Mining cryptocurrencies is not currently profitable using commodity clouds like AWS, Google Cloud, or Azure. But there's a breakeven point where it is (1) profitable on rented (cloud) commodity hardware and (2) profitable on consumer devices. What then?
I benchmarked a modified Ethereum miner on my new M1 Max. First, let's do a quick profit calculation.
Hashing Power: 10.33 MH/s
Wall Power: 120 W
Cost per KWh (California Average): $0.19
Pool fee: 1%
On average, we'd make about $0.79 / day, but pay $0.54 extra on our electricity bill. That's a profit of about $0.23 / day.
On the surface, "profitable," but once you consider depreciation and the upfront hardware costs – it's not. But imagine the high prices and shortages of GPUs over the years but applied to even more commodity consumer hardware.
Cloud services. Mining cryptocurrencies is already against the terms and conditions of nearly every cloud provider. It's a quick way to get permanently banned (which is why I didn't run a benchmark on AWS). But I could imagine miners will find a way to run their hashes on cloud servers as it becomes more lucrative.
Developing regions. Cloud can be prohibitively expensive on a global scale, so if prices are lowered for regions in developing countries, do those data centers become targets for cryptocurrency mining abuse?
If AWS doesn't do it, someone else might. Imagine it's extremely profitable to mine cryptocurrencies on commodity hardware. Cloud providers have strong barriers to entry because of the high fixed costs associated with building a data center. With a short enough payback period, a challenger could bootstrap their data centers by allowing cryptocurrency mining for a short while.