TL;DR: Post-merge, Ethereum will utilize approximately ~99.95% less energy.
Ethereum is set to finalize its shift to Proof-of-Stake in the coming months, bringing forth an array of enhancements long speculated. With the Beacon chain operational for several months now, we can start examining the figures. One particular aspect we are eager to analyze includes updated energy-use forecasts, marking the conclusion of expending a country’s worth of energy on consensus.
Currently, there are no definitive statistics regarding energy usage (nor details about the hardware employed), so the subsequent information represents a rough estimation of the energy usage of Ethereum’s future.
Since many individuals operate numerous validators, I’ve opted to adopt the quantity of unique addresses that have made deposits as a proxy for the number of servers currently active. While various stakers may have utilized multiple eth1 addresses, this largely offsets the numbers against those with duplicate configurations.
At the moment of writing, there exist 140,592 validators from 16,405 unique addresses. Clearly, this is heavily influenced by exchanges and staking platforms; consequently, excluding them results in an estimated 87,897 validators presumed to be staking from personal setups. As a sanity check, this infers that the average home staker operates 5.4 validators, which appears to be a plausible estimation.
Power Requirements
What quantity of power is needed to operate a beacon node (BN), 5.4 validator clients (VC), and an eth1 full node? Based on my personal configuration, it requires around 15 watts. Joe Clapis (a developer from Rocket Pool) recently operated 10 VCs, a Nimbus BN, and a Geth full node using a 10Ah USB battery bank for 10 hours, averaging 5W for the setup. Given that the typical staker likely doesn’t operate such an optimized configuration, let’s estimate a total of 100W.
Multiplying this by the previously mentioned 87k validators indicates that home stakers consume approximately ~1.64 megawatts. Estimating the power usage by custodial stakers is somewhat more complex, as they operate tens of thousands of validator clients with redundancies and backups.
To simplify matters, let’s presume they use 100W for every 5.5 validators. According to the staking infrastructure teams I’ve engaged with, this is a gross overestimation. The actual figure is likely something like 50 times less (and if you are part of a custodial staking team consuming above 5W/validator, feel free to reach out; I’d be glad to assist).
In total, a Proof-of-Stake Ethereum thus consumes around 2.62 megawatts. This level of consumption is not comparable to that of countries, provinces, or even cities, but rather resembles the energy needs of a small town (approximately 2100 American households).
For context, Proof-of-Work (PoW) consensus on Ethereum presently consumes energy equivalent to a medium-sized nation, a necessity to ensure the integrity of a PoW chain. As the term indicates, PoW generates consensus based on which fork has conducted the most “work.” There are two methods to enhance the rate of “work” being performed: boost mining hardware efficiency and utilize more hardware concurrently. To safeguard a chain from successful attempts at compromise, miners must perform “work” at a pace exceeding that of potential attackers. Since an attacker is likely to have similar resources, miners must maintain substantial quantities of efficient equipment operational to deter attackers from out-mining them, which inherently consumes a significant amount of power.
Under PoW, the price of ETH and the hash rate are positively interrelated. Hence, as the price rises, the energy consumed by the network likewise increases in equilibrium. Under Proof-of-Stake, when the price of ETH escalates, the network’s security improves as well (the value of the ETH at stake goes up), while energy requirements stay constant.
Some Comparisons
Digiconomist estimates indicate that Ethereum miners currently use 44.49 TWh annually, equivalent to a continuous consumption of 5.13 gigawatts. Consequently, PoS is about ~2000 times more energy efficient based on the conservative estimates mentioned earlier, representing a total energy consumption reduction of at least 99.95%.
If energy consumption per transaction is of particular interest to you, that translates to roughly ~35Wh/transaction (average ~60K gas/transaction) or around 20 minutes of television viewing. In contrast, Ethereum PoW consumes the same energy as a house for 2.8 days per transaction, while Bitcoin uses the equivalent of 38 house-days.
Looking Ahead
Although Ethereum continues with PoW for the time being, this situation will not persist much longer. In recent weeks, we have witnessed the introduction of the first testnets for The Merge, the term assigned to the moment Ethereum transitions from PoW to PoS. Numerous engineering teams are working diligently to ensure that The Merge occurs as swiftly as possible, without compromising safety.
Scaling solutions (such as rollups and sharding) will further aid in lowering the energy consumed per transaction by leveraging economies of scale.
Ethereum’s days of excessive power consumption are limited, and I hope this holds true for the broader industry as well.
Thanks to Joseph Schweitzer, Danny Ryan, Sacha Yves Saint-Leger, Dankrad Feist, and @phil_eth for their contributions.