Apple is reportedly close to finalizing a deal with Google that would cost around $1 billion annually to integrate Google’s Gemini AI model into Siri. The Gemini model, featuring 1.2 trillion parameters, will temporarily enhance Siri’s capabilities, including complex planning and query summarization, while Apple continues developing its own AI system with 1 trillion parameters.
According to Bloomberg, the integration is intended to boost Siri’s intelligence and responsiveness, helping the assistant catch up to competitors in context understanding and flexibility. Gemini will run on Apple’s Private Cloud Compute servers to safeguard user data and prevent direct access to Google’s infrastructure.
Internally, the project is known as Glenwood, overseen by Apple’s VP of Software Development Craig Federighi and Vision Pro lead Mike Rockwell. The updated voice assistant, codenamed Linwood, is expected to launch in spring 2026 with iOS 26.4.
Apple considers this collaboration a temporary solution, as it is developing its own AI system projected to match or exceed Gemini’s capabilities. Google, in turn, strengthens its position as a leading AI provider, with Gemini 2.5 Pro already ranking highly among language models and adopted by companies like Snap for advanced AI functionalities.
Due to regulatory restrictions in China, the updated Siri likely won’t use Gemini there. Instead, Apple is testing a localized AI version in collaboration with regional partners, including Alibaba and Baidu, to comply with local laws.
The tech giants are keeping financial details and the exact scope of the partnership private, treating Google strictly as a technology supplier rather than a strategic partner akin to their long-term Safari collaboration.
Previously, Google incorporated Gemini into Google Maps to enhance AI-driven features for users, demonstrating the model’s growing adoption across platforms.
