Independent of whether this is true (and I'm sceptical), Bitcoin mining requires computation (and lots of it). Computation requires power. Even if that power is renewable and would be difficult to use for things other than computation, there is still a huge demand for computation. Hence it's still a choice to spend that power on Bitcoin mining rather than other computation tasks such as simulations for scientific research.

Yes that’s a thing. The problem for cryptocurrency is there’s no free lunch, and to gain significant security of a public ledger you have to spend some resources to do so. Bitcoin spends electricity, Chia.Network is working to spend unused disk space instead which is an interesting approach. Others like Algorand, Dfinity, Ethereum 2.0 etc are working to hopefully prove that rule can be overcome with less resource-hungry sources of cryptographic randomness. Hopefully they’ll succeed, but overcoming that rule would be a significant technical accomplishment, and not a given.

> The problem for cryptocurrency is there’s no free lunch

There absolutely is a free lunch.

In renewable energy, there is a massive amount of "free" energy that is not being used for anything, and is literally being thrown away, because it can't be stored or used during the time of day that it is generate.

That energy is free. There is zero downsize to spending it on guessing number. It might even be negative value energy, actually.

Did you know that sometimes energy prices go into the negative, and people will pay you to use it? Something to do with the electric grid, and difficulty of destroying energy, I am not sure.

You're asserting that there's a huge demand for remote offline computation by scientists, but no such market exists as far as I can tell. Cryptocurrencies might end up creating such a market eventually, since they're researching solutions to the problems of verifiable and secure decentralized computation (e.g. Golem) and incentive for file storage (e.g. Filecoin).

Are you claiming that crypto-mining is pricing out useful scientific research? If so, can you provide evidence for that claim? Keep in mind the geographic nature of electricity generation and transmission.

I don't have a citation either, but you can empirically determine this to be likely by just comparing the reward per kW spent and noticing that it has become less than the cost per kW of non-renewable energy.

Update: better analysis here - https://hackernoon.com/the-reports-of-bitcoin-environmental-...