When you talked with customer beside their big idea, they may be has two basic concern. How much will it cost? Usually, cash is a limitation. Most customers want to figure out how much they'll have to spend. And know what they spends will turn into an even bigger pile How long will it take ? The other big constraint is time. No customer ever says "Take as long as you want" And lots of the time, you have a specific event or date the customer wants their software ready for.
Google excited to announce their second-generation Tensor Processing Units (TPUs) are coming to Google Cloud to accelerate a wide range of machine learning workloads, including both training and inference. We call them Cloud TPUs, and they will initially be available via Google Compute Engine. They’ve designed, built and deployed a family of Tensor Processing Units, or TPUs, to allow you to support larger and larger amounts of machine learning computation, first internally and now externally. resources: https://blog.google/topics/google-cloud/google-cloud-offer-tpus-machine-learning/