Apple and OpenAI were, not long ago, partners. Apple integrated OpenAI’s technology into its devices starting in 2024, a deal that looked like a comfortable arrangement between a hardware giant and an AI powerhouse. Then OpenAI started building its own hardware. Now Apple is suing.

On July 10, 2026, Apple filed a 41-page federal lawsuit against OpenAI in the Northern District of California, alleging the company ran a coordinated campaign to steal trade secrets tied to its consumer hardware ambitions. OpenAI responded by saying it takes the lawsuit seriously but believes the allegations lack merit, and that it stands behind employee mobility and open competition.

What Apple is actually alleging

The complaint is not a vague accusation. Apple’s lawyers put specifics on the table.

The lawsuit names Tang Tan, OpenAI’s Chief Hardware Officer and a former Apple executive, as a central figure in the alleged scheme. Apple claims OpenAI solicited confidential information from both current and former Apple employees, including internal documents and physical prototypes.

Here is where it gets unusual: Apple alleges that job candidates were asked to bring Apple device parts to interviews. The complaint also references informal internal communications in which individuals expressed amusement about unauthorized access to Apple’s internal network storage.

Some media observers have described Apple’s response as a “thermonuclear” legal strike, borrowing the phrase Steve Jobs once used to describe his feelings about Android.

How the partnership unraveled

The partnership that began in 2024 was, at the time, framed as a natural complement. Apple had the distribution and the devices; OpenAI had the model. The relationship cooled as OpenAI’s hardware ambitions became clearer.

Tang Tan’s role sits at the center of this tension. He went from a senior position at Apple to leading OpenAI’s hardware division, and Apple is now alleging his transition was not clean. The lawsuit argues his hiring was part of a broader pattern, not an isolated executive move.

Elon Musk, who has his own complicated relationship with both OpenAI and Apple, weighed in publicly on the dispute.

What this means for OpenAI’s hardware play

Litigation of this scale, especially one that targets the head of your hardware division by name, creates real operational risk. Tang Tan’s ability to lead product development while simultaneously being a named defendant in a high-profile federal case is a question OpenAI will need to manage carefully.

There is also the discovery process to consider. Federal lawsuits come with broad discovery rights. If Apple can compel OpenAI to produce internal communications, hiring records, and product development documents, the reputational exposure could extend well beyond whatever the court ultimately decides on the merits.

OpenAI’s statement leaned on two principles: it doubts the claims have legal merit, and it supports workers’ right to choose their employers. Trade secret cases often hinge on the line between an employee’s general knowledge and skills, which travel freely with the person, and specific proprietary information, which does not. California is generally more protective of employee mobility than most states, which may give OpenAI some structural advantage in the litigation. But “we hired people legally” is a harder argument to sustain if the evidence shows those people were actively solicited to bring physical hardware components to job interviews.