Published 13 February, 2018; last updated 16 February, 2018

Tensor Processing Units (TPUs) perform around 1 GFLOPS/$, when purchased as cloud computing.

In February 2018, Google Cloud Platform blog says their TPUs can perform up to 180 TFLOPS, and currently cost $6.50/hour.¹ This gives us $171,000 to rent one TPU continually for a roughly three year lifecycle² Which is 1.05 GFLOPS/$.

This service apparently began on February 12 2018.³ So this does not appear to be competitive with the cheapest GPUs, in terms of FLOPS/$, or even the cheapest cloud computing.

1. “Built with four custom ASICs, each Cloud TPU packs up to 180 teraflops of floating-point performance…”

   “Cloud TPUs are available in limited quantities today and usage is billed by the second at the rate of $6.50 USD / Cloud TPU / hour.”

   Google Cloud Platform Blog, https://cloudplatform.googleblog.com/2018/02/Cloud-TPU-machine-learning-accelerators-now-available-in-beta.html [Accessed Feb 13 2018]

2. We do not know the lifecycle of TPUs, but usually assume a lifecycle of three years for converting per hour and per computer prices for computing hardware.
3. “Starting today, Cloud TPUs are available in beta on Google Cloud Platform (GCP) to help machine learning (ML) experts train and run their ML models more quickly.”

   Google Cloud Platform Blog, https://cloudplatform.googleblog.com/2018/02/Cloud-TPU-machine-learning-accelerators-now-available-in-beta.html [Dated Feb 12 2018, Accessed Feb 13 2018]