this looks like it's for GCP. TPUs are used for most internal workloads. It's available externally but some of the papercuts and devex without the TPU/TF team h

this looks like it's for GCP. TPUs are used for most internal workloads. It's available externally but some of the papercuts and devex without the TPU/TF team helping you can be more painful than using Nvidia/CUDA