Apple Accuses Former Engineer of Using Proprietary Hardware Schematics to Train OpenAI Agents.

Apple Escalates Trade Secret Lawsuit Against OpenAI Over Stolen Hardware Schematics

Apple has dramatically escalated its high-stakes intellectual property litigation against artificial intelligence rival OpenAI. According to new federal court filings submitted on Monday, Apple alleges that a former senior electrical systems engineer illegally accessed proprietary circuit schematics for power converters shortly after leaving the company to join OpenAI.

The Cupertino tech giant claims the decisive evidence was recovered from a MacBook that OpenAI surrendered on August 21 as part of the ongoing legal discovery process. According to the filings, the former employee, Chang Liu, accessed Apple’s highly classified hardware schematics and utilized the proprietary data to train autonomous AI agents in March 2026.

In response to the escalating legal threat, Apple is formally petitioning the federal court to grant an expedited discovery process to uncover the full extent of the data transfer. OpenAI previously attempted to have the lawsuit dismissed entirely, arguing to the court that its engineering efforts are focused on building entirely new, unrelated products that do not rely on Apple’s intellectual property. OpenAI has not immediately responded to requests for comment regarding this latest evidentiary filing.

A federal judge will review Apple’s request for expedited discovery ahead of a scheduled procedural hearing on September 15.

How is artificial intelligence (AI) changing the nature of trade secret theft? Historically, business espionage involved stealing designs to produce counterfeit goods. In this case, Apple alleges that the data was used to train AI agents. If an AI model accepts proprietary engineering data to learn how to design circuits independently, it creates a huge legal gray area regarding how courts can order a company to “forget” or remove specific weights from a trained neural network.

While power converters might sound commonplace compared to neural processors, current power efficiency is the single biggest bottleneck in AI hardware design. Whether OpenAI is developing proprietary server infrastructure for its data centers or quietly working on consumer-grade hardware, acquiring a power management architecture on par with Apple would save years of fundamental research and development time.

Apple and OpenAI have a complex “friend-and-enemy” relationship. While Apple recently integrated ChatGPT into its operating system, the two companies have been fiercely poaching each other’s top-tier engineers to build their own AI hardware and software ecosystems. Such high-profile lawsuits serve two purposes: protecting intellectual property and acting as a strong deterrent for current employees considering moving to a competitor’s AI lab.

 

💬 AI Content Assistant

Ask me anything about this article. No data is stored for your question.

Apple Escalates Trade Secret Lawsuit Against OpenAI Over Stolen Hardware Schematics

Apple has dramatically escalated its high-stakes intellectual property litigation against artificial intelligence rival OpenAI. According to new federal court filings submitted on Monday, Apple alleges that a former senior electrical systems engineer illegally accessed proprietary circuit schematics for power converters shortly after leaving the company to join OpenAI.

The Cupertino tech giant claims the decisive evidence was recovered from a MacBook that OpenAI surrendered on August 21 as part of the ongoing legal discovery process. According to the filings, the former employee, Chang Liu, accessed Apple’s highly classified hardware schematics and utilized the proprietary data to train autonomous AI agents in March 2026.

In response to the escalating legal threat, Apple is formally petitioning the federal court to grant an expedited discovery process to uncover the full extent of the data transfer. OpenAI previously attempted to have the lawsuit dismissed entirely, arguing to the court that its engineering efforts are focused on building entirely new, unrelated products that do not rely on Apple’s intellectual property. OpenAI has not immediately responded to requests for comment regarding this latest evidentiary filing.

A federal judge will review Apple’s request for expedited discovery ahead of a scheduled procedural hearing on September 15.

How is artificial intelligence (AI) changing the nature of trade secret theft? Historically, business espionage involved stealing designs to produce counterfeit goods. In this case, Apple alleges that the data was used to train AI agents. If an AI model accepts proprietary engineering data to learn how to design circuits independently, it creates a huge legal gray area regarding how courts can order a company to “forget” or remove specific weights from a trained neural network.

While power converters might sound commonplace compared to neural processors, current power efficiency is the single biggest bottleneck in AI hardware design. Whether OpenAI is developing proprietary server infrastructure for its data centers or quietly working on consumer-grade hardware, acquiring a power management architecture on par with Apple would save years of fundamental research and development time.

Apple and OpenAI have a complex “friend-and-enemy” relationship. While Apple recently integrated ChatGPT into its operating system, the two companies have been fiercely poaching each other’s top-tier engineers to build their own AI hardware and software ecosystems. Such high-profile lawsuits serve two purposes: protecting intellectual property and acting as a strong deterrent for current employees considering moving to a competitor’s AI lab.