Apple's next-generation Siri will run on Nvidia's Blackwell B200 data center chips accessed through Google Cloud, The Information reported, a move that puts two of Apple's biggest rivals at the center of its AI strategy. The Gemini-powered Siri, expected to launch with iOS 27 later this year, will tap into Google's fleet of Nvidia Blackwell B200 GPUs for processing user queries. Apple has already established that the new Siri will be based on Google's Gemini AI model, making the Nvidia hardware play a logical next step.
What makes this notable: Apple is outsourcing critical infrastructure for its most important AI product in years, a departure from its historical insistence on controlling every component. The Information noted the "move diverges from [the company's] strategy of attempting to control all the critical ingredients to its products."
Apple will enable Nvidia's confidential compute feature, a hardware-based security system that encrypts data while it's actively being processed on the GPUs. Nvidia says the feature "preserves the confidentiality and integrity of AI models deployed on Rubin, Blackwell, and Hopper GPUs," allowing workloads to run securely "at scale with near-native performance, even in shared or cloud environments." The Blackwell B200 is Nvidia's flagship data center GPU, designed for large-scale AI training and inference with support for trillion-parameter models. It succeeds the Hopper architecture with improvements in memory bandwidth, inference speed, and multi-GPU scaling. The new Siri is expected to ship as a dedicated chatbot app with iOS 27, capable of remembering personal context and reading anything displayed on screen. Apple has not officially shown the new Siri yet.
It remains unclear how Apple's existing Private Cloud Compute server system fits into the upcoming Siri launch. Apple is expected to reveal details at WWDC 2026, which opens June 8.
