Google’s Gemini AI is set to become the $1 billion-a-year “behind-the-scenes” engine for Apple’s Siri. This deal, a major victory for Google, will embed its 1.2 trillion parameter model deep inside its rival’s ecosystem.
Apple is using the technology as an “interim solution” for its “Glenwood” project, the internal plan to fix Siri. The new “Linwood” Siri, due in the spring, will be a hybrid, using Google’s AI for complex “planner” and “summariser” functions.
This partnership was formed after Google’s “ultrapowerful” model won an extensive “bake-off” against OpenAI’s ChatGPT and Anthropic’s Claude. The move is a reluctant admission from Apple that its 150-billion parameter models are no longer competitive.
Overseen by top executives Craig Federighi and Mike Rockwell, Apple is pushing its own teams to build a 1T+ model replacement. However, catching Google’s rapidly advancing AI is an immense challenge.
Privacy is a non-negotiable component. The Gemini model will be hosted on Apple’s “walled-off” Private Cloud Compute servers. This guarantees that Google gets its $1 billion fee but zero access to Apple’s user data.