Apple’s Siri isn’t as good as any other AI assistant, and hence, the US tech giant is taking a drastic measure to catch up in the AI race. Apple has inked a partnership deal worth $1 billion with Google to use its advanced Gemini large language model (LLM) to power the Siri assistant. According to a report from Bloomberg, the partnership is a temporary measure that will see Apple pay Google $1 billion per year to utilise a 1.2 trillion-parameter AI model that surpasses the capabilities of Apple’s current in-house AI models.
This partnership follows a long delay in Apple’s development of its next-generation smart assistant, which was initially planned for debut in an earlier iOS update. With rivals like OpenAI’s ChatGPT and Google’s own Gemini advancing at a rapid pace, Apple has opted to integrate a proven, powerful third-party solution to deliver the “Apple Intelligence” features promised to users.
Apple falls back on Google to power Siri, Apple Intelligence
The deal provides a win-win situation for both parties. Apple gets access to world-class LLM capabilities immediately for powering functions like AI content summarisation and multi-step task execution, whereas Google secures a huge financial win. It is said that the Gemini model provided to Apple will run on Apple’s Private Cloud Compute servers, thus ensuring that Google does not gain access to user data, which has been a key privacy commitment for Apple.
This new Gemini-powered, more capable Siri is now expected to be introduced in the upcoming iOS 26.4 update, which is expected to debut in Spring 2026.
Apple-Google competition downed?
Apple’s decision to side with Google for its AI partnership comes after considering internal models and testing alternatives from companies like OpenAI and Anthropic. Ultimately, the existing multi-billion-dollar relationship with Google, which already pays Apple approximately $20 billion annually to be the default search engine on its iPhones, iPads and Macs, proved decisive.
Industry Insiders suggest this partnership is a short-term fix. Apple plans to aggressively continue development on its own large language models, including a proprietary 1 trillion-parameter cloud model that could be ready by 2026. Once Apple’s in-house LLMs are deemed capable enough to support Siri’s new features, Apple is expected to switch from the Gemini partnership entirely.
With Gemini onboard for now, Apple expects Siri and other Apple Intelligence features to be able to handle complex queries and execute tasks across multiple apps, finally bringing Apple’s digital assistant closer to the performance of rivals like ChatGPT and Perplexity.
