Proposed change could save Ethereum from L2 ‘roadmap to hell’

If you believe Ethereum’s legion of critics on X, the blockchain is dying. On one side, it’s being beaten as ultra-sound money by Bitcoin, and on the other, it is outdone by faster and cheaper transactions by Solana.

Fee revenue for the layer 1 has “collapsed” by 99% in the past six months as “extractive L2s” steal all the users, transactions and fee revenue while paying virtually nothing to the base layer. That’s made ETH slightly inflationary again, with its price sliding.

The L2s are siloed from one another, and each has a financial incentive to maximize its own revenue at the expense of ETH. Base charged users $2.5 million in fees last month and gave just $11,100 back to the L1. OP mainnet is reportedly raking in $321.31 for every $1 it funnels back to Ethereum.

“ETH is in a death spiral,” proclaimed Bitcoin maximalist Fred Kreuger, who claims the network’s plummeting fee revenue of just $73 million a year spells certain doom and cannot justify its $300-billion market cap.

So, given this grim picture, it may be a surprise to discover that key figures in the Ethereum ecosystem aren’t more worried. They believe that despite some challenges, the Ethereum roadmap is still on track, even if critics are painting it as a highway to hell.

Maybe that’s just because they’re a lot more confident in where the roadmap will end up. Ethereum Foundation researcher Justin Drake told The Rollup podcast this week that if Ethereum can make a few course corrections to its roadmap, it’ll be a case of “winner takes all.”

“We can get to 10 million transactions per second (TPS) within the next 10 years. And that’s enough for the whole world. It’s 100 transactions per day, per human,” he said, predicting a multitrillion-dollar market cap.

Skip ahead to the proposed solution to fix the roadmap

Ethereum roadmap has already seen a 20x increase in transactions

The TL;DR for the roadmap is that Ethereum’s painfully slow base layer will focus on being the most decentralized and bomb-proof layer possible while making changes to allow L2s to execute transactions fast and cheaply. 

The L2s, based on optimistic rollups or zero-knowledge (ZK) tech, inherit Ethereum’s security (and hopefully ethos), and the number of transactions in the ecosystem has grown to peak at around 16 million per day at a very low cost for users.

“L2s have led to a huge increase in the capacity of the Ethereum stack,” says Ed Felten, co-founder of Arbitrum, at Korean Blockchain Week. Arbitrum is the most successful L2 at present, though Base is nipping at its heels. Felten says the argument that L2s are stealing transactions from the now desolate L1 doesn’t make any sense, as the vast majority of those transactions just couldn’t occur on the base layer.

“If that 20x demand shows up and Ethereum L1 is all we have, 19x of that won’t happen, or it will happen on a different chain,” he says. 

“So, the idea that the alternative to scaling up Ethereum with L2s is that people will just stay on Ethereum, and Ethereum will somehow magically serve much more traffic than it currently can — I don’t think that’s realistic. So, I think this was the path to scaling; it’s been very successful.”

Felten actually started work on scaling Ethereum with rollups at Princeton University before the Ethereum network even launched, so he’s been all in on the L2 vision for a decade now.

Not everyone agrees the right balance has been struck. Ethereum researcher Max Resnick thinks the roadmap is now too weighted toward the L2s and should be reorientated to focus at least some effort toward scaling the base layer so it can be used more effectively for high-volume DeFi.

Other chains are also grudgingly conceding that monolithic blockchains may not be able to do everything, and a modular design like Ethereum isn’t as bad as they once thought.

While Solana is successfully pursuing fast and cheap scaling on the base layer, it recently started using L2s — which it calls “network extensions” — to enable a broader range of use cases. 

Dan Albert from the Solana Foundation tells Magazine:

“Solana does not need L2s to scale… but certain apps on Solana may choose to build an L2 or may need something like an L2 or a network extension to provide the very particular kind of product or user experience that they want for their customers.”

Are the fees Ethereum charges L2s too low?

Much of the current bearishness around the roadmap is due to the plunge in fee revenue, which is kind of ironic, given that for years, Ethereum was criticized for its ludicrous $100 transaction fees during times of high demand.

Felten says lowering fees was a deliberate design decision to attract users.

