The planned new AI iteration of Apple’s voice assistant, Siri, may rely on Nvidia chips hosted within the Google Cloud. According to the well-informed US magazine “The Information”, Apple is considering combining Google’s Gemini AI models with Nvidia’s powerful hardware, opting for partnerships rather than building the entire system internally.
This report comes just ahead of the annual WWDC 2026 developer conference, where Apple plans to debut the revamped Siri. Following several related previews, it has become clear that Siri is evolving into a sophisticated “AI Agent” capable of autonomously extracting tasks from searches across emails, messages, files, and photos. The new Siri is expected to recognize on-screen content, save contacts directly from message threads, and share photos upon request.
Despite possessing substantial financial resources, Apple has historically lagged in AI innovation, a gap that experts suggest the company is now addressing through strategic acquisitions and partnerships.
For the extensive AI processing required, Apple is reportedly going to utilize Nvidia’s Blackwell B200 chips, which are available through the Google Cloud infrastructure. The Blackwell B200 is Nvidia’s flagship model for data centers, specifically designed for the training and operation of large AI models. These chips, based on the newest Blackwell architecture, promise significant improvements in performance, memory bandwidth, and AI processing compared to previous generations, enabling the advanced functions Apple has planned for Siri.
However, it remains unclear how this integration will align with Apple’s own “Private Cloud Compute” infrastructure. Apple introduced this system as part of its “Apple Intelligence” platform, promising secure cloud-based AI processing with no compromises to data privacy. It is rumored that a specific feature of Nvidia’s technology will be used to ensure that data remains encrypted even while being processed by the chips.
