An important thing to realize about proof-of-work is that at the moment, it's constrained by availability of hardware: both SHA256² ASICS (Bitcoin) and GPUs (ot

An important thing to realize about proof-of-work is that at the moment, it's constrained by availability of hardware: both SHA256² ASICS (Bitcoin) and GPUs (other coins) are in short supply right now. Additionally, the quality of that hardware is still not that good, with frequent failures when used for PoW mining (though that has improved a lot). Historically, that's given miners in low labour cost jurisdictions a surprisingly big advantage. I've personally visited mining in China a few years back: it was amazing the degree to which the place was a death trap. Huge piles of flammable packaging just left in big piles, high voltage bus bars just sitting out in the open with just signs to "protect" workers from being electrocuted, etc. But as things calm down, PoW mining becomes dominated by the availability of cheap power and cheap cooling. Ideally negative - IE doing something useful with the waste heat (harder than it sounds as you're competing against heat pumps). And when rapid growth isn't as important, you can afford to spend the relatively small amount of time and effort making the equipment last much longer. Like all electronics, there's no fundamental reason why it can't last ~forever, and the extra up-front cost isn't that much if you're using the equipment for a few years. Finally, because renewables have an enormous range of production capacity between different weather conditions, we can expect a renewable only grid - especially one dominated by solar and wind - to have far more capacity than needed for much of the year unless storage becomes much cheaper than it is. That's why PoW mining using up that excess is plausible. But it remains to be seen if it can compete with stuff like long-term storage in the form of hydroelectric, hydrogen, etc.