Cambridge Study: Ethereum Merge Cut Energy Use 99.9%

Cambridge update: Ethereum’s Merge slashes electricity use

The Cambridge Centre for Alternative Finance (CCAF) has released an updated estimate showing that Ethereum’s transition from proof-of-work to proof-of-stake — the Merge on September 15, 2022 — reduced the network’s electricity demand by more than 99.9%. The study places Ethereum’s current annual electricity consumption at about 7.87 gigawatt-hours (GWh), equivalent to a continuous load of approximately 0.90 megawatts (MW), and estimates annual greenhouse gas emissions of roughly 2.37 kilotonnes CO2e. These figures reflect direct node measurements and a market-value-adjusted energy intensity comparison across major proof-of-stake blockchains.

Key findings at a glance

• Annual electricity consumption: 7.87 GWh (≈0.90 MW continuous)
• Estimated emissions: ~2.37 kt CO2e per year
• Discoverable full nodes measured: 8,522
• Network-weighted average power per node: ~105 watts
• Renewable share of supply: 39.4%; nuclear: 17%; fossil fuels: 43.6% (natural gas 27.7%)

Methodology: direct node tests and real-world hosting data

A distinguishing element of the CCAF analysis is its reliance on direct power measurements instead of a single assumed wattage for every node. Researchers ran 20 combinations of Ethereum execution and consensus clients across two representative hardware configurations. The lighter residential setup drew a median of 18 watts, while a professional workstation used for enterprise-style deployments consumed about 152 watts. Weighting those results against the observed node population produced an average of roughly 105 watts per node.

Cambridge found 8,522 discoverable full nodes. About 36% of those ran on residential hardware and 64% operated in cloud or enterprise data centers. Node geography was concentrated: the United States hosted 31% of nodes, Germany 16%, Finland 8%, and France 6% — these four countries together accounted for nearly 62% of the measured full-node network.

Energy mix now determines emissions

With consensus secured by validators rather than energy-intensive miners, Cambridge notes that electricity is no longer the primary "price of security" for Ethereum. Instead, grid generation sources now shape most of the network’s remaining carbon footprint. The study estimates that renewable sources supplied 39.4% of Ethereum’s energy, nuclear 17%, and fossil fuels the remaining 43.6% — dominated by natural gas at 27.7%. The final emissions tally depends on the specific electricity grids that serve each node.

Alexander Neumüller, research lead at Cambridge’s digital assets energy program, summarized the shift: "Under Proof-of-Stake, electricity is no longer the price of security." The report corroborates that continuous power demand fell from around 2.4 GW pre-Merge to about 0.90 MW after the transition.

How Ethereum compares to other proof-of-stake networks

In absolute energy terms, Ethereum still consumes more electricity than most peers in Cambridge’s top-tier proof-of-stake comparison. Solana ranked highest at roughly 13.48 GWh per year, with Ethereum second. However, when adjusting energy use for market value, Ethereum performed near the low end of energy intensity among major PoS chains. Cambridge found Ethereum used about 33 kWh per $1 million of market capitalization — the second-lowest rate in the comparison, behind BNB Chain. Solana’s adjusted figure was about 283 kWh per $1 million, roughly 8.5 times Ethereum’s rate.

Across the top-tier PoS networks included in the analysis, combined energy use was about 38 GWh. Other networks ranged: NEAR, Tron, and TON each consumed approximately 3.6–5.1 GWh annually, while Cardano and BNB Chain remained below 1 GWh.

Why Cambridge avoided per-transaction estimates

The report deliberately did not provide a per-transaction energy figure for Ethereum’s mainnet. Cambridge notes that about 92% of transactions in the Ethereum ecosystem now settle on scaling networks (layer-2 solutions and rollups), making a mainnet-only per-transaction calculation incomplete and potentially misleading. The study also highlights that technologies like stateless light verification could reduce hardware requirements, even as broader node participation might offset such efficiency gains.

Implications for crypto energy reporting and policy

The CCAF update offers a more granular, hardware- and hosting-informed snapshot of Ethereum’s post-Merge energy profile. For policymakers, exchanges, and institutional crypto users assessing blockchain carbon footprints, the report underscores two points: (1) consensus mechanism matters — PoS drastically reduces baseline electricity demand versus PoW — and (2) emissions depend heavily on the electricity grid mix supporting node infrastructure. While Ethereum’s absolute electricity use remains higher than many peers, its market-value-adjusted energy intensity and the dramatic drop after the Merge strengthen claims that proof-of-stake is a substantially lower-energy model for securing public blockchains.

The Cambridge findings update earlier energy narratives and provide a practical dataset for ongoing assessments of blockchain sustainability, validator hosting practices, and the evolving relationship between cryptocurrency networks and decarbonization efforts.