Apple Puts Google's Gemini Inside Siri, Rewriting the AI Assistant Landscape

Tim Cook walked onto the stage at Apple Park on June 8 for the last time as chief executive and announced that the company's most recognizable product — Siri — would be rebuilt around Google's Gemini model. The decision ends years of speculation about whether Apple would develop its own large language model capable of competing with GPT and Claude, or partner with an external provider. Apple chose partnership, and it chose Google.

The new Siri, branded as part of the Apple Intelligence platform, replaces the underlying model that has powered the assistant since its introduction in 2011. Alongside the Gemini integration, Apple announced a multi-AI extension system that allows third-party models to connect directly to iPhone functions. Anthropic's Claude becomes available to iPhone users through this framework for the first time, giving Apple an architecture that can route different types of requests to different models depending on the task.

Cook also announced iOS 27 Beta 1, released the same afternoon as the keynote, a move designed to signal development velocity and give developers immediate access to the new AI capabilities ahead of the fall consumer release.

The EU responded within 24 hours. European regulators confirmed that none of the Gemini-powered Siri features will reach users in European Union member states, citing ongoing concerns about data handling and the market dynamics of embedding a dominant search provider's model into the world's most widely used smartphone platform. Apple's characterization of the regulatory standoff as nearing resolution was directly contradicted by the EU announcement, leaving European iPhone users on a separate feature track from the rest of the world.

The strategic logic behind the Google partnership is straightforward. Apple's core strength has always been hardware integration and user experience, not model training at scale. Building a frontier language model internally would require infrastructure investment and talent acquisition that would compete directly with Apple's existing capital allocation priorities. Licensing Gemini gives Apple a competitive assistant without that overhead.

For Google, the deal is equally compelling. Embedding Gemini inside iOS gives the model exposure to over one billion active iPhone users, a distribution channel that no amount of direct consumer marketing could replicate. The arrangement also positions Google favorably as it prepares to respond to competitive pressure from OpenAI's consumer products and from Microsoft's continued integration of AI into Windows and Office.

The multi-AI architecture Apple unveiled may prove to be the more consequential long-term announcement. By treating AI models as interchangeable components that can be selected based on task type, Apple is creating a marketplace dynamic where model developers compete for placement inside one of the most valuable distribution systems in consumer technology. That competition has implications for pricing, capability development, and the pace at which AI improvements translate into features that ordinary users can see and use.