Generally available: Compute Engine supports the Google's custom-developed accelerator Tensor Processing Unit (TPU),...

Generally available: Compute Engine supports the Google's custom-developed accelerator Tensor Processing Unit (TPU), providing a converged experience across AI accelerators on Google Cloud. You can use the Compute Engine instance API and managed instance group (MIG) API to create and manage TPU VMs. You can perform standard VM configurations such as using a custom OS or configure boot disk size. Compute Engine APIs support the creation and management of TPU slices across all consumption options, enabling small-scale experimentation and large-scale training and inference workloads.

For more information, see TPU resources in Compute Engine.