Google launches two new AI chips to cut Nvidia reliance

Google unveils two new AI chips

Google on Monday unveiled TPU 8tx and TPU 8i Max, new variants of its tensor processing units designed for AI workloads that broaden its chip portfolio to additional use cases.

TPU 8tx: security built into hardware

TPU 8tx, designed with AMD, is based on the TPU 8t chip and is geared specifically for confidentially training and running AI models. The chip builds security directly into the hardware layer.

“builds security directly into the hardware layer, keeping sensitive data and model weights protected and isolated even while in active use”

TPU 8tx enables on-device model training

The chip has a custom inferencing network that lets it run reasoning alongside training, so it can be used to distill and train small models on the fly without moving data between devices.

TPU 8i Max doubles memory and power

TPU 8i Max combines two inference-focused TPU 8i chips launched in April, doubling the memory capacity, processing power and bandwidth in a single package. It lets AI models handle larger information ranges during queries.

Part of eighth generation launched in April

The new units are part of Google's eighth generation of chips introduced in April this year and build on what those chips offered: more inference capacity and support for large training requirements.

Hyperscalers race to reduce Nvidia dependence

The chip launch comes as Microsoft, Amazon, OpenAI and Anthropic work on their own AI-specific semiconductors and accompanying software to reduce reliance on Nvidia's dominance in the market.