Apple’s long-awaited overhaul of Siri may be powered by a surprising partner behind the scenes…Google.

The company is reportedly paying Google to build a custom Gemini AI model to run on Apple’s servers, set to debut in next version of iOS, according to Bloomberg’s Mark Gurman.

The new Siri architecture will reportedly feature three key components: a query planner, a knowledge search system and a summariser.

Google’s Gemini models are said to handle the planner and summariser tasks, while Apple’s own Foundation Models will process personal data on-device, ensuring privacy remains intact.

Apple is believed to have considered Anthropic’s Claude models but ultimately chose Gemini due to financial reasons and the companies’ existing relationship – Google pays Apple to remain the default search engine on iPhones.

The arrangement allows Apple to leverage Gemini without publicly acknowledging it, keeping all branding firmly under Apple’s name.

The Gemini-backed backend could enable Siri to answer questions more intelligently, potentially reducing reliance on basic web results or third-party integrations.

However, users will see only Apple’s interface, with Gemini’s role hidden entirely.

Bloomberg suggests this move mirrors strategies seen in other tech companies, such as Samsung, which layers its Galaxy AI features over Google Gemini.

The decision reflects Apple’s practical need to close gaps in its own AI capabilities and finally deliver the AI-powered features promised last year.

Apple is also targeting a new display designed to showcase the updated Siri.

While the partnership remains unconfirmed and financial details undisclosed, a Gemini-powered Siri could mark one of Apple’s most significant AI moves yet.