The Merge: A Week Later, Visualized by Data

Introduction

The Merge finally arrived at 2:43 am EST on September 15th, 2022 without a hitch. While the average user hasn’t noticed much, a great deal has changed behind the scenes. My last article explains what the Merge is, as well as some of these changes including the network’s energy usage and ether’s supply growth. However, one can’t easily view these changes. To visualize the impact of the Merge, it’s better to let the data speak for itself. In this article, I will explore and visualize how Ethereum’s energy usage, client diversity, block production, and supply schedule have changed.

Energy Usage

The Bottom Line

The most significant change that has been captured by the media after switching from Proof of Work (PoW) to Proof of Stake (PoS) is Ethereum’s heavily reduced energy usage. According to the Crypto Carbon Ratings Institute’s detailed report, this transition has reduced Ethereum’s energy consumption by 99.98%. To visualize this difference in size we can compare something like the Eiffel Tower to a LEGO figurine. Additionally, Ethereum’s annual CO2 emissions have reduced from 11 million tons to just 870 tons. This is equivalent to that of under 100 United States households.

Energy Usage Estimation: PoW

There are a few different ways to estimate energy usage since it is unrealistic to measure the direct output of each node. To estimate energy usage under Proof of Work, one can use a top-down method or bottom-up method. A top-down method looks at the total income of miners and estimates the share of that income spent on electricity. This method is illustrated by the yellow line on the graph. A bottom-up approach looks at the total hash rate, the amount of each type of hardware used, and then sums up the average energy usage of all devices considered. This has been found to be a more reasonable estimation method and is indicated by the orange and grey lines on the graph.

Energy Usage Estimation: PoS

The energy usage of Proof of Stake is estimated by measuring the hardware requirements of each client, estimating each client’s energy usage, then weighting the usage by the distribution of clients among the total number of nodes in the network. CO2 emissions can then be estimated by using this weighted data and multiplying it by a carbon intensity factor that is regionally adjusted by the distribution of network nodes. Below, tables illustrate the calculation of the PoS energy usage and a comparison between PoW and PoS. Both are annualized using data from August and September 2022. This impact is so significant, it has been estimated by Ethereum researchers such as Justin Drake to have reduced the total global energy consumption by 0.2%, according to Business Insider.

Increased Client Diversity

Another effect of the Merge is increased client diversity. A client, according to Wikipedia, is “a computer or program that, as part of its operation, relies on sending a request to another program or computer hardware or software that accesses a service made available by a server”. Simply put, they are the different software types ran by Ethereum nodes to validate transactions. Some networks, like Bitcoin, only use one client. Ethereum uses several, and it is up to each node operator to choose which to use. Under PoW, Ethereum’s client usage was dominated by Geth, with over 80% of the nodes using it. Under PoS, client diversity has heavily increased, meaning more nodes are using different types of clients, with Prysm leading at 44% of the network, followed by Lighthouse at 34% of the network. While this distribution is much more favorable, it is still not ideal, and the hope is that node operators will spread to other clients as they are released.

Client diversity makes networks more resilient to bugs and attacks. Bugs usually only affect one client and can cause validators to go down. A severe bug could cause a large number of validators to be removed from the validator set, which could make the network more vulnerable to an attack, resulting in the activation of the Inactivity Leak mechanism, or an unrecoverable fork. If 1/3 or less of the ETH staked is affected, the Beacon Chain can still continue to finalize blocks, and affected validators have time to wait for the bug to be fixed or switch to another client. If greater than 1/3 of the ETH stake is affected, the Beacon Chain cannot be finalized, meaning that transactions aren’t guaranteed to be permanent and immutable. Validator penalties are also more severe, and the Inactivity Leak mechanism is activated. With this mechanism, the affected staked ETH is slowly burned until the affected ETH drops below the 1/3 threshold. Thus, the ideal scenario is to have client distribution not exceed 33% of nodes for any singular client. Ideal distributions compared to current ones are illustrated by this graphic from clientdiversity.org

More Consistent Block Times

Switching from the hashing race of PoW to the more organized queue format of PoS has resulted in much more consistent block times. Before, blocks were solved based on hash power, meaning that block times varied heavily — they could be almost instant due to pure luck or take several minutes. Now, the time separating validators in the queue is just 12 seconds. This means that we have a consistent interval between blocks, and if a validator misses a block, the time jumps to just 24 seconds. A validator may miss a block for a few reasons. This could be that the validator was down, a malicious block was produced that was not approved, or otherwise. This difference in block times is illustrated clearly in this graph produced by Coin Metrics. Post-Merge, the Ethereum network is much more predictable with consistent block times.

Supply Schedule

One of the most significant changes to the Ethereum network as a result of the Merge is the 90% reduction in ether emissions. Before the Merge, issuance was comprised of 1,600 ETH per day from the Beacon Chain consensus layer and 13,000 ETH per day from the PoW execution layer. Now that these chains have merged and the PoW chain is no longer issuing rewards, new ether issuance is solely 1,600 ETH per day. This, along with EIP-1559’s burning function, means that daily net ether issuance is less than 1,600 ETH. This gives ether the potential to be a deflationary asset depending on network activity. If gas fees are above 15 gwei, then more ether per day is burned than is issued. According to ultrasound.money, ether has remained inflationary after the merge due to low network activity. However, ether was deflationary for the first day or so after the merge since gas fees were above 15 gwei. Since then, we have had a few spouts of deflation, as evident by the hourly chart, which illustrates the number of blocks which were deflationary and the base fees burned associated with each block.

Since the Merge, ether supply has grown by 4,700 ETH, putting us at an average inflation rate of about 0.19% per year. If we were still running PoW, the supply since the date of the merge would have increased by 93.3k ether. This is a significant difference, as a PoW Ethereum would have resulted in a 3.78% inflation per year since the merge. This graph demonstrates the difference between PoS Ethereum, PoW Ethereum, and Bitcoin’s inflation rates. Since PoS is less resource intensive than PoW, validators require just a fraction of the compensation that miners require in order to remain profitable. This allows ether to have a low issuance, and this combined with EIP-1559 allows for deflationary properties and the ultrasound money meme to propagate. Currently, the applications which contribute to the highest ETH burn rates are OpenSea at 600 ETH, Uniswap at 516 ETH, and ether transfers at 434 ETH since the Merge.

Conclusion

The Merge is one of the most significant events to date in the blockchain space, which is ironic since its changes aren’t noticed by the average user. Even if one was using the network during the Merge itself, one wouldn’t notice anything, besides a massive spike in fees due to a large number of people trying to be included in the first PoS block. In fact, the validator of the first PoS block collected over 45 ETH in fees. Apart from this, looking at data is the best way to visualize these momentous changes in the Ethereum network. While these changes aren’t visible now, their impacts will certainly show up over the years in Ethereum’s energy usage, client diversity, block times, and supply schedule.