: If Apple secures its injunction based on lingering iCloud synchronization and unreturned hardware, enterprise companies across tech will likely implement far stricter offboarding restrictions, forensic auditing, and non-compete scrutiny.
- Protection of AI Training Workflows: The scrutiny surrounding Liu’s alleged AI agent—using proprietary schematics to train machine learning models on circuit design—sets an important legal precedent for how courts treat trade secrets incorporated into automated development tools.
- Timeline for OpenAI Hardware: Should Judge Davila grant Apple’s preliminary injunction, OpenAI’s hardware rollout could face substantial engineering delays or redesign requirements. For consumers eager to see what an OpenAI and Jony Ive collaboration produces, market availability could be pushed back significantly.
What Happens Next: The Road to the October 1 Hearing
The immediate question before Judge Davila is whether to grant OpenAI’s motion to strike, disregard the disputed declarations, or admit Chang Liu's responsive declaration into the formal record.
All eyes now turn to the scheduled October 1, 2026 court hearing. During this proceeding, Judge Davila will hear oral arguments concerning both Apple’s preliminary injunction bid and its request for expedited discovery into OpenAI’s confidential internal hardware blueprints. As the court sorts through hundreds of pages of competing technical testimony, the ruling will determine whether Apple can halt OpenAI’s hardware momentum or if OpenAI successfully forces Apple’s trade secret claims to trial under standard discovery rules.
FAQ
What did OpenAI ask the court to do regarding Apple's evidence? OpenAI and its co-defendants filed a motion asking the court to strike two expert declarations and disregard new evidence that Apple introduced in its reply brief, arguing the submissions violate procedural rules by denying defendants a fair opportunity to respond.
Who are the individual defendants in Apple's lawsuit? The lawsuit names former Apple engineers Chang Liu, a senior system electrical engineer, and Tang Tan, Apple's former vice president of hardware design who now serves as OpenAI's chief hardware officer.
What trade secrets is OpenAI accused of taking? Apple alleges the defendants misappropriated proprietary circuit schematics for power-converter systems, internal simulation parameters, confidential metal-finishing processes, and trade secrets related to battery and display architectures.
Why is Jony Ive's startup involved in the lawsuit? OpenAI acquired io Products, the hardware company co-founded by Jony Ive, in 2025 to develop consumer AI hardware; Apple named io Products as a defendant after multiple Apple designers and engineers migrated to the venture.
Has the court ruled whether OpenAI or its employees stole trade secrets? No, the court has not issued any final finding of liability or trade secret theft against OpenAI, io Products, or the individual engineers; the case remains in preliminary procedural and discovery stages.
