Why Should an L1 Blockchain Choose to be a Rollup?

This question was the essence of a recent conversation between some of the most influential builders in crypto. It’s a debate that raises questions about the future of blockchain architecture. But before we get into the details of that conversation, here are some things you know know:

Monad is a highly optimized Layer 1 (L1) blockchain that is EVM-compatible, meaning it can run Ethereum smart contracts seamlessly. Its selling point is speed and performance, boasting 10,000 transactions per second, 1-second block times, and single-slot finality. It also requires relatively low hardware, making it accessible to a wide range of users and validators. The project aims to improve both throughput and latency, two key issues facing blockchain networks today.

Celestia is taking a different approach. It’s building a data availability (DA) layer designed to be modular. Instead of focusing on being a full blockchain with execution and consensus, Celestia handles only the data availability and ordering, while rollups manage the execution. This allows Celestia to scale faster and more efficiently since it doesn’t carry the full weight of traditional blockchains. Many believe rollups like those on Celestia represent the future of blockchain scalability.

EigenDA*, a project under EigenLayer, offers a radically minimalistic design for data availability. The idea is to strip away traditional blockchain features like peer-to-peer (P2P) networking, consensus, and replication to achieve hyperscale throughput. EigenDA uses a model where data is not fully replicated but encoded with erasure codes, allowing only necessary chunks to be distributed. This design is aimed at removing bottlenecks related to bandwidth and scaling while keeping the network censorship-resistant.*

The Debate: Rollups vs. Layer 1

So, back to the conversation. It started off with a tweet from Nick White of Celestia, suggesting that Monad might be more scalable if it ditched its L1 architecture and went full-on rollup. In Nick’s view, you either build a scalable data availability layer (like Celestia) or a scalable execution layer. He believes trying to do both in a “monolithic” way is a dead end.

However, Shamim from Monad doubled down on their decision to focus on building an optimized Layer 1. Monad’s mission is to enhance performance directly at the base layer. They believe a strong, efficient L1 is critical for long-term scalability. Their argument? Even if rollups are part of the future, you need a robust foundation for dApps and other innovations to build on, and Monad’s goal is to deliver that foundation by improving throughput, latency, and fee markets.

The tension between modular approaches (like rollups) and monolithic ones (where everything is handled by the Layer 1) speaks to a larger philosophical debate in crypto: Do we want to optimize everything at the base layer, or do we want to distribute responsibilities across different layers?

The Architecture Trade-offs

Another crucial topic in this conversation is data availability (DA), a blockchain’s ability to store and retrieve data efficiently. As blockchain ecosystems grow, DA becomes a bottleneck for scaling. Enter EigenDA and Celestia, two projects focused on building high-performance data availability solutions.

Celestia’s approach revolves around maximizing the benefits of rollups. It focuses on data availability and pushes execution to rollups. This means faster finality (12-second blocks) and lower latency. However, the trade-off is that Celestia’s throughput is currently lower than EigenDA’s, at just 0.16 MB/s. But Celestia’s team insists that the gains in decentralization and permissionless rollup capabilities outweigh the current limitations in throughput, which they plan to increase responsibly.

EigenDA, on the other hand, prioritizes hyperscale throughput. By minimizing the system requirements and eliminating consensus and replication, EigenDA boasts an impressive 15 MB/s throughput. However, it doesn’t offer finality or consensus internally, meaning you’ll need to write your DA certificates to another layer, like Ethereum or an L2, for finality. It also has some limitations, such as reliance on a single disperser, which can become a bottleneck until further decentralization steps are implemented.

Why Does This Matter?

For anyone interested in the future of blockchain, this discussion offers a glimpse into the future direction of the space. If rollups are the future, modular systems like Celestia may become more dominant. If performance at the base layer remains key, monolithic approaches like that of Monad or Solana could set a new standard. EigenDA’s unique design offers a completely new perspective on how to handle data availability, which could also revolutionize how blockchains scale. Each approach has its own merits and trade-offs. Whether it's through highly performant L1s, hyperscale DA layers, or rollups, the goal remains the same: a more scalable, decentralized, and efficient blockchain ecosystem.

These conversations allow us to better understand the direction of crypto innovation and even spot opportunities early.