The energy criticism of proof-of-work mining used to be simple and quantitative: it consumes a lot of electricity, comparable to a mid-sized country. That number is still roughly right, and it has stopped being the interesting part of the argument.
What changed the conversation
Grid operators discovered that mining is unusually flexible demand. A data centre serving users cannot switch off. A mining operation can, within seconds, and does so when power prices rise.
That makes it useful in a way most large consumers are not. Operators can contract with miners to shed load during peak demand, effectively treating them as a reserve that pays for itself rather than sitting idle.
The stranded energy case
The stronger argument is about location. Generation exists in places where transmission does not – remote hydro, flared gas at oil sites, wind capacity that outruns the lines connecting it. That energy is currently wasted.
Mining goes where the power is, because the only thing it needs to export is a network connection. That turns wasted generation into revenue and, in the flared gas case, converts methane into a less harmful emission.
The counter-arguments
Flexibility is only valuable if it is contracted and enforced, and much of the industry operates without such agreements. Stranded energy is a real category, but a large share of mining runs on ordinary grid power in ordinary places, competing with everything else for it.
And the local effects are genuine. Communities near large operations report higher electricity costs and constant noise from cooling systems, which is a cost borne by people who receive none of the revenue.
What proof-of-stake changed
Ethereum’s move away from mining removed a large share of the sector’s energy use in a single step, which demonstrated the alternative works at scale. Bitcoin will not follow – proof-of-work is not a bug in its design but the whole security model – so the debate stays live for the largest network specifically.
Where it stands
Better than the crude version, worse than the industry’s defence. Mining can be a useful grid participant and often is not. The variation between operations is larger than the average, which makes any single figure misleading.