What new changes in public chain infrastructure like ETH are worth paying attention to recently?

Written by: Lao Bai

Recently, the hottest track in the primary market is undoubtedly AI, followed by BTC. About 80% of the projects discussed daily are concentrated in these two areas. Personally, I can discuss 5 or 6 AI projects in a single day at most.

It is foreseeable that the AI bubble will peak in the next year or two. With hundreds of new AI projects coming online, the market value of the AI track will reach its zenith. When the final bubble bursts, leaving a mess behind, it will also give birth to true unicorns that find the intersection of AI and Crypto, pushing this track and the entire industry forward.

So, in the current overheated AI environment, let's take a moment to look at the changes that have occurred in the infrastructure layer over the past few months, especially in the public chain infrastructure track. Some new developments are indeed worth mentioning.

ETH, or the further deconstruction of single-layer chains

When Celestia first proposed the concept of modularity and the DA layer, the market actually took quite some time to digest and understand it. Now, this concept has already taken root in people's minds, and various RaaS infrastructures have flooded to a stage where the number of infrastructures > the number of applications > the number of users. (RaaS: Rollup-as-a-Service, which refers to providing ready-made Rollup products and services to help application developers quickly launch Rollups.)

In the past few months, there have been various technological advancements in the execution layer, DA layer, and settlement layer, each giving rise to new technical solutions. Even the concept of the settlement layer is no longer exclusive to ETH. Let's briefly discuss representative technologies from each layer.

Execution Layer

The hottest concept in the execution layer is undoubtedly parallel EVM, represented by Monad, Sei, and MegaETH. Existing projects like FTM and Canto are also planning upgrades in this direction. However, just as not all ZK projects protect privacy, projects labeled as parallel EVM actually differ in their technical routes and ultimate goals.

Using a diagram from Sei for an intuitive display, it is evident that in an optimistic scenario, the shift from sequential processing to parallel processing significantly enhances performance.

Parallel EVM can actually be divided into several different technical routes:

1) From the perspective of how transactions are parallelized - there’s nothