Apple and Google Confirm Multi-Year AI Deal: Siri to Move to Gemini and Google Cloud

Apple and Google have announced a multi-year collaboration, under which the next generation of Apple Foundation Models will be built on Google Gemini models and related cloud technology. These models are intended to help power future Apple Intelligence features, including a more personalized Siri, which the partners say is coming this year.

What this means in practice: Apple is effectively admitting that for a key part of the AI layer it can no longer rely solely on in-house development. If it wants to catch up with the pace of modern LLM assistants, it needs a partner with a ready, top-tier technology and the infrastructure to operate it at scale.

Why this is a big deal

Siri has been on the market for over ten years, and it has long been the case that compared with today's language models it often feels more limited — especially on more complex queries and multi-step tasks. Apple wanted to change that with the Apple Intelligence project, which at WWDC 2024 promised Siri with better “context” and the ability to work with what the user has on screen.

However, parts of these features were reportedly delayed and the user experience was criticized for inaccuracies. That's why the move to Gemini is now seen as an attempt at a restart – Apple keeps its own product and interface but builds them on a significantly stronger AI foundation.

Extra detail: OpenAI remains, but is not the main choice

Important detail: this is not an 'Apple = OpenAI' story. According to available information, Apple considered multiple players when choosing a solution, including OpenAI, but chose Google Gemini as the main basis for its Foundation Models.

OpenAI (ChatGPT), according to media reports, is not entirely ending in Apple's ecosystem, but will rather function as optional (opt-in) support for more complex queries, while the 'default' AI layer for Apple Intelligence and the future Siri is to be based on Gemini.

What about privacy?

Apple also emphasizes that Apple Intelligence is meant to run directly on devices and, where necessary, via its Private Cloud Compute, and that its long-term commitment to privacy is not intended to be compromised by this.

Privacy will remain a key issue — because working with a cloud giant naturally raises questions about exactly where data is processed and what real guarantees users have.