Two years ago, Apple put ChatGPT inside the iPhone. Now the company is suing OpenAI in federal court in Northern California — over alleged theft of trade secrets. And in the same move, it becomes clear: the revamped Siri arriving this fall won’t rely on OpenAI, but on Google Gemini.
What Apple alleges
The language in the complaint is unusually sharp. OpenAI, Apple says, has been taking its confidential information «at every level, from members of its Technical Staff to its Chief Hardware Officer» — to build its own consumer hardware.
At the center sits Tang Tan, OpenAI’s hardware chief and a former Apple vice president. He allegedly directed Apple employees interviewing at OpenAI to share internal secrets as part of the process. One former Apple employee is said to have downloaded dozens of confidential hardware files — including detailed information on unreleased products, engineering presentations, and technical specs.
The real bombshell: Siri moves to Gemini
Almost more interesting than the lawsuit is the confirmation buried alongside it: Apple’s new Siri, due this fall, is built on Google’s Gemini models — not on ChatGPT anymore. That ends the speculation over whether OpenAI stays a core part of Apple Intelligence. The answer is a firm no.
It’s a striking reversal. Back in 2024 the two companies signed a high-profile partnership. Now it’s a lawsuit on one side and Google on the other.
My take
Two things stand out. First: the AI race has long been a hardware race too. It’s no longer just about models — it’s about devices, and Apple is defending its home turf hard. You don’t pick a complaint with this tone lightly.
Second, and more relevant for us here: Anthropic doesn’t show up in this poker game at all. Apple picks Google, not Claude. The big platform deals — who sits inside which operating system — are being handed out right now, and Anthropic is deliberately betting on other channels: enterprise, coding, government. Whether that’s the smarter call, we’ll find out in a year or two.
Sources: