Google Gemini to secretly power part of Apple’s revamped Siri, report claims

Apple is reportedly partnering with Google to integrate aspects of its Gemini AI models into the upcoming revamped version of Siri, expected to launch with iOS 26.4, according to Bloomberg’s Mark Gurman.

In his latest Power On newsletter, Gurman says that Apple will use a custom Gemini model running on its Private Cloud Compute servers to handle parts of Siri’s new intelligence system — specifically for planning and summarisation tasks.

While the collaboration will reportedly play a significant role in how Siri processes and fulfils requests, Apple does not plan to publicly acknowledge the partnership, unlike its high-profile integration of OpenAI’s ChatGPT within Apple Intelligence.

As previously outlined by Bloomberg, the new Siri architecture will feature three main components — a query planner, a knowledge search system, and a summariser. The Gemini models are expected to assist with the planner and summariser capabilities, while user privacy will remain protected through Apple’s infrastructure, ensuring no external data sharing.

The knowledge search component could also draw on Gemini, allowing Siri to better understand world topics and trivia — helping it answer questions more naturally, without defaulting to third-party results or basic web searches.

Despite Gemini’s extensive role behind the scenes, the report claims Apple will continue to market Siri as entirely Apple-developed, powered by Apple’s servers and presented through its own interface.

The arrangement mirrors a similar approach used by Samsung, whose “Galaxy AI” features are partly underpinned by Google’s Gemini models.