Apple may use Nvidia chips to power its Gemini-backed Siri upgrade

Apple’s next Siri upgrade may lean on more than just Google’s Gemini model. According to The Information, Apple plans to use Google Cloud servers powered by Nvidia chips for parts of the new Siri experience expected in September.

The report says Apple will use Nvidia’s Blackwell GPUs inside Google Cloud for heavier Siri requests. Apple has not confirmed the plan, so this remains a reported arrangement, not an official Siri announcement.

Why Nvidia chips matter here

The chip detail matters because Apple has to make outside AI compute fit its privacy pitch.

Apple already uses Private Cloud Compute for more complex Apple Intelligence requests, with the company saying those requests run on Apple silicon servers built around strict privacy protections. If Apple is using Google Cloud for Gemini-backed Siri work, it needs a way to keep user data protected even when the model and hardware are outside Apple’s own stack.

Your Complete iOS 26 Playbook:

Get our exclusive Ultimate iOS 26 Guide 📚 — absolutely FREE when you sign up for our newsletter below.

That is where Nvidia’s confidential computing comes in. Nvidia says its newer data center chips can run AI workloads in protected environments with hardware-backed checks, which could help Google Cloud process sensitive requests without exposing the underlying user data in the usual way.

For iPhone users, the point is simple: Apple appears to be looking for extra AI horsepower while still trying to make Siri feel like an Apple-private experience.

Siri’s reset is getting more complicated

Apple’s delayed Siri overhaul is expected to be one of the biggest parts of its next Apple Intelligence push. Earlier reports have pointed to a more capable assistant that can understand personal context, handle more app actions, and respond more naturally across the system.

This reported Google-Nvidia setup would add another layer to that story. Apple may still present the feature as part of its own iOS 27 Siri upgrade, but the backend could involve Google models, Google Cloud infrastructure, and Nvidia chips for some requests.

That does not mean every Siri interaction will leave the device or run on Nvidia hardware. Apple’s AI system already splits work between on-device processing, Apple’s cloud servers, and outside models such as ChatGPT when users allow it. The new report suggests the Gemini-backed version of Siri may follow a similar split, with cloud compute reserved for requests that need more power.

September is the reported target

The Information says Apple is aiming to launch the new Siri in September. That timing would put the assistant alongside the next major iPhone software release, though Apple could still preview parts of the plan at WWDC26 on June 8.

The safer read for now is that Apple’s next Apple Intelligence update may depend as much on cloud architecture as on flashy demos. Siri needs to become smarter, but Apple also has to convince users that outside AI partners do not make the assistant feel less private.

That is a harder pitch than simply saying Siri uses Gemini. It explains why Apple may be spending so much effort on the compute layer behind the feature.