Cambridge study: Ethereum’s Merge slashed electricity use by more than 99.9%, but the network still uses more power in absolute terms than many PoS peers
A new estimate from the Cambridge Centre for Alternative Finance finds that Ethereum’s shift from proof-of-work to proof-of-stake (the Merge, Sept. 15, 2022) cut its electricity demand by an almost unimaginable amount — from roughly 2.4 gigawatts to about 0.90 megawatts continuous power. That’s a reduction of over 99.9%.
Key findings
- Annual electricity use: 7.87 GWh (about 0.90 MW continuous).
- Estimated annual emissions: ~2.37 kilotonnes CO2e.
- Network-weighted average power draw per full node: ~105 watts (based on direct measurements).
- Discoverable full nodes: 8,522 (36% on residential hardware, 64% in cloud/enterprise datacenters).
- Node geography: United States 31%, Germany 16%, Finland 8%, France 6% — those four countries account for nearly 62% of discoverable nodes.
How Cambridge measured it
Rather than applying a single assumed figure per node, researchers directly measured power use across 20 combinations of Ethereum execution and consensus clients on two hardware profiles. A lighter residential setup drew a median 18 W; a professional workstation-style node used about 152 W. These measured values were combined with the observed node population to produce the ~105 W network-weighted average.
Energy intensity vs. absolute use
Although Ethereum’s total annual consumption (7.87 GWh) was higher than many PoS rivals, the network performs well when energy use is adjusted for market value. Cambridge calculated roughly 33 kWh of electricity per $1 million of market capitalization — the second-lowest rate in their top-tier PoS comparison (BNB Chain ranked lowest). By contrast, Solana used an estimated 13.48 GWh per year and about 283 kWh per $1M, roughly 8.5x Ethereum’s intensity. The PoS networks in Cambridge’s top-tier comparison used about 38 GWh combined.
Energy mix and emissions
Cambridge’s emissions estimate depends on the electricity grids tied to each node location. They estimate 39.4% of Ethereum’s power comes from renewables and 17% from nuclear, so non-fossil sources make up about 56.4% of supply. Fossil fuels contribute the remaining 43.6%, with natural gas the largest single fossil source at 27.7%.
Context and caveats
Cambridge’s report underscores the dramatic efficiency gain from moving to proof-of-stake. “Under Proof-of-Stake, electricity is no longer the price of security,” said Alexander Neumüller, research lead at Cambridge’s digital assets energy program. The paper does not claim Ethereum is the lowest-energy PoS network in absolute terms — its annual total remains higher than most peers studied — but its per-market-value energy intensity is among the lowest.
The researchers also avoided a per-transaction energy figure, noting that roughly 92% of activity in the Ethereum ecosystem now settles on scaling networks rather than the mainnet, making a mainnet-only per-transaction metric misleading. They add that protocol changes such as lighter stateless verification could lower individual node requirements, but greater node participation could offset those hardware savings; the study treats future demand as uncertain rather than assuming efficiency gains will automatically reduce total consumption.
These updated hardware, hosting and energy-mix data give one of the clearest pictures yet of Ethereum’s post-Merge footprint, confirming the enormous drop in electricity demand while highlighting how geography and grid mixes now drive most of the protocol’s remaining emissions.
Read more AI-generated news on: undefined/news
