Apple’s Lawsuit Alleges OpenAI Stole Trade Secrets
Apple’s trade secret lawsuit against OpenAI, as reported by TechCrunch, paints a picture of coordinated, almost casual corporate espionage. The complaint alleges that OpenAI systematically recruited Apple employees and instructed them to steal confidential information — from exploiting an authentication bug to access network storage, to asking job candidates to bring actual hardware parts and CAD artifacts to interviews. The tension is not just about individual rogue behavior; Apple argues this misconduct is “normalized and exemplified by leadership,” suggesting a culture of theft at the top. The sheer scale is striking: over 400 former Apple employees now work at OpenAI.
The complaint details specific operational tactics. One message reads, “LOL, I found out I can access the [network storage], so funny,” sent by a former Apple engineer who accessed systems after leaving. Apple alleges OpenAI coached departing employees on how to evade the “dreaded walkout” — Apple’s security procedure that immediately terminates access — and told them not to sign exit documents. OpenAI also allegedly used Apple’s internal terminology when questioning suppliers about power and battery components, implying deep knowledge of proprietary design. A key acquisition, io (founded by Jony Ive), is accused of misleading an Apple partner into using a confidential metal-finishing technique.
For a technical builder, this case is a stark reminder that security is as much about people and process as it is about code. The allegations that OpenAI’s nascent hardware business rests on “shaky foundations, rotten to its core” by misappropriated secrets should make any company hiring from a direct competitor rethink its onboarding and compliance procedures. The lawsuit also underscores that insider threat detection and robust exit processes are non-negotiable when dealing with proprietary hardware IP. Whether or not all allegations hold up in court, the operational details here are a blueprint for what not to do.
