Investing in L2 vs Investing in ETH: Which Has a Brighter Future?

Author: James Ho

Compiled by: Shenchao TechFlow

Investment in L2 vs ETH

In the past few years, Layer 2 (L2) solutions on Ethereum have made significant progress. Currently, the total value locked (TVL) in Ethereum L2 exceeds $40 billion, compared to just $10 billion a year ago. On @l2beat, you will find over 50 L2 projects, but the top 5-10 projects account for more than 90% of the TVL.

After the implementation of the EIP-4844 proposal, transaction fees have dropped significantly, with transaction costs on platforms like Base and Arbitrum even falling below $0.01.

Despite the tremendous progress in technology and usage for L2, the performance of L2 tokens as liquidity investments has generally been poor (although they have performed well as venture capital). You can find many jokes and anecdotes about L2 tokens underperforming relative to ETH.

We reviewed the valuation of major L2s relative to ETH. A notable observation is: despite the increase in the number of listed L2s, their total fully diluted valuation (FDV) as a percentage of ETH has remained unchanged.

Two years ago, the only listed L2s were Optimism and Polygon, with their FDV accounting for 8% of ETH. Today, we have L2 projects like Arbitrum, Starkware, and zkSync, with their FDV accounting for 9% of ETH.

Each new L2 token listing effectively dilutes the valuation of previously listed L2 tokens.

The result of investing in L2 tokens is a significant underperformance relative to ETH. The returns over the past 12 months are as follows:

• ETH: +105%

• OP: +77%

• MATIC: -3%

• ARB: -12%

For a long time, the FDV of major L2 tokens has been around $10 billion. To some extent, this is quite arbitrary, as market participants do not have strong reasons to explain why it is $10 billion instead of $20 billion or $3 billion. Ultimately, there is significant supply pressure due to demand liquidity and/or large unlocks.

The aforementioned L2s generate fees of $20 million to $30 million per month. Since the implementation of EIP-4844, fees have dropped to $3 million to $4 million per month, with annual fees around $40 million to $50 million.

Including: Optimism, Arbitrum, Polygon, Starkware, zkSync

Currently, the total FDV of major L2 tokens is about $40 billion, with annual fees of $40 million, resulting in a valuation multiple of about 1000 times.

This stands in stark contrast to large DeFi protocols, which typically have valuation multiples ranging from 15 to 60 times (based on last month's annual fees):

• DYDX: 60 times

• SNX: 50 times

• PENDLE: 50 times

• LDO: 40 times

• AAVE: 20 times

• MKR: 15 times

• GMX: 15 times

As more L2 projects are listed, the FDV of L2 tokens may continue to face pressure and dilution. The market is oversupplied, making it difficult for the liquidity market to support easily.

Conclusion

1. In the long run, L2 may generate considerable fee income. L2 generates $150 million in fees annually (including Base, Blast, Scroll), and this figure could grow significantly with increased L2 activity.

2. The above content is not aimed at specific L2 projects but is a broad observation of the entire category. Buying a basket of L2 tokens with an FDV of about $40 billion and fees of about $40 million (1000 times) and expecting it to outperform ETH in the long term seems challenging.

3. Clearly, there is no shortage of block space between L2s and high-throughput chains (such as Solana, Sui, Aptos, etc.). The limiting factor lies in the applications using this block space. I hope that more focus will be placed on the application layer in the future, and that the liquidity market will reward the application layer rather than the infrastructure layer in the coming years.

4. In the previous cycle, it was more common for projects to be listed significantly ahead of time. MATIC was listed in the liquidity market with an FDV of less than $50 million and has now surpassed $5 billion, growing over 100 times. However, the recent cases of $OP, $ARB, $STRK, $ZK, and most other L2 tokens that may eventually be listed are not the same.