Skip to content
Tech News
← Back to articles

The Tensor G6’s new TPU is secretly the best Pixel 11 upgrade no one’s talking about

read original more articles
Why This Matters

The Tensor G6’s upgraded TPU in the Pixel 11 series significantly enhances on-device AI capabilities, enabling faster and more efficient machine learning tasks. This improvement benefits consumers by delivering more advanced AI features like improved computational photography and real-time voice recognition, all while maintaining privacy and reducing reliance on cloud services. For the tech industry, it signals a shift towards more powerful, integrated AI hardware in smartphones, fostering innovation in AI-driven applications.

Key Takeaways

Joe Maring / Android Authority

Scouring the new Pixel 11 series for significant hardware changes this year, the biggest standout is the Tensor G6’s new and improved Tensor Processing Unit (TPU). Google says the Tensor G6’s TPU delivers 50% more compute power and up to 3.5x faster performance in some workloads. Impressive numbers, sure — but what does that actually mean when you pick up a Pixel 11?

For starters, let’s get to grips with what the TPU actually does. While Tensor’s CPU handles your apps and its GPU mostly handles games, the TPU is primarily tasked with number-crunching for machine learning workloads, featuring dedicated dot-product matrix math units and high-speed memory access. In other words, it’s the main AI processor inside Google’s Pixel handsets. It’s used for a wide variety of on-device tasks that don’t require the cloud, such as speech recognition, voice transcription, and text suggestions.

If you own a Pixel 10 Pro, are you upgrading to the Pixel 11 Pro? 6 votes Yes, I'm upgrading to the Pixel 11 Pro. 33 % No, I'm upgrading to another Pixel 11 device (elaborate in the comments). 17 % No, I'm upgrading to another Android phone (elaborate in the comments). 0 % No, I'm sticking with the Pixel 10 Pro. 50 %

Google’s TPU also covers a wider range of functions than just booting up Gemini Nano. The TPU is deeply integrated with the Tensor’s image processing pipeline, allowing the Pixel to leverage it as part of a heterogeneous system for computational photography tools like Night Sight, Super Res Zoom, portrait segmentation, video processing, and more. This year’s upgrade is one of the factors behind Google’s ability to extend long-range zoom out to 120x, up from 100x last year.

Bottom line: a more powerful TPU in this year’s Pixel 11 series means more horsepower to run increasingly demanding AI tools without the costs, latency, and privacy issues associated with cloud computing.

New on-device AI features for the Pixel 11

Joe Maring / Android Authority

As you’d expect, the Pixel 11 series has a few new AI tricks up its sleeve, powered by its more powerful processor. For starters, Google claims that Gemini Nano delivers higher-quality, lower-latency performance on its newer phones, resulting in more accurate, faster responses to small queries. This is arguably where Pixel 11 owners will feel the TPU upgrade most, not necessarily in 50% faster responses, but in more capable, powerful use cases.

The phones also introduce new voice-to-text capabilities that automatically remove filler words like “um,” helping to form your spoken thoughts into more polished text. The Pixel Camera app now also supports automatic sign language to text translation. Furthermore, Magic Capture takes up to 500 pictures and uses Gemini to determine the best six suggestions — five photos and one video — for you to keep. Live Translate can translate videos, podcasts, and voice messages into your chosen languages as you watch and listen.

... continue reading