The January 12, 2026 joint statement was notably direct: “After careful evaluation, we determined that Google’s technology provides the most capable foundation for Apple Foundation Models and we’re excited about the innovative new experiences it will unlock for our users.” At Google Cloud Next ’26, Google Cloud CEO Thomas Kurian reaffirmed the partnership, stating the companies are “collaborating as Apple’s preferred cloud provider to develop the next generation of Apple Foundation Models based on Gemini’s frontier capabilities.”
Running a 1.2-trillion-parameter model for consumer queries at Apple’s scale requires immense compute. For the cloud-heavy queries Siri will field, Apple has turned to Nvidia’s latest Blackwell GPUs. Reports from The Information indicate Apple has placed large orders for Nvidia hardware, using the chips to handle inference for the Gemini-powered Siri rather than relying solely on Google’s own data centers for every request.
Complex Siri queries will route to Google Cloud running the licensed Gemini model on Nvidia hardware, while Apple employs Nvidia’s confidential computing capabilities to encrypt data during cloud processing. This means Apple retains full user-data control and privacy guarantees even when tapping Google’s cloud infrastructure.
Despite using Google’s AI models and cloud resources, Apple processes all Siri queries through its own Private Cloud Compute infrastructure. Instead of sending user data directly to Google’s servers, queries are routed through Apple’s privacy-preserving layer, maintaining the on-device-first privacy guarantees Apple emphasizes. Smaller, distilled on-device models handle simpler tasks locally, while Gemini is used for cloud-scale queries and to train those smaller models through distillation.
At a functional level, the overhaul transforms Siri from a voice assistant into a full-fledged AI chatbot designed to compete with ChatGPT, Claude, and Google’s own standalone Gemini app. According to Bloomberg’s Mark Gurman and corroborating reports, the new Siri introduces:
In a notable strategic shift, Apple is opening Siri to third-party AI chatbots through an Extensions system coming in iOS 27. While Apple Intelligence already supports a ChatGPT handoff, users will now be able to select Claude, Gemini, or other competing assistants as their preferred chatbot inside Siri, turning the assistant into a distribution platform for AI models rather than a walled-garden service. Tim Cook confirmed on a Q4 2025 earnings call that the company is pursuing additional AI partnerships beyond OpenAI and Google.
This overhaul did not follow a smooth product roadmap. Apple first unveiled Apple Intelligence at WWDC 2024 with a sizzle reel promising a personalized Siri capable of personal context, on-screen awareness, and deeper app integration. Those features were delayed in March 2025, promised again at WWDC 2025, pushed to an internal iOS 26.4 target, and ultimately scrapped after Apple determined its in-house models were not progressing fast enough.
By November 2025, Apple finalized an agreement for Google to supply Gemini models. Internally, the project is reportedly code-named “Campos,” and it represents Apple essentially rebuilding Siri from the ground up to deliver the conversational AI experience it could not achieve alone. The revamped assistant is expected to ship with iOS 27, iPadOS 27, and macOS 27 in September 2026, and will also power broader Apple Intelligence features across the ecosystem.
Comments
0 comments