Why EVM L1s are ending: Celo's move to an OP Stack L2
Quote692 words
Backs the claim that EVM-compatible L1s lose to L2s, using Celo: an OP Stack L2 with EigenDA would cut block time from 5s to 2s, raise throughput 50% and keep 1-block finality and fee abstraction. Celo chose L1 because 2020 L2s were too costly.
Key claims
Celo L2 block time 2s vs 5s, throughput +50%
Celo proposes EigenDA for sub-cent fees
Celo L1 launched in 2020 before L2s were live
The era of EVM L1s is coming to an end!
@marek_ made a strong case against EVM compatible L1s last week, arguing that with improving modular infrastructure & tooling, L2s are superior.
Is that true?
In my opinion: Yes.
Let's look at the why in the case of @Celo. 🌱👇
🟡 What is Celo?
Originally, @Celo was born as the first carbon-negative, EVM-compatible L1 with a specific focus on sustainability and regenerative finance (ReFi).
The latter is a concept of an alternative financial system that promotes and restores environmental, social, and financial stability, fostering financial inclusion globally.
🟡 Why go for an L1 initially?
To build out the above vision, @Celo had to build an execution layer that maximizes accessibility and provides a seamless user experience. Back in early 2020 when Celo mainnet went live, scaling EVM execution on L2 was very hard.
The first L2s (Arbitrum & Optimism) didn't go live on mainnet until a year later, and even then transaction fees on L2 would still often amount to a few dollars. Too much to serve the world's unbanked population with scalable, on-chain financial services.
The only option to provide the fast finality and low cost that @Celo was striving for really was building a standalone L1.
🟡 Celo L1 vs. Ethereum L1
Celo is originally a fork of the Ethereum Go client (geth), but differed significantly from Ethereum back in 2020 by using a proof-of-stake based PBFT consensus mechanism way before Ethereum's PoS transition.
Although Celo's on-chain assets are not ERC-20 tokens on the Ethereum Mainnet, they have ERC-20 compliant interfaces, allowing compatibility with existing Ethereum tools and code.
Additionally, Celo, like Ethereum, supports EVM smart contract functionality. This enables developers to use familiar programming languages (Solidity or Vyper), developer tools, and standards from Ethereum (e.g. Truffle).
🟡 Celo L2 Architecture
Okay, now we know how @Celo looks today. What does Celo look like tomorrow?
The Celo L2 was first proposed at EthCC 2023. @cLabs and the @Celo community have since spent the past year deciding on how to implement this L2 technically.
The final decision resulted in leveraging @Optimism's OP Stack and a current proposal to leverage @eigencloud's @Eigen_DA to achieve sub-cent transaction fees.
The Dango testnet that is now live, is the first opportunity for core infrastructure providers to familiarize themselves with the @Celo L2 codebase!
🟡 What are the benefits?
Okay cool, but how is this superior to @Celo's current L1 setup?
First of all, this is a question of alignment. Ethereum has always been @Celo's home. The reason why they built alongside Ethereum rather than on it back in 2020 simply is that it was not possible back then.
Nevertheless, @Celo always strived to provide an execution environment as equivalent to the Ethereum L1 as possible.
By transitioning to an L2, @Celo can deepen this alignment and even build on Ethereum's battle-tested security guarantees.
A core focus when designing @Celo's L2 was maintaining the key benefits that the Celo L1 introduced.
With an OP Stack based rollup that leverages @eigen_da, this is not only possible but actually allows @cLabs to further improve the UX through:
✦ Fast, secure and cheap bridging of wETH & ERC-20 tokens between Ethereum & @Celo.
✦ Decreasing the block time to 2s from 5s
✦ Increasing the throughput by 50%
Additionally, @Celo is able to:
✦ Maintain 1-block finality
✦ Enable fee abstraction, which means that multiple ERC-20 tokens can be used as gas currencies incl. Celo-native USDT & USDC)
✦ Allow for user onboarding with a phone number, email address, or other social identifier via SocialConnect & more!
✦ Likely repurpose the L1 validator set in the context of a decentralized sequencer.
🧠 Conclusion
So if you can build EVM execution layers on top of Ethereum that are more secure, faster and user-friendly than alternative L1s, why go for the latter?
Hence totally agree with @marek_ that it doesn't make much sense anymore to most EVM-compatible chains to build outside of Ethereum. While there might be cases where certain execution layer innovation makes a sovereign L1 more suitable, the era of EVM-compatible alt-L1s does seem to come to an end.
QUOTING @xkonjin
"If you're an EVM compatible L1, you have no business being an L1 anymore." - @marek_ This bold claim from @Celo CEO is sparking heated debates at @EthCC. But is he right? Let's dive into Celo's audacious L1 to L2 transition. 🧵 t.co/98DzJyquyX
"If you're an EVM compatible L1, you have no business being an L1 anymore." - @marek_ This bold claim from @Celo CEO is sparking heated debates at @EthCC. But is he right? Let's dive into Celo's audacious L1 to L2 transition. 🧵 t.co/98DzJyquyX
What came next · since 15 Jul 2024
Since then
Zero has written 1 more piece on Celo since this one.