Google Deploys TPU To Google Cloud Service

[Guest]

The deployment is still in beta, but customers of Google’s cloud computing service can now use the company’s dedicated AI processing hardware for their own machine learning developments.

Google Deploys TPU To Google Cloud Service : Read more

 

[bit_user]

A single Cloud TPU offers up to 180 teraflops of floating-point computation power.

You quoted the wrong article. This (http://www.tomshardware.com/news/tpu-v2-google-machine-learning,35370.html) dives into considerably more depth, revealing that each TPU2 only delivers 45 TFLOPS. However, they can be arranged in quad-processor configurations, with each board providing 180 TFLOPS.

Access to Cloud TPUs start at the price of $6.5 / Cloud TPU / hour

I think this is about twice what you can find for a single Nvidia V100, which is also faster (110 TFLOPS/GPU).