Google Cloud's 8th generation TPUs: The computing foundation for the agentic era

Accelerate the AI lifecycle with specialized architectures purpose-built for frontier-model training and real-time reasoning.
Connect with a Google Cloud specialist to learn more.

Request more information.

First name

Last name

Business email

Calling Code

Business phone

Job title

Company name

Sign me up to receive news, product updates, event information, and special offers about Google Cloud from Google.

I understand my personal data will be processed in accordance with Google's Privacy Policy.

The way we build and deploy AI is undergoing a profound shift. As models evolve from providing simple predictions to executing multi-step reasoning loops, the architectural requirements for training and inference have diverged sharply. Training demands massive compute throughput and scale-up bandwidth, while real-time inference requires massive memory bandwidth and ultra-low latency.

To lead in the agentic era, you cannot rely on one-size-fits-all hardware. Our 8th generation TPU family introduces two purpose-built architectures: TPU 8t for training and TPU 8i for inference. Hosted for the first time on our own Axion ARM-based processors, they provide a fully optimized, co-designed foundation to help your teams build what comes next.

Here is how we empower your teams to drive rapid innovation:

Performance without compromise: accelerate the AI lifecycle with infrastructure purpose-built for frontier-model training and real-time reinforcement learning for inference.
Sustainable economics at scale: deliver unrivaled price-performance through system-level co-design that optimizes the entire infrastructure stack.
Open, flexible, and portable operations: speed up development with familiar open-source frameworks and a portable ecosystem for global scaling.

Ready to scale your AI operations? Connect with our experts to build your future on Google Cloud’s 8th Generation TPUs.

Cloud AI products comply with our SLA policies. They may offer different latency or availability guarantees from other Google Cloud services.

Google Cloud's 8th generation TPUs: The computing foundation for the agentic era

Accelerate the AI lifecycle with specialized architectures purpose-built for frontier-model training and real-time reasoning. Connect with a Google Cloud specialist to learn more.

Request more information.

Accelerate the AI lifecycle with specialized architectures purpose-built for frontier-model training and real-time reasoning.
Connect with a Google Cloud specialist to learn more.