Ethereum Fees: Sustainable Blockchain Revenue?

In our January 5, 2023 macro research report titled “The Case for Real World Adoption That Could Unlock a $242bn Revenue Opportunity in 5 Years,” we presented an important crypto theme for 2023: the need for blockchain to serve real world use cases.

Ethereum is one of the leading candidates to be the universal settlement layer for tokenized assets. And Ethereum’s security comes from its transaction fee revenue, which is a measure of network usage.

We believe blockchains need demand-driven fee revenue in order to be economically sustainable and secure. Too many blockchains and projects rely on token inflation to subsidize security, which is a long-term risk. Ethereum, via the Merge (to reduce issuance) and EIP-1559 (to burn ~80% of transaction fees), has a self-driving monetary policy.  

Although our 2023 theme will identify future revenue opportunities for the blockchain space, it is worth analyzing Ethereum fees throughout ETH’s history.

By comparing fees in Ethereum’s early days (2016-17), fees during ETH’s first bull run and subsequent crash (2018-2019), and fee revenues in the most recent cycle from 2020-2022, we can develop a framework for valuing ETH based on future fees.

We also raise the question of what happens to ETH fee revenue with the proliferation of L2s. Ultimately, we believe transaction fees are the only sustainable blockchain revenue to carry the space toward mainstream adoption.

Ethereum’s Present Day Revenue Snapshot

• Before diving into historical fee trends, let’s look at a snapshot of Ethereum today. We are both in a post-Merge, Proof of Stake world and in a bear market, meaning transaction fees and demand for blockspace are low relative to the mania we saw in 2021 (examined later).
• If we assume a conservative 750k ETH in annualized transaction fees (a sharp dropoff from 2020-22 levels), that equates to ~$1.125bn in blockchain revenue (of which 80% is burned rather than being paid to validators). Total revenue (transaction fees plus issuance to validators) is ~$2.212bn, although we do not count issuance revenue as blockchain revenue as it represents a block subsidy issued by the protocol for security rather than a measure of external activity that generates blockchain revenue via external utility.

The Early Days: Ethereum Fee Revenue Near the Beginning

• ETH’s market cap vs its transaction fee revenue may seem relatively expensive for Ethereum (although there is an argument that transaction fee revenue is incredibly high margin, since PoS costs are negligible, and fees can potentially be considered net profit).
• Nevertheless, it is important to look at the trajectory of transaction fees in addition to the outright fee revenue, as Ethereum is still in its adoption growth phase. In the early days of Ethereum (2016), before any useful applications were built and used to scale, annual fees were a mere 14,975 ETH. In 2017, as the first crypto mania began and limited applications like the first NFTs (CryptoKitties) and ICOs were on the rise, ETH transaction fees increased by 8x. When coupled with the increase in ETH price, fees denominated in USD increased by 314x YoY.

The First Bear Market Cycle: ETH Fees in 2018-2019

• 2018 marked the end of the first hype cycle for ETH – driven largely by unsustainable ICO revenue. Nevertheless, demand for ETH blockspace remained high for half of 2018, resulting in 268k ETH in annual fees (2.2x YoY in ETH terms), or $160mm in USD denominated fees (3.5x YoY). This increase is notable as price and activity entered a deep bear market in mid-2018 onward.
• 2019 was a true bear market trough, with ETH’s average price lower than 2017 and 2018 and fee revenue dropping due to limited use cases. Nevertheless, the important point to note is that aggregate ETH fees in 2019 were still 1.56x higher (in ETH terms) in the 2019 doldrums as they were in 2017 – suggesting increased stickiness in demand for Ethereum blockspace.

Ethereum Fees During the 2020-2022 Crypto Cycle

• The thesis for upward trending demand for ETH blockspace (despite the volatile cyclical nature of crypto) was corroborated during this past bull / bear cycle of 2020-2022.
• While the fee driver for ETH in the previous cycle centered on ICOs, 2020 and 2021 brought two important new product-market fit applications for the Ethereum blockchain: DeFi and NFTs. Summer of 2020 marked the start of the DeFi movement, and ETH fees significantly increased (even though the biggest bull market price action did not occur until 2022).
• Fees in ETH terms and in USD terms skyrocketed from 2019 to 2020 to 2021. Even in 2022, where the second half saw a brutal bear market washout, aggregate ETH and USD fees were higher than in 2020.

Demand Drivers for Ethereum Fees: Present Day and Future

• Fee revenues for 2020 - 2022 were multiples of the fees paid in the 2017-2018 cycle  – and the 2022 figure was during a brutal crypto bear market. The ongoing resilience and stickiness in ETH fees simply highlights that underlying demand for Ethereum blockspace persisted throughout a cycle.
• Additionally, this transaction fee revenue occurred without the real world adoption that we believe is likely to accelerate in the next 5 years. From the graphic below, the largest contributors to ETH transaction fees over the past  1.4 years are (1) NFTs, (2) trading using decentralized exchanges (Uniswap, MetaMask router), (3) sending ETH between parties, and (4) stablecoins.
• We expect these existing use cases to be bolstered by new real world use cases (tokenization, institutional DeFi) that could usher consistent, sustainable fee revenue for Ethereum going forward.

L2s, Transactions, and Gas Prices: An Uncertain Outcome

• While Ethereum fee revenue signals healthy demand for the network, transaction fees themselves cannot remain expensive over the long run. Ethereum’s fee revenue is higher when its ”gas price” (cost per transaction) is high. However, high fees will price out users over the long run.
• The solution to high Ethereum gas fees are L2s, which take batches of user transactions, compress them together, and publish onto the ETH L1. As a result, L2 transactions are as secure as L1 but are magnitudes cheaper.
• An open question: if most transaction volumes migrate to L2s to onboard the masses (a good outcome), that means ETH L1 gas prices will likely fall. What does that mean for ETH L1 revenue? Is lower revenue, but higher adoption, enough to sustain monetary premium for Ethereum? Our future research will examine the effect of L2s to see how much L2 adoption would be needed to drive congestion (high gas prices) on ETH L1.

‍