Cambridge: Ethereum PoS energy consumption falls below the lower limit of forecast

University of Cambridge: The combined energy consumption of Ethereum is lower than low-end expectations

A new report released by the University of Cambridge's Center for Alternative Finance points out that the combined energy footprint of Ethereum is among the lower-energy-consuming category among proof-of-stake networks, although its absolute consumption is still higher than that of most single network peers. The study estimates that Ethereum consumes approximately 7.87 GWh of electricity per year. More importantly, for sustainability comparisons, the University of Cambridge also calculated the energy intensity of Ethereum relative to market value and found that every $1 million in market value corresponds to approximately 33 kilowatt-hours. This value ranks second lowest among the proof-of-stake networks evaluated, second only to Binance Intelligent Chain.

Key Points

The University of Cambridge estimates that Ethereum consumes approximately 7.87 GWh per year based on a wall-end power model. Ethereum's energy intensity is estimated to be approximately 33 kilowatt hours per $1 million, ranking the second lowest among the proof-of-stake networks studied. Solana is believed to have the highest electricity consumption among the proof-of-stake networks considered, at approximately 13.48 GWh per year. The University of Cambridge attributed Ethereum's remaining emissions mainly to the power grid that powers its validators and node operators, rather than to mining-based network operations.

How Cambridge University builds an Ethereum power model

Researchers at the University of Cambridge estimated electricity consumption by measuring the power consumption of Ethereum nodes under different software client configurations. According to the report, they simulated the power needs of 20 combinations of the Ethereum network's main software clients and used assumptions related to typical home and professional hosting environments. In the study's model, a "typical home" node configuration consumes approximately 18 watts of power, while a more powerful workstation configuration consumes approximately 153 watts. Subsequently, the University of Cambridge used the mixing ratio of residential and professional hosting nodes observed by Ethereum to estimate that the average power consumption per node was approximately 105 watts. To extend this data to network-wide consumption, the report counted approximately 8522 discoverable full nodes, 64% of which were running in cloud or enterprise environments, and the remaining 36% used residential network connections. These ratios are important for sustainability because hosting environments can differ in terms of power sources and energy efficiency.

Energy intensity and market value: Why change the comparison criterion

Although absolute electricity consumption is a dimension of sustainability, Cambridge University also calculates energy intensity adjusted for market value. This approach aims to answer a practical question for investors and policymakers: how much electricity the network uses relative to its economic value. Based on this criterion, the estimated market value of Ethereum of approximately 33 kilowatt hours per US$1 million is at the lowest level among the proof-of-stake systems evaluated, second only to Binance Smart Chain. In contrast, Solana ranks first in the Cambridge University data set in terms of both annual absolute power consumption and energy intensity. The report estimates that Solana uses approximately 13.48 gigawatt hours per year, with a market value adjusted energy intensity of approximately 283 kilowatt hours per $1 million. The University of Cambridge points out that this is about 8.5 times that of Ethereum, and the combined consumption of the networks in comparison is about 38 gigawatt hours. For readers, this means that rankings may vary depending on whether the measure is raw consumption or market capital-adjusted intensity. Cambridge University's dual perspective suggests that even if both indicators ultimately reflect actual electricity consumption, sustainability arguments based solely on electricity may ignore the important dimension of "efficiency per dollar market value."

Impact of post-merger transition on emissions

The framework of this research is closely related to Ethereum's process from proof-of-work to proof-of-stake. Ethereum transitioned from mining-based verification to pledge-based verification through a September 2022 merger. After this change, the energy-intensive mining process that used dedicated hardware to compete for block rewards was removed and replaced by pledging ether to protect network verifiers. The University of Cambridge report builds on previous estimates that the merger has significantly reduced Ethereum's electricity consumption. After the upgrade, network power estimates showed a decrease of more than 99.9%, reflecting the elimination of proof-of-work operations. As mining disappears, the University of Cambridge says Ethereum's remaining emissions are now mainly driven by the power grid that powers its nodes and validators. The report estimates that approximately 56.4% of the Ethereum power portfolio comes from renewable energy and nuclear energy, with the remaining 43.6% coming from fossil fuels. This distinction is crucial for the future. Even if energy consumption remains relatively stable, emissions may improve or deteriorate as the node's area changes power generation methods, or companies and cloud operators shift to cleaner sources of electricity.

What these numbers mean for investors and policymakers

Cambridge University described the work as one of the most detailed assessments to date of the combined energy footprint of Ethereum. This is important because Ethereum is often used as a case study in the broader sustainability debate surrounding the proof-of-stake blockchain-especially when regulators and institutional participants are weighing environmental and operational factors while also considering financial risks. By estimating node-level power consumption, scaling based on the number of nodes that can be observed, and using a consistent methodology to compare results from different proof-of-stake ecosystems, the report provides stakeholders with a more current basis for evaluating blockchain energy use claims after major agreement changes. Still, the study's methodology also hints at areas where uncertainty may exist. The results depend on assumptions about node power consumption between client combinations and hosting types, and how the number of nodes can be found to represent the broader network. The direction of change after the merger is clear, but the refinement of the current benchmark will continue to depend on the evolution of node operations. Looking ahead, as node distribution, hosting practices, and grid energy mix change over time, readers should pay attention to updated third-party assessments-especially under proof of stake, as the sustainability debate increasingly shifts from "mining infrastructure" to the sources of electricity that power verifiers and node operators.