Should aid organizations accept ETH donations?

If India's COVID-19 relief fund is similarly impactful to GiveWell's Maximum Impact Fund, it can use $2 million to give people thousands of total years of healthy life (or "save" hundreds of lives). 

It's hard to imagine that the risk of losses to crypto investors has much ethical importance relative to those years — everyone who buys or sells ETH should understand that they are willingly taking on risk.

As for electricity: Via this calculator, it seems like $2 million in ETH requires roughly 60,000 kWh of electricity to mine. (The transfer takes roughly 100 additional kWh.) For comparison, a single intercontinental flight requires ~12,000 kWh per person.

If Vitalik Buterin had access to lifesaving medicine for hundreds of people, but he had to fly across the globe with four other people to deliver it, to what extent would it be unethical for him to do so?

The situation is more complicated than that — most instances of crypto use, especially high-profile ones, presumably make it more likely that others will use crypto —but I don't think it's so much more complicated that it becomes unethical for the relief fund to accept the money.

*****

$2 million is an unusually large crypto donation. Might it be unethical to accept donations below a certain size, since the transfer cost is constant?

I'd leave it up to the charity in question to determine how much good they can do with a given donation, and how to weigh that against the environmental cost. A blanket answer of "no, it's always unethical" seems very unlikely to be right.

*****

That's without getting into the question of what counts as an "ethical" source of money. Many people would argue that Bill Gates's donations are unethical because of his business practices, or that anyone who earns-to-give as an investment banker is doing something unethical. But because of those donors, millions of people are alive who otherwise would have died; ethical tradeoffs are part of life, and these seem like good trades to make.

While Ethereum currently uses proof-of-work they are working on transitioning to proof-of-stake, so I would book the electricity it currently uses as growing pains of a new technology. 

So far, the only real "application" of Ethereum is perhaps speculation.

Ethereum currently has real world usages such as https://kleros.io/ that does neutral conflict arbitration. 

While Ethereum's current prediction markets aren't very good at providing probability for events they will mature and provide information about events that's valuable.

Once Ethereum switches to proof-of-stake the transaction costs will also go down and projects that currently aren't viable will become viable.