Apple Shifts AI Strategy to Google’s Nvidia Chips Amid Privacy Concerns

Apple's long-awaited Siri overhaul will now rely on Google's Nvidia Blackwell B200 data center chips after internal infrastructure proved inadequate. The pivot comes nearly two years after Apple first promised a redesigned AI assistant, with the official reveal slated for June 8 at WWDC alongside iOS 27 previews.

Originally, Apple planned to process cloud-based queries exclusively through its Private Cloud Compute (PCC) system built on custom Mac silicon. Craig Federighi emphasized this architecture would prevent third-party data access during WWDC 2024. But technical limitations emerged when attempting to run Gemini models on PCC, forcing the compromise with Nvidia hardware.

To mitigate privacy risks, Apple will implement Nvidia's confidential compute feature—encrypting data during cloud processing. The move highlights the growing computational demands of generative AI, even for trillion-dollar tech giants.