Most big labs have large computing clusters and upgrade them from time to time. We've almost always needed huge computing power in the scientific domain, no?

Anyway these days I see a lot of industrial investment in middle size computing datacenters full of GPUs. Sure Google scale is not within reach but I'm sure there's room for scrapier algorithms and training methods to demonstrate feasibility and then paying a lump sum to AWS (or some ml-expert-for-cloud-training SME) for the final step.

Anyway, I thought the expensive part was data acquisition and labeling? I like the 'surrogate network' approach of learning a very-expensive-to-compute simulation or model, that doesn't need data, but the output of a costly simulation.