# Study says useful proof-of-work could raise electricity use

Published 2026-11-11 · Crypto.world Team · Crypto.world

> A preprint finds that replacing conventional proof-of-work with useful computation could increase total electricity consumption under certain market conditions. The result depends on assumptions about mining rewards, demand and the cost of producing proofs.

Replacing conventional cryptocurrency mining with useful computation could increase total electricity consumption under certain market conditions, a new research preprint finds. The paper by Isaiah Hull and Or Sattath, submitted to [arXiv](https://arxiv.org/abs/2611.13131) on Nov. 9, examines how demand for computing work might respond when mining rewards lower its effective price.

Proof-of-useful-work has been proposed as an alternative in which mining performs computations with uses beyond the mining process. The paper uses matrix multiplication as a concrete example of such work.

## Demand can offset savings

The authors model a cryptocurrency before and after a switch from proof-of-work to proof-of-useful-work, keeping protocol parameters, including the mining reward, fixed. In their model, the reward becomes a subsidy for producers of useful work, lowering the price paid by customers.

That lower price can encourage customers to demand more computation. A sufficiently strong response can cause the electricity used for the additional work to exceed the electricity saved by ending conventional hashing, the researchers find.

The paper describes demand as elastic when a percentage discount produces a larger percentage increase in the quantity demanded. When demand is inelastic, the modeled electricity saving is partly offset by additional useful work, rather than eliminating the consumption associated with the mining reward.

In a simplified example with a linear demand curve, total electricity consumption rises by 20% after the transition, even though proof-of-work hashing is eliminated. That example is a model of a possible outcome, not a measurement of an operating network.

## Conditions behind the result

The main model assumes that costs are spent on electricity, that proof-of-useful-work adds no overhead to the useful computation and that the mining subsidy is unchanged. The paper also examines less restrictive settings, including cases where only a fraction of costs are electricity costs.

In its linear-demand case, the threshold for an increase depends on both demand elasticity and the size of the mining reward, rather than elasticity alone. The paper also examines more general demand curves and finds an increase with elastic demand when the mining reward is sufficiently small relative to the useful-work market.

The researchers say they are not aware of an empirical estimate specifically for demand elasticity for matrix multiplication. They discuss cloud-demand studies as possible historical comparisons, while noting that those studies do not directly identify matrix-multiplication demand elasticity.

## Proof costs and public goods

The paper also considers the extra computation needed to make useful work verifiable. In the matrix-multiplication construction discussed by the authors, that can include perturbation and hashing steps.

If the reward is too small to cover proof overhead for every producer in the modeled market, only some producers make proofs, the analysis finds. Competition then keeps the useful resource's price unchanged, customers receive no discount, and total electricity consumption remains at its pre-transition level.

The main analysis applies most directly to useful work sold as a privately demanded market good. For work useful only as a public good, the authors say their induced-demand mechanism does not apply because there is no consumer price reduction that prompts additional private demand.

## Sources

1. Original source · Preprint [arXiv](https://arxiv.org/abs/2611.13131) arXiv:2611.13131

## Related stories

- [Researchers Model How Capital Costs Delay Bitcoin Mining Gaps](https://crypto.world/news/researchers-model-capital-costs-bitcoin-mining-gaps)
- [Researchers propose splitting Bitcoin mining rewards in fee-only era](https://crypto.world/news/bitcoin-fee-only-mining-reward-split-proposal)
- [Solana study finds no clear early advantage for graph-based rug detection](https://crypto.world/news/solana-early-rug-pull-detection-study)
- [LlamaRisk proposes Aave V3 borrowing rate limits](https://crypto.world/news/aave-llamarisk-proposes-dynamic-borrow-caps)
- [Helius puts programmable privacy rings on Solana mainnet](https://crypto.world/news/helius-solana-privacy-rings-mainnet)

## Explore

[More articles](https://crypto.world/news)

---

Canonical: https://crypto.world/news/useful-proof-of-work-energy-use-study
