Measuring the Power Consumption after "the Merge"

Abstract

Blockchain networks have raised growing public concerns due to their substantial electricity consumption. The transition from Proof-of-Work (PoW) to Proof-of-Stake (PoS) on the Ethereum network is widely regarded as a landmark event in reducing blockchain energy use, with prior studies commonly reporting energy savings exceeding 99%. However, existing estimates vary substantially because of the strong assumptions embedded in the dominant top-down and bottom-up approaches. The top-down approach assumes that miners' electricity costs are closely tied to mining revenue under market equilibrium, whereas the bottom-up approach relies on the assumed average efficiency of the mining fleet, which is unobservable and highly sensitive to assumptions regarding hardware composition and utilization. "The Merge'' provides an observable profitability-based sorting mechanism that helps identify the efficiency distribution of mining hardware. By observing which miners could profitably migrate to Ethereum Classic after "The Merge'', we infer the efficiency threshold of economically viable machines and reconstruct the pre-Merge mining fleet more realistically. Using this framework, we estimate Ethereum's pre-Merge PoW electricity demand at 2.98 GW. The Ethereum Classic midpoint residual post-Merge PoW demand of 0.099 GW implies net electricity savings of 96.67%; including the broader Ethash-family residual yields savings of approximately 93.7-96.3%. To further investigate the determinants of estimation divergence, we estimate a VAR model and find that fluctuations in Ethereum prices significantly affect mining equilibrium and implied energy consumption. Overall, the paper provides a transparent, behaviorally grounded framework for estimating blockchain electricity use and offers refined evidence on the energy implications of consensus-mechanism design.