Apple’s upcoming iOS 27 Siri AI system will reportedly route some of its complex queries through Google Cloud servers. These requests will be processed by Google’s high-end NVIDIA Blackwell B200 GPU clusters, marking a significant partnership in the artificial intelligence landscape.
While Apple plans to maintain a privacy-first approach by running basic AI operations locally on user devices and private cloud servers, heavy-duty workloads will require external computing power. To balance performance and server load, Apple will utilize authorized Gemini models hosted on Google’s cloud infrastructure.
Addressing Privacy Concerns with Confidential Computing
Sending user requests to third-party cloud servers naturally raises privacy questions. To mitigate these concerns, Apple and Google are implementing hardware-level security measures. Apple will leverage NVIDIA’s confidential computing technology, which keeps data encrypted even while it is being processed by the GPU.
Confidential computing ensures that sensitive information remains secure in shared cloud environments, protecting both the privacy of user data and the integrity of the AI models. NVIDIA has confirmed that this hardware-level security is fully supported on its Hopper, Blackwell, and upcoming Rubin GPU platforms.
The Power of the Blackwell B200
The NVIDIA Blackwell B200 is one of the most powerful data center GPUs designed for high-performance AI training and inference. By leveraging these advanced GPU clusters, Apple’s iOS 27 Siri AI can handle complex, multi-trillion parameter model calculations with lower latency, ensuring a faster and more capable assistant experience for users when the update rolls out.
