Apple vs. OpenAI: New Evidence and an Allegation of a Cover-Up

In the legal fight between Apple and OpenAI over allegedly stolen hardware secrets, Apple has filed new court documents that the company itself describes as shocking. They rest on a forensic analysis of a work computer that a former Apple engineer’s lawyers handed over in August. Apple accuses the man and other former staff not only of stealing confidential circuit schematics but also of trying to erase the traces. OpenAI rejects the allegations as meritless and has again asked the court to dismiss the case.

Key takeaways

• In late August, Apple filed new forensic evidence said to have been found on the old Apple MacBook of former employee Chang Liu.
• According to the complaint, Liu downloaded confidential circuit schematics and had an AI agent run the circuit simulations on its own.
• Apple accuses Liu of deleting forensic data in June, together with another defendant, after he learned about the internal investigation.
• The company has expanded the suit to eleven former employees and is seeking expedited discovery plus a preliminary injunction against OpenAI’s hardware development.
• OpenAI calls the allegations meritless and blames Apple’s trouble retaining staff. A verdict is not expected before 2027.

What Apple accuses the former engineer of

At the center of the new filing is Chang Liu, described by Apple as a former senior electrical systems engineer who moved to OpenAI in early 2026. Before leaving, he is said to have loaded dozens of confidential files onto his Apple laptop, including a schematic marked highly confidential for future Apple hardware. He allegedly imported that schematic into the widely used simulation program LTspice and then instructed an AI agent to run the circuit simulations autonomously. Apple argues that this step is especially damaging: once trade secrets flow into an AI system, the consequences are hard to undo, because the knowledge spreads in an uncontrolled way through machine learning. The company speaks of irreversible and continually propagating uses.

Apple also alleges that Liu obtained unauthorized access. According to the complaint, he exploited a flaw in the network sign-in to keep reaching company cloud storage after his departure. In chat messages that Apple has cited in earlier briefs, he allegedly joked about it and marked the fact with laughing emoji. OpenAI, Apple says, was aware of this continued access.

The evidence-destruction allegation

The most serious new claim concerns June 2026. After Liu allegedly learned that Apple was investigating him, he is said to have discussed with Yu-Ting Peng how to delete precisely the forensic data Apple would need to reconstruct events. He also allegedly told a colleague to reset Apple devices to factory state. By Apple’s account, the documents now on file only became available because Liu’s lawyers handed over his old MacBook for examination in August. In discovery, Apple is also demanding access to a Mac mini used at OpenAI.

In parallel, Apple has broadened the suit: instead of two defendants, it now names eleven former employees. Among those identified is Tang Yew Tan, who previously oversaw the design of the iPhone and Apple Watch. He is said to have used job interviews to press Apple staff still on the payroll for project code names, CAD data, prototypes, and supplier information, and to have advised candidates to stay at Apple as long as possible before disclosing their move. Apple calls this recruiting as espionage.

What is really at stake: the race for AI hardware

The dispute is part of a bigger picture. In 2025, OpenAI acquired io, the hardware startup of former Apple design chief Jony Ive, and is working on its own AI device meant to give users access to ChatGPT away from the smartphone. Against that backdrop, Apple reads the departures as a pattern: about 400 former Apple staff now work at OpenAI, according to the complaint. The original July suit, which we covered when it was filed, already accused OpenAI of deliberately harvesting knowledge and helping departing employees bypass security processes.

OpenAI pushes back: the alleged trade secrets remain vague, it says, and Apple is substituting lawyerly rhetoric for missing facts. Residual access, the company argues, was the result of Apple’s own poorly managed offboarding. Liu, it says, only accessed files after his move at the request of former colleagues. In an earlier statement, OpenAI called the accusations careless, aggressive, and oddly personal, and tied them to Apple’s difficulty keeping talent and using AI effectively.

Assessment: a test case for the industry

Whatever the outcome, the case highlights two shifts that reach beyond Apple and OpenAI. First, AI changes the impact of a leak. A schematic that flows into model training or an agent pipeline cannot be pulled back as easily as a copied folder, and courts will have to work out how to measure that kind of harm at all. Second, it puts the AI industry’s talent war on display: when hundreds of engineers travel the same path from one company to the next, the line between an ordinary job change and systematic knowledge transfer blurs. Similar conflicts already exist between other providers.

For OpenAI, the near-term stakes go beyond damages. If the court grants the requested preliminary injunction, it could slow the company’s hardware plans for the duration of the case. Both sides have already filed repeated motions and counter-motions, and a verdict is not expected before 2027. Until then, the case is likely to deliver one thing above all: a public look at how carelessly, or how deliberately, the AI industry handles other people’s intellectual property.