← Back to Insights

Apple WWDC 2026: Siri Gets a New Brain, Powered by Google Gemini

Nils Liu
Apple Siri Google Gemini WWDC iOS 27 AI News

TL;DR

At Tim Cook's final WWDC keynote, Apple announced a rebuilt Siri running on a custom 1.2-trillion-parameter Google Gemini model at roughly $1B/year. iOS 27 also lets users swap Siri for ChatGPT, Claude, or Gemini via a new Extensions system.

Apple WWDC 2026: Siri Gets a New Brain, Powered by Google Gemini

Apple WWDC 2026 opened in Cupertino today with what will be Tim Cook’s final keynote as CEO, and the headline announcement was not a new device: Siri has been rebuilt from the ground up, with Google’s Gemini as its new core engine.

Bloomberg reported that Apple has signed a multi-year deal worth roughly $1 billion annually for a custom version of Google’s Gemini AI, featuring approximately 1.2 trillion parameters. That makes it one of the largest foundation models ever deployed in a consumer product. Rather than training a frontier model internally, Apple chose to license Google’s best work and build around it.

What the New Siri Actually Does

The redesigned Siri introduces three capabilities its predecessor could not handle:

Personal context: Siri can now access your emails, messages, photos, and files to construct meaningful answers. Ask it who you’re meeting next week and it searches your calendar and inbox to respond, instead of directing you to open an app.

Onscreen awareness: Siri understands what is currently on your display and acts on it directly, with no need for users to describe the context.

Cross-app task execution: Multi-step actions spanning multiple apps can now be delegated to Siri end-to-end.

Apple is also launching a standalone Siri app with a ChatGPT-style chat interface, supporting both text and voice with conversation history synced across devices. History retention can be set to 30 days, one year, or indefinite.

Extensions: Let Users Pick Their AI

iOS 27 introduces an Extensions framework that lets users route requests to ChatGPT, Claude, or Gemini instead of Siri, with each service assignable to its own voice.

The feature reads as strategic hedging. Google Gemini holds the core Siri license, but Anthropic and OpenAI both appear in the options list. All three are effectively paying Apple for user surface access while Apple controls the default. Apple also announced an AI Agent App Store, letting third-party services plug into reservation booking, smart home control, and other delegated task flows.

Why Google Got the Deal

Apple’s prior arrangement with OpenAI was a commercial partnership for ChatGPT integration, not a core infrastructure deal. The Google agreement is categorically different: it replaces Siri’s underlying model.

Several factors likely tipped the balance. Gemini’s multimodal training data and infrastructure scale are unmatched. The Gemini 3.5 Flash announcement at Google I/O 2026 in May demonstrated dramatic improvements in speed and cost-efficiency, making large-scale on-device deployment practical. Apple and Google also have over a decade of search licensing history, meaning the legal and commercial frameworks for a new AI deal were already well-established.

Tim Cook’s Final Bet

Cook has led Apple since 2011, growing its market cap from roughly $350 billion to over $4 trillion.

His final major product call before handing over to a new CEO: make Google the engine behind Apple’s AI future. Whether Gemini-powered Siri can close the gap with ChatGPT in real-world use remains the open question the next Apple leadership team will have to answer.

If this was useful, subscribe to the newsletter for weekly AI PM insights and GenAI case studies.

Get the latest insights

Join the newsletter to receive my latest articles on GenAI, AI Agents, and architecture.

No spam. Unsubscribe anytime.