Apple is working to shrink Gemini and pack it into the iPhone to power the new Siri

Apple is reportedly adapting Google Gemini for a rebuilt Siri, but the model is too large to run entirely on-device.

The new Siri would likely split work between the iPhone and cloud processing, using NVIDIA confidential computing instead of Google servers.

This approach underscores Apple’s balancing act between AI performance, privacy, and on-device capability.

WWDC26 is expected to be the stage for Apple’s Siri overhaul.