Cloud TPU v4 is now generally available! To start using Cloud TPU v4 Pods for your AI workloads, please fill in this form

Cloud TPU

Speed up machine learning workloads with Google’s custom-developed hardware accelerators.

View documentation for this product.

Accelerate machine learning models with Google supercomputers

Cloud TPU empowers businesses to speed up their machine learning models, including Natural Language Processing (NLP), ranking and recommendations, and computer vision. Developers, researchers, and businesses can tap into the same custom-designed machine learning ASICs (application-specific integrated circuits) that power Google’s Search, YouTube, and LaMDA AI model. 

Quickly train and iterate on machine learning models

Cloud TPU minimizes the time-to-accuracy when you train large, complex neural network models. Models that would have taken weeks to train on other hardware. 

Handle large-scale workloads with flexibility

TPUs were specifically designed for models with matrix computations and large models with large effective batch sizes. TPU VMs make it easy to employ popular ML frameworks

Reduce costs and increase sustainability

Cloud TPU provides low-cost performance per dollar at scale for various ML workloads. And now Cloud TPU v4  gives customers 2.2x and ~1.4x more peak FLOPs per dollar vs Cloud TPU v3.

Cloud TPUs for every workload and budget

Cloud TPU is designed to run cutting-edge machine learning models with AI services on Google Cloud. And its custom high-speed network offers over 100 petaflops of performance in a single pod—enough computational power to transform your business or create the next research breakthrough.

Full backwards compatibility

Cloud TPU v4 Pods are the latest generation of Google’s custom ML accelerators and are now available in GA. It retains backwards compatibility with Cloud TPU v2 and v3, but has a >2x increase over Cloud TPU v3 in raw compute performance per chip. Each TPU v4 chip also contains a single logical core, enabling utilization of a full 32 GiB of memory from one program, compared to 8 GiB on v2 and 16 GiB on v3. Learn which of Cloud TPU products works best for your unique project needs.

Run machine learning workloads on Cloud TPUs using machine learning frameworks such as TensorFlowPytorch, and JAX. Our quickstarts provides a brief introduction to working with Cloud TPU VMs and explains how to install an ML framework and run a sample application on a Cloud TPU VM.

Save money by using preemptible Cloud TPUs

You can save money by using preemptible Cloud TPUs for fault-tolerant machine learning workloads, such as long training runs with checkpointing or batch prediction on large datasets. Preemptible Cloud TPUs are 70% cheaper than on-demand instances, making everything from your first experiments to large-scale hyperparameter searches more affordable than ever. Visit our pricing page to get a sense of how Cloud TPU can process your machine learning workloads in a cost-effective manner.

Learn how TPU v4 has enabled our customers

Take the next step

Start building on Google Cloud with $300 in free credits and 20+ always free products.

Need help getting started?
Work with a trusted partner
Continue browsing

Take the next step

Start your next project, explore interactive tutorials, and manage your account.

Need help getting started?
Work with a trusted partner
Get tips & best practices

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