Apple is reportedly preparing to pay Google around $1 billion annually to use its powerful Gemini AI model in the upcoming version of Siri. According to sources close to the matter, both companies have been negotiating for months, and the deal is now nearing completion. This move marks Apple’s most significant step yet toward revitalizing its voice assistant and competing in the rapidly advancing AI race.
Through the partnership, Apple will gain access to Google’s 1.2-trillion-parameter Gemini model — an enormous leap compared to its current 150-billion-parameter system. The company plans to use Gemini as an interim solution while continuing to develop its own in-house AI technologies. With this integration, Siri’s core framework will be rebuilt to enable deeper contextual understanding and more natural interactions.
Bloomberg reports that the new Siri is expected to launch alongside iOS 26.4 in the spring, though Apple remains cautious about potential schedule changes. Internally known as “Glenwood,” the project is led by Mike Rockwell, who previously worked on Vision Pro, and software engineering head Craig Federighi. The upgraded Siri itself carries the codename “Linwood.”
Once Gemini is integrated, Siri will go beyond simple commands — gaining abilities to summarize information, plan complex tasks, and assist with decision-making. While Apple will rely on its own models for certain Siri features, Gemini will handle key aspects like summarization and planning. The model will run on Apple’s Private Cloud Compute servers, keeping all user data fully isolated from Google’s systems.
Notably, this collaboration is independent of any plans to integrate Gemini’s chatbot features directly into Siri. Meanwhile, Apple continues to expand its partnership with OpenAI for ChatGPT integration.