Hacker News Reader

Some workloads I can think of beyond crypto cases:

Distributed compilation of large code bases, offline rendering, baking lighting for AAA, media transcoding.

Basically, anything batch-style that doesn't demand reliable latency and can tolerate some amount of failure in the cluster.

Most simply it could be used for crypto mining (except the utility would have to upgrade the hardware very often)

The compute rig would need to be the size of the water tank to make the business even. Maintenance cost would be much higher too. Better to keep all compute in house sized DC units and just sell hot water to neighbors.

These guys [1] used to be called Nerdalize and iirc their first iteration looked a lot like this. I think they have since moved to a model of a few select locations where nodes are closer to each other.

[1] - https://www.leaf.cloud

EDIT: fixed link