Ethereum Energy Use Plunges 99.9% After The Merge, Cambridge Research Confirms
New research from the Cambridge Centre for Alternative Finance (CCAF) shows that Ethereum’s energy consumption has fallen by more than 99.9% since The Merge, with the network now using approximately 7.87 GWh of electricity annually. The figure represents a dramatic decline from Ethereum’s final pre-Merge baseline of about 2.4 gigawatts of continuous power.
The CCAF report, produced by the Cambridge Judge Business School, also estimated Ethereum’s yearly emissions at about 2.37 ktCO₂e, a roughly 99.98% decline from the network’s final Proof-of-Work baseline. The findings provide the most comprehensive third-party audit of Ethereum’s post-Merge environmental footprint to date.
Also read: Ethereum Open Interest on Binance Hits Two-Year Low as OKX Spot Volume Surges
How The Merge Changed Ethereum’s Energy Profile
The Merge, completed in September 2022, transitioned Ethereum from a Proof-of-Work (PoW) consensus mechanism to Proof-of-Stake (PoS). Under PoW, miners competed using specialized computing hardware, consuming vast amounts of electricity. PoS replaces miners with validators who secure the network by staking ETH, dramatically reducing energy demand.
Cambridge reports that Ethereum now draws about 0.90 megawatts of continuous power, compared to approximately 2.4 gigawatts before The Merge. To contextualize the current scale, the report notes that Ethereum’s 7.87 GWh annual demand is less than half the British Museum’s reported 16.18 GWh yearly electricity use.
Also read: Ethereum Faces $1,800 Wall as 4.3M ETH Holds Key to Next Major Price Breakout
Cambridge’s Bottom-Up Audit Methodology
The CCAF based its estimates on real network infrastructure rather than theoretical models. The audit covered approximately 8,522 nodes — computers that maintain copies of the blockchain and relay data — across the Ethereum network. Researchers used wall-plug measurements from 20 different client combinations and hosting data to calculate power draw.
The report distinguishes between nodes and Ethereum’s roughly 894,000 validators. While nodes process data, validators provide economic security through staked ETH. Cambridge estimated average power use at about 105 watts per node, with residential setups averaging a median of 18 watts.
Grid Energy Now Shapes Ethereum’s Carbon Footprint
With mining eliminated, the carbon intensity of local power grids now determines most of Ethereum’s remaining environmental impact. The report found that the United States hosts 31% of Ethereum full nodes, followed by Germany (16%), Finland (8%), and France (6%). Together, these four countries host about 62% of all full nodes.
Cambridge reported that 36% of nodes run on residential hardware, while the remaining 64% use cloud or enterprise data centers. Hetzner, Amazon Web Services, and OVH collectively host approximately 40% of all nodes. The research found that 56.4% of Ethereum’s electricity comes from sustainable sources, with renewables supplying 39.4% and nuclear power providing 17.0%. Natural gas remains the largest single source at 27.7%.
The findings underscore that node location — and the local grid’s energy mix — now matters more than the consensus system in determining Ethereum’s carbon footprint.
Frequently Asked Questions
How much energy does Ethereum use now?
Ethereum now uses about 7.87 GWh of electricity per year, which is less than half the energy used by the British Museum.
Why did Ethereum’s energy use drop so much?
The drop is due to The Merge in 2022, which transitioned Ethereum from Proof-of-Work (mining) to Proof-of-Stake (validators), eliminating the need for energy-intensive computational mining.
What is the Cambridge Centre for Alternative Finance?
The Cambridge Centre for Alternative Finance (CCAF) is a research institute at Cambridge Judge Business School that studies emerging financial technologies and their environmental impact.
What are Ethereum’s carbon emissions now?
Cambridge estimates Ethereum’s annual emissions at about 2.37 ktCO₂e, a 99.98% drop from the network’s final Proof-of-Work baseline.
