Apple picks Google's Gemini to power next-gen AI features, including Siri
Apple and Google have officially teamed up, with Gemini set to power the overhauled Siri and next generation Apple Intelligence features rolling out later this year
 "Apple picks Google's Gemini to power next-gen AI features, including Siri")
After months of speculation and rumours, Apple and Google have finally confirmed that the latter’s Gemini will be powering Siri and future Apple Intelligence features. Apple will use Google’s Gemini as the underlying technology for its foundation models, which will power next-generation Siri, with the change scheduled to roll out to users later this year. Both companies issued a joint statement confirming this, where they said that after careful evaluation, Apple determined that Google’s AI technology provides the most capable foundation for Apple's foundation models.
Apple and Google partnership: What will it result in
According to a joint statement by the companies, Apple and Google have entered into a multi-year collaboration. The key result of this partnership will be that the next generation of Apple foundation models will be based on Google's Gemini models and cloud technology. This includes the personalised version of Siri and Apple Intelligence features, some of which were showcased at WWDC 2024 and are set to be launched later this year.
Although Apple rolled out most of the Apple Intelligence features it unveiled at WWDC 2024, the redesigned version of Siri was notably held back. In a statement shared with John Gruber in May, Apple acknowledged that development was taking longer than anticipated, leading to delays in bringing those promised Siri upgrades to users.
Neither company has clarified exactly which Siri capabilities will rely on Gemini. However, a report by 9To5Google suggests Apple will continue to use its own models for certain functions, while tasks such as summarisation and planning are expected to be handled by Google’s AI.
Gemini-powered Siri: What to expect
According to Bloomberg, the upcoming version of Siri is said to be built around three main layers: a query planning system, a knowledge retrieval engine, and a summarisation component. Reports suggest that Google’s Gemini models will power the planning and summarising tasks, helping Siri structure responses and generate clear summaries from user prompts.
Despite Gemini’s involvement, these operations will run on Apple’s Private Cloud Compute platform, which is designed to safeguard user privacy and ensure that data is not shared externally.
Gurman added that Gemini could also be used for the knowledge search layer. If implemented, this would significantly expand Siri’s general knowledge capabilities, going beyond its current approach, where more complex or contextual questions are handed off to ChatGPT with user consent. The Gemini-based setup is expected to be more deeply integrated into Siri’s core functions rather than acting as an optional hand-off.
ALSO READ: Apple Health app redesign and new features may arrive with iOS 26.4: Report