Apple will rely on Google’s cloud infrastructure powered by Nvidia’s latest chips to run its next-generation Siri, marking a significant shift in the company’s AI strategy. The tech giant will use Google’s servers equipped with Nvidia’s Blackwell B200 data center chips for complex AI requests that require more processing power than Apple’s own hardware can handle.
The partnership comes after Apple’s Private Cloud Compute system proved too slow during testing for the new AI models. Apple announced earlier this year that its next-generation Apple Foundation Models would be based on Google’s Gemini technology, but new details reveal the extent of the collaboration.
This development highlights Apple’s pragmatic approach to the AI race. While the company has long prided itself on controlling the entire user experience through its own hardware and software, the computational demands of modern AI are forcing even Apple to seek external partnerships. The move also shows how Nvidia’s dominance in AI chips extends beyond just powering ChatGPT and other consumer AI services – it’s now essential infrastructure for Apple’s core products.
Apple plans to run as much of the new Siri as possible directly on users’ devices, but more demanding requests will be processed on Google’s cloud servers. This hybrid approach aims to balance performance with Apple’s privacy commitments. The company will use Nvidia’s “confidential computing” technology, which encrypts data while it’s being processed, along with other security measures to protect user information.
The reliance on external infrastructure represents a notable departure from Apple’s previous strategy. The company’s Private Cloud Compute system, unveiled at WWDC 2024, was designed to run on Apple’s own server hardware using Apple Silicon chips. However, the new AI models apparently require more computational power than Apple’s current server infrastructure can provide at the necessary speed.
Apple is expected to showcase the redesigned Siri at its WWDC keynote on Monday, June 8. The new virtual assistant will reportedly offer more personalized responses and capabilities powered by the Google Gemini models running on this cloud infrastructure.
The partnership underscores several important trends in the tech industry:
- Even Apple acknowledges it can’t build all AI infrastructure in-house
- Google’s cloud business continues to win major enterprise customers
- Nvidia’s chips have become indispensable for serious AI applications
- Privacy-focused companies are finding ways to use cloud AI while protecting user data
For Apple users, this means Siri should become significantly more capable, but it also means some interactions will depend on internet connectivity and Google’s servers. Apple will need to clearly communicate how this affects user privacy and what data, if any, might be processed outside of Apple’s direct control.
The collaboration spans multiple years, suggesting this isn’t a temporary solution while Apple builds its own infrastructure. Instead, it appears Apple has decided that partnering with established cloud providers makes more sense than trying to match Google’s server capacity and Nvidia’s specialized AI chips on its own.
