The energy question
Proof of work consumes electricity deliberately. The argument about whether that is acceptable is worth having properly rather than by slogan.
Proof of work secures a network by making block production expensive in electricity. The expenditure is not a side effect of an unoptimised system — it is the mechanism. Understanding that is necessary before the argument makes sense in either direction.
The case for concern
- Consumption is large in absolute terms, comparable to that of a mid-sized country on most estimates.
- Electricity used for mining is electricity not used elsewhere, or generated additionally, with whatever emissions that implies.
- Difficulty adjustment means the consumption scales with the reward rather than with the number of transactions, so efficiency improvements in hardware do not reduce total use — they are competed away. See mining.
- Mining hardware is single-purpose and becomes electronic waste when superseded.
The case in response
- Miners are uniquely mobile and price-sensitive, so they locate where power is cheapest — which is frequently where it is stranded, curtailed or otherwise wasted.
- This makes mining a buyer of last resort for energy that has no other customer, which can improve the economics of renewable projects with intermittent output.
- A meaningful share of mining uses low-carbon generation, although estimates vary widely and the methodology is contested on all sides.
- Comparisons should be against the alternative rather than against zero. The conventional financial system consumes substantial energy in branches, data centres, card networks and cash logistics, though direct comparison is genuinely difficult.
Estimating global mining consumption requires assumptions about the hardware mix and the electricity mix that nobody can observe directly. Ranges from credible sources differ substantially. Anyone quoting a precise figure with confidence — in either direction — is quoting an estimate as a measurement, and the underlying methodology usually acknowledges uncertainty the citation does not.
What proof of stake changed
This demonstrated that a large network can secure itself without the electricity expenditure, which reframed the debate considerably. It does not settle it: whether staking provides security equivalent to burning external energy is exactly the dispute between the two approaches, and reasonable people hold both positions. See the Merge.
What the argument is really about
Nobody disputes the electricity is consumed. The dispute is whether what it buys — a settlement network nobody controls — is worth it. That is a value judgement about the usefulness of censorship-resistant money, not a technical question, and it will not be resolved by better measurement.
Which is worth naming explicitly, because both sides tend to argue as though the other has simply not seen the data.