Layer 2 tokens and ETH: Which one is easier to make money investing in?

Layer 2 on Ethereum has made significant progress in the past few years.

Today, ETH L2’s TVL exceeds $40 billion, up from $10 billion a year ago.

On @l2beat you’ll find a list of 50+ L2s, though the top 5-10 account for over 90% of TVL.

After EIP-4844, Txns fees are very cheap, with fees for Base, Arbitrum, etc. being less than $0.01.

Still, L2 tokens have generally performed pretty badly as liquidity investments compared to ETH (they are clearly excellent venture investments).

Let’s go back and plot the valuations of the major L2s vs ETH. One notable observation:

Even though the number of listed L2s has increased, their percentage of ETH’s total FDV has remained the same.

Two years ago, the only listed L2s were Optimism and Polygon. They accounted for 8% of ETH.

Today, we have Arbitrum, Starkware, zkSync all joining the L2 ranks. They now account for 9% of ETH.

Every new L2 token that enters the market actually dilutes the valuation of previously listed tokens.

Compared to ETH, the returns on investing in L2 tokens are quite significantly lower than ETH.

Returns over the past 12 months (as of today):

ETH +105%;

OP +77%;

MATIC -3%;

ARB -12%.

The leading L2 token has long had a FDV of ~$10B. In a sense this is fairly arbitrary and I don’t think there is a good reason from market participants why the valuation is $10B vs $20B vs $3B. Ultimately there is significant supply pressure in the form of vesting, unlocking, etc. - from people who need liquidity and/or a surge.

Why isn't there more liquidity buying these L2s? The aforementioned L2s generate $20-30 million in fees per month. Since eip-4844, fees have dropped and now total $3-4 million per month. This equates to $40-50 million in fees per year.

Blobs are dirt cheap right now. Most of the revenue goes to the bottom line.

(Including optimization, arbitrum, polygon, starkware, zksync)

The combined value of these leading L2s is now about $40 billion FDV, with an annualized cost of $40 million. That’s about 1,000x.

This is in stark contrast to the large DeFi protocols, which have valuations much lower than DeFi protocols (based on annualized fees in the last month). Typically in the 15-60x range.

DYDX - 60x;

SNX - 50x;

PENDLE - 50x;

LDO - 40x;

AAVE - 20x;

MKR - 15x;

GMX-15x.

There are many more L2s in the pipeline. As these products come to market, it will likely continue to put pressure and dilute the FDV category of L2s. The supply is simply too large for the liquidity market to easily support.

A few final thoughts:

1. In the long term, L2 can generate significant fee income. L2 generates $150M in fees per year (including Base, Blast, Scroll). This can grow meaningfully with L2 activity.

2. The above comment is not about any specific L2, this is more about the category. It seems difficult to buy a basket of L2s with ~$40B FDV for ~$40M in fees (1000x) and expect to outperform ETH in the long term.

3. It’s clear that high throughput chains like L2, Solana, Sui, Aptos, etc., are not lacking block space. The limitation is mainly on the use of this block space on applications. I expect to see an increasing focus on applications and a liquidity market that rewards the application layer rather than the infrastructure stack in the coming years.

4. In the last cycle, it was more common for projects to go public earlier. MATIC had a FDV of less than $50 million on liquid markets and is now over 100x higher at over $5 billion. This appreciation was achieved through liquid markets. However, recently, this has not been the case for $OP $ARB $STRK $ZK and most other L2 tokens that may eventually go public.