Google’s latest Pixel AI push keeps your data on your phone

TL;DR

• Google is shifting Pixel intelligence from the cloud to the device, enabling faster performance and offline functionality.
• Gemma 4 E2B for TPU is built specifically for Pixel’s Tensor chip, allowing apps to process requests locally.
• Pixel 10 gains practical offline capabilities, including trip planning, recipe suggestions, smart home controls, and Mobile Actions for hands-free device commands.

Google’s been hard at work making Pixel phones smarter with AI, but its next big push is to cut the cloud out of the equation. The company is doubling down on on-device AI, offering developers new tools to build apps that run locally on Pixel hardware for faster responses, better privacy, and offline access.

The company’s latest announcement is the Gemma 4 E2B for TPU, a lightweight version of Google’s open AI family that’s been optimized for the Tensor Processing Unit inside Pixel devices. The model runs AI tasks on the phone itself, eliminating the need to send requests to remote servers, so apps can continue to function even without an internet connection, while sensitive data remains on the device.

Google also demonstrated what the new technology looks like on the Pixel 10. During the recently concluded Google I/O India, Google showcased offline AI agents that can help plan trips, suggest recipes, and control smart home devices. Another feature, Mobile Actions, lets users enable or disable Wi-Fi or launch Google Maps with voice or text commands processed entirely on the device.

Multimodal AI is another big area of focus. The Pixel 10 can have AI conversations without an internet connection, identify landmarks or plants in photos, and transcribe lectures or voice notes on the device itself, per Google’s announcement. These capabilities could be particularly useful when traveling, flying, or working in areas with unreliable connectivity.

The company also pointed to enterprise use cases such as creating offline shopping maps for retail stores and helping mechanics identify defective car parts using photos taken on a Pixel device.

Thank you for being part of our community. Read our Comment Policy before posting.