“When you scale, when you increase your capacity, when you increase your gas limit, the price of gas will go down, and there will be less revenue collected by your chain. And you can view that as a failure, or you can view it as a success, depending on your goal.”

While some Ethereum community members have proposed increasing the fees charged to L2s to improve revenue, Rushi Manche, co-founder of Movement — an L2 that’s compatible with Ethereum and the Facebook-developed Move language that hit 12,000 transactions per second on testnet recently — says that’s short-term thinking.

“I think the Ethereum Foundation is looking at this from a 30-year time horizon, which teams don’t think about; they’re thinking about two years from now. But in a 30-year time horizon, what you need is low fees, low barriers to access and entry, let people try crypto. One that early network effect is there, you can increase, so actually, I think it’s the right call.”

Will Ethereum’s fee revenue rise again?

Fee revenue is set to increase when the current amount of space for data availability (known as blobs) gets used up. Felten says that’s not far off. 

“If you look at the percentage of the blob target that is used, it’s consistently in recent months, above 80%, often above 90%,” he says. “If the demand for blob space grows above 100% of the target, then the blob data price will increase a lot.”

On the day of the LayerZero airdrop, demand on the network was so great that fees skyrocketed. “The price of blob space went up by 13 orders of magnitude in one day,” he says, adding that when the cost of L2 transactions increases, then “users back off and use less.”

So, if the fee revenue problem will fix itself as more people use the ecosystem and the number of transactions is increasing exponentially, why has the roadmap received so much criticism?

One big reason is that many people don’t see L2s or rollups as an extension of Ethereum but rather as a competitor to it. They see a bunch of siloed rollups, each with their own agenda, liquidity and incentives, which may not align well with Ethereum’s goals.

While Drake points to a dramatic reduction in the base fee and the creation of infrastructure to enable the full danksharding upgrade as proof the roadmap is on track, he concedes its designers didn’t foresee that issue.

“From a pure kind of throughput perspective, TPS perspective, we have scaled dramatically. The problem is that some other issue kind of came up that we didn’t 100% appreciate or foresee, is that the quality of the blob space is kind of lower than what we would want. Basically, we have these L2s, these rollups that are to a very large extent siloed and that creates fragmentation for Ethereum.”

How the Ethereum ecosystem is tackling interoperability

Greater interoperability is the obvious way to make a bunch of disparate L2s feel like Ethereum again, but there’s no easy way to make that happen. At present, users can transfer between rollups via Ethereum (which can take a week), or they can use bridges, which are honeypots for hackers.

Various solutions are being worked on so that users don’t feel, to quote Vitalik Buterin, “like you’re jumping between 34 blockchains. It should feel like you’re using Ethereum.”

Magic Account has just unveiled a wallet that uses account abstraction/smart accounts to essentially hide the underlying infrastructure from the users and allow them to spend funds on any chain instantly. Based Rollups that are more aligned with the base layer and address liquidity fragmentation are also becoming popular, although they are hampered by Ethereum’s 12-second block times.

The various rollups are all working on interoperability solutions that connect various chains in their own ecosystems. Optimism has the Superchain to connect linked L2s like Base and Optimism; Arbitrum Orbit connects its L2s and L3s; and Polygon has the Agglayer, which connects its various ZK solutions and proof-of-stake network.

“The way we view it is that everybody is trying to make an intranet right: Optimism, Superchain, Agglayer,” says Bryan Pellegrino, co-founder and CEO of interoperability solution LayerZero. Intranet usually refers to a private network like those used within companies to share internal information and communication.

“Intranets have their place in the real world; intranets are still used at scale today, but like, the internet is still the internet.”

“You get some benefits. You have shared sequencing. You basically have better composability across all these assets, but you lose a little bit of sovereignty. And still, when you’re trying to go between this intranet and that intranet, you need some way to communicate externally there.”

LayerZero is seen by many as the best solution in the bridging space, designed in such a way that it hasn’t been hacked yet, while other bridges have collectively lost billions.

But even good bridges and message passing can’t solve interoperability on its own. Many critics argue that L2s have a financial incentive to try and keep users and their fees in their own walled garden and don’t actually want genuine interoperability.

