Apple is reportedly working with Google and Nvidia to power its next-generation Siri, marking a major shift in the company’s usual approach of keeping key technologies in-house.
The upgraded Siri is expected to use a licensed version of Google’s Gemini model for some queries, while Google Cloud infrastructure would rely on Nvidia Blackwell chips to process AI workloads.
The new Siri is expected to handle some requests through Google Cloud using Gemini, rather than relying only on Apple’s own servers or on-device processing.
The move would give Apple access to a more powerful AI model while it continues building out its own Apple Intelligence features. It also suggests that Apple may be willing to use outside AI infrastructure for more complex assistant tasks.
Google’s data centers are expected to use Nvidia’s Blackwell B200 chips for Siri-related AI processing.
The chips support confidential computing, which encrypts data while it is being processed. This could help Apple maintain stronger privacy protections even if some Siri requests run through third-party cloud infrastructure.
Nvidia has also highlighted its work with Google Cloud around Gemini, Blackwell infrastructure and confidential computing for sensitive AI workloads.
Apple has built much of its AI messaging around privacy, including Private Cloud Compute, which is designed to process more complex AI requests in the cloud without storing user data.
Apple says data sent to Private Cloud Compute is used only to complete the request and is not retained.
Using Google Cloud and Nvidia hardware would create a different setup, so privacy protections will be closely watched. The use of confidential computing appears aimed at reducing that concern by keeping data encrypted during processing.
The first version of the upgraded Siri is expected to arrive in September, likely alongside Apple’s next iPhone software and hardware cycle.
Apple has not officially confirmed the reported Google and Nvidia arrangement, so final details may change before launch.