Google has secured a monumental $1 billion-per-year contract that will see its Gemini AI model become the new “behind-the-scenes” engine for key functions in Apple’s Siri. This deal places Google’s “ultrapowerful” 1.2 trillion parameter model at the heart of its main competitor’s ecosystem.
The partnership was finalized after an extensive “bake-off,” where Google’s advanced AI was pitted against models from OpenAI and Anthropic. Gemini emerged as the clear winner, demonstrating superior capabilities that Apple needed for its next-generation assistant.
For Apple, this is an “interim solution” to power its “Glenwood” project, the internal initiative aimed at fundamentally fixing Siri’s long-standing issues. The new assistant, codenamed “Linwood” and planned for a spring release, will be a hybrid system.
This hybrid model means Google’s AI will handle the most complex “summariser” and “planner” functions, enabling Siri to understand and execute multi-step tasks. Simpler requests and other Apple Intelligence features will continue to run on Apple’s in-house 150-billion parameter models.
Privacy remains a cornerstone of the agreement. The 1.2 trillion parameter Gemini model will be hosted on Apple’s Private Cloud Compute servers, not Google’s. This “walled-off” architecture ensures Google never gets access to Apple user data, satisfying Apple’s non-negotiable privacy-first stance.
Google Banks $1 Billion Annually as Gemini AI Becomes Siri’s New Engine