Magazine points out to Felten that Arbitrum seems unlikely to join its competitor Optimism’s Superchain any time soon, even though they both use similar technology.

While he doesn’t really address why not, he does say that Arbitrum will compete by providing a “better product, lower cost, better security, more reliability, and so I don’t think we should be afraid of users leaving.” 

“We should want the doors to be open so people can come in,” he says.

Interoperability challenges between Ethereum L2s

Regardless of the motivation, Felten points out there are big practical problems with making the different L2s interoperable. Chains don’t operate in lockstep, and they all have different block times. 

“You won’t reduce the friction in crossing chains to zero,” Felten warns. “As long as there are different trust assumptions, different governance on different chains and so on, the movement of an asset or the coordination across chains is a decision that potentially carries risk.”

“You’ll get something that is more like using the internet, where you have different servers or services on the internet that interact, but they don’t interact simultaneously or within a transaction atomically. Instead, there are short delays.”

“The speed bumps will be much smaller than they are now, but they won’t be nothing.”

The seamless operation within a monolithic chain is known as “synchronous,” while the internet-style delays are called “asynchronous.” Pellegrino says that decentralized app designers will need to radically rethink how things operate for this version of interoperability.

“Historically, everything was built for one chain, synchronous composability,” he says. “Applications are designing with async patterns now, but it’s a really big design shift.”

Ethereum L1 shared sequencer could Make Ethereum Great Again

But there is a different approach Ethereum could embrace as part of its roadmap to enable L2 interoperability and composability, says Drake.

Composability is one of the reasons DeFi took off on Ethereum. It refers to things like “Money Legos,” where anyone could build a finance DApp on Ethereum that can seamlessly interact with any other DApp, clicking together like pieces of Lego.

A credibly neutral, decentralized, shared sequencer run by Ethereum validators would enable synchronous transactions and composability between different types of rollups. Ethereum researchers have been discussing the idea for some time, but it’s not yet part of the official roadmap.

ZK-rollups have additional paths to interoperability, but ZK tech isn’t as fast or cheap just yet. Projects including Espresso, Astria and Radius are working on shared sequencing but they don’t rely on Ethereum’s validators and have different trust assumptions.

But as Collective 2077 director Emmanuel Awosika asks, why would L2s be willing to give up control over their own centralized sequencer (or the intranet’s sequencer) if it means they have to forego sequencer fees?

Drake argues that about 80% of L2 fees are congestion fees, while the other 20% is contention fees related to MEV, or Maximal Extractable Value.

There are different methods, but broadly speaking, MEV is where bots manipulate and reorder transactions in the queue with priority fees so they can profit from the trade first.

But Drake argues that MEV will soon shrink down to less than 1% of the total due to private mempools hiding transactions (making MEV impossible) and decentralized exchanges returning MEV to users. MetaMask is already using a virtual mempool to combat MEV.

He argues the benefits and increase in use they’d get from higher composability would outweigh the MEV loss, and they could maintain congestion fee revenue.

If a base layer decentralized shared sequencer was embraced, instead of a bunch of siloed L2s, the community would have a genuine Ethereum ecosystem that could scale up to the size of the globe while retaining composability. 

The bull case for Ethereum: Multi-trillion dollar market cap

That’s how Drake sees 10 million TPS happening on Ethereum. He argues that enabling the capacity would induce demand, rather like how faster broadband speeds encouraged companies like Netflix to make use of it.

If that happens, back-of-the-envelope math suggests that even data availability fees of one-10th of a cent would see Ethereum rake in $1 billion per day, while fees of half a cent would create $5 billion a day, potentially leading to an Nvidia-style market capitalization.

“On top of this base valuation of trillions of dollars, there is this opportunity for Ether to be money, to be collateral as economic security, but also as economic bandwidth for decentralized stablecoins,” he says. “Long story short, you know, there’s a very exciting potential roadmap over the next 10 years.”

That’s the bull case from an Ethereum stan. But it’s still going to take years before the problems are solved and we see a decentralized sequencer realize this vision. In the meantime, Solana and Sui are scaling already. And there’s no point in having the right roadmap if it happens so slowly that it’s no longer required.

For corrections or clarifications please email andrewfenton at cointelegraph.com

This article first appeared at Cointelegraph.com News