This might be the biggest partnership of the year — and it was confirmed almost casually. Google Cloud chief Thomas Kurian stated at Cloud Next ‘26 that Gemini will power a new, context-aware version of Siri. Later this year.
What we know
Details are still thin, but the core message is clear: Apple will use Google’s Gemini models to fundamentally improve Siri. Kurian talked about a ‘personalized’ version that understands context on the user’s device — not just a voice-activated search engine, but an assistant that actually knows what you’re doing.
This fits the rumors that have been circulating for months: Apple Intelligence needs better models than what Apple has built internally so far. And Google has the infrastructure and model family with Gemini to deliver that.
Why this matters
Siri is installed on over a billion Apple devices. If Gemini runs under the hood there, it becomes the largest single deployment surface for a language model ever. For Google, that means massive reach without having to sell its own hardware. For Apple: finally an assistant that isn’t embarrassing.
For Anthropic and OpenAI, this is a problem. Both tried to get into the Apple ecosystem slot. Google has apparently won that race — at least for Siri.
My take
This is a deal that could shift the AI power balance. Not because of the technology — Gemini, Claude, and GPT are all at a similar level. But because of distribution. If you run on a billion iPhones, you win access to the most users. And that’s what the AI race is really about right now: not just having the best model, but getting it into the hands of the most people.
Sources: