Ethereum L1 vs L2 Gas Prices

This morning, I found myself mulling over the relationship between the cost of gas on Ethereum L2 networks in relation to mainnet gas costs.

The idea that L2 fees will all continue to be a fraction of mainnet fees makes intuitive sense, but it may not always be the case over the next few years. Currently, prices on L2 roll-ups are considerably lower than on mainnet, and certainly, it makes intuitive sense that they would be.  Most people in the space believe this L2 gas discount should continue into the future. But is this really true? Will L2 gas prices remain so low compared to mainnet?

A counter-argument is that users should be willing to pay similar prices for similar security guarantees. From a user’s economic perspective, if I can do the same things I want to on mainnet with similar guarantees, but with greater convenience (i.e. near-instant settlement), shouldn’t I be willing to pay more, or at least the same amount?

Rollups Will Drive Mainnet Premiums, but They Don’t Yet

While end users may choose where they transact, there are rollups and major cornerstone protocols that must spend gas in order to continue operating, and who will always be willing to pay a premium where a typical end user may not. We can assume that an L1 gas premium will emerge over time as these ‘mandatory’ settlement expenses begin to account for a significant portion of the fees spent on the base layer. However, these functions still account for a very small portion of the total fees spent, and it will likely be some time before they take up significant space on L1 and have a significant impact on fees.

Rollups are currently spending around $300k per day in gas, or about 1.5% of the total gas spent.

Other Drivers of Mainnet Premiums

Mainnet NFTs are Superior

Upwards 30% of fees spent in the last 24 hrs are NFT-related, as we continue to see the effects of the belief that NFTs launched on mainnet should be worth more. While this belief may sound weak, cost as a quality filter definitely works! We have only to look at all the spam in our Polygon wallets to appreciate this.

I would not be surprised to see this trend continue for quite some time until a cross-chain NFT standard becomes widely adopted, and in the meantime this will continue to prop up high gas prices.

Fungibility

While depositing to Curve on Arbitrum likely surpasses the experience on mainnet (less fees and less time spent waiting for confirmations), your money is likely also slightly less fungible. Few exchanges have adopted L2 off ramps and it’s time-consuming finding a bridge and moving money across chains. In many cases, users must still bridge via mainnet, driving up mainnet block space demand, and thus L1 gas prices.

What about Uptime? Security Guarantees Aren’t Everything.

Being able to exit funds to mainnet through an escape hatch is great, but many DeFi users require the ability to manage loans or funds in real-time. Ethereum’s has proven itself with uptime over the course of countless market events, but an L2’s validator network is still relatively untested, as we have seen with recent downtime from Arbitrum and Optimism’s upgrade last week. This results in mainnet transactions to be in demand and raises mainnet gas prices.

There are likely other reasons that a typical user might be willing to pay a premium on mainnet, but does it really outweigh the convenience of near-instant settlement?

What About the Future?

We’re scaling quickly, but it will be an eternal race. As @HenriLieutaud put it recently on Twitter:

He goes on to say “Chose your fee level depending on your distance to L1,” which nicely sums up the likely long-term state of gas fees on Ethereum.

StarkWare’s Fractal Scaling blog post goes into some more high-level thoughts about the many-layer future.

Where Does this Leave us Now?

As we move further into 2022, I would not be surprised to see higher L2 gas prices than most people expect. There may even be times when gas prices on certain L2s exceed mainnet costs, i.e. during a big protocol launch, NFT drop, or simply if a killer app launches on an L2.

Luckily, most of the hard work has been done and we can expand the number of rollups as demand increases, and validiums/plasma, volitions, and adamantiums will increase scalability dramatically.