Apple’s ambitious plans for a smarter, more capable Siri will rely on Google’s cloud infrastructure and Nvidia’s latest data center chips, marking a significant shift from the company’s traditional approach of keeping user data processing in-house. The partnership signals Apple’s recognition that its own server hardware isn’t powerful enough to handle the most demanding AI workloads at scale.
The collaboration builds on an earlier joint announcement between Apple and Google, where the companies revealed that Apple’s next-generation Foundation Models would use Google’s Gemini technology. These models will power the new Siri and other Apple Intelligence features expected to be unveiled at Monday’s WWDC keynote.
While Apple will still try to process as many Siri requests as possible directly on users’ devices, more complex queries requiring significant computational power will be handled by Google’s cloud servers equipped with Nvidia’s Blackwell B200 data center chips. This represents a notable departure from Apple’s previous strategy of relying primarily on its own hardware infrastructure.
The decision comes after Apple’s own Private Cloud Compute system, announced at WWDC 2024, proved inadequate for the new AI models during testing. That system was designed to run on Apple’s custom Silicon chips, but apparently couldn’t deliver the performance needed for the more sophisticated Siri capabilities Apple wants to offer.
Privacy concerns – always a key consideration for Apple – will be addressed through multiple layers of protection:
- Nvidia’s “confidential computing” technology that encrypts data during processing
- Additional privacy and security measures from both Apple and Google
- Integration with Apple’s existing privacy frameworks
This partnership highlights the intense computational demands of modern AI applications and the challenges even tech giants face in scaling advanced AI services. For Apple, known for its tightly controlled ecosystem and emphasis on user privacy, relying on external cloud infrastructure represents a pragmatic acknowledgment of current technical limitations.
The move also underscores the growing importance of Nvidia’s data center business, as companies across the tech industry compete to offer more sophisticated AI-powered features. Google’s cloud infrastructure, combined with Nvidia’s specialized AI chips, provides the computational muscle Apple needs to deliver on its AI ambitions.
Apple is expected to showcase the redesigned Siri at WWDC on Monday, June 8, where the company will likely detail how the new AI capabilities will work in practice and address user concerns about data privacy in this new cloud-based architecture.