Apple is reportedly paying Google $1 billion annually for its Gemini artificial intelligence model to power Siri, underscoring the iPhone maker’s rapid efforts to catch up in the competitive AI landscape.
This significant expenditure highlights Apple’s strategic reliance on a rival to boost its voice assistant capabilities. The deal provides access to an “extremely powerful” model featuring 1.2 trillion parameters.
Google’s Gemini will enhance Siri’s summary and planning functions, enabling the assistant to synthesize information and execute complex tasks. Apple intends to keep Google’s involvement as a “behind-the-scenes” technology provider.
This approach means iPhone users will likely remain unaware that Apple is leveraging technology from the creator of the Android operating system to power its artificial intelligence.
The 1.2 trillion-parameter custom Gemini system represents a substantial upgrade from the 150 billion-parameter model currently used for Apple Intelligence’s cloud-based features.
Critically for data privacy, Google’s model will operate on Apple’s Private Cloud Compute servers, ensuring separation from Google’s own infrastructure.
Internally, the new Siri, planned for iOS 26.4, is codenamed “Linwood,” while the project to integrate a third-party model is known as “Glenwood.”
The enhanced Siri is expected to launch in spring 2025. This adoption of Gemini is characterized as a “temporary solution,” as Apple acknowledges it lags in AI development.
Apple is simultaneously developing its own 1 trillion-parameter cloud-based model. The company hopes this in-house solution will be ready for consumer applications next year and could eventually replace Gemini.
In China, where Google services are prohibited, Siri’s Gemini integration is unlikely to be available. Apple is exploring alternatives for this market.
These alternatives include using its own models in conjunction with an Alibaba-developed filter. Apple is also reportedly considering a partnership with Baidu for its AI offerings in China.