Apple is reportedly preparing to use Nvidia’s Blackwell B200 processors to power a long-awaited overhaul of Siri. According to a recent report, these chips will be hosted in Google’s data centers, marking a significant shift in Apple’s approach to AI infrastructure. The move comes after Apple’s tests on its own Private Cloud Compute servers proved too slow for the upgraded assistant’s needs.
The company had originally planned to rely solely on its in-house hardware for cloud-based Siri tasks, but performance issues forced a reevaluation. Apple experimented with a modified version of Google’s Gemini model on its own servers, but the processing speed was insufficient for practical deployment. As a result, Apple will now route some Siri requests through external infrastructure powered by Nvidia technology.
Despite this reliance on third-party hardware, Apple aims to maintain user privacy. The company is expected to leverage Nvidia’s confidential computing capabilities to encrypt data during processing. This approach could help address privacy concerns associated with using external data centers. However, the final implementation details remain unclear, as Apple has not disclosed how its Private Cloud Compute branding will apply to Nvidia chips hosted by Google.
The revamped Siri, set to be previewed at Apple’s Worldwide Developers Conference in June, represents the assistant’s biggest redesign since its inception in 2011. New features may include personal context awareness and the ability to understand on-screen content. A dedicated Siri app could also be introduced to compete with other AI assistants. The update is expected to roll out to users in September.
This partnership with Nvidia is a significant win for the chipmaker, adding one of the world’s most valuable companies to its customer base. Apple has not officially confirmed the arrangement, but the upcoming WWDC keynote is anticipated to reveal the extent of the Siri overhaul after years of delays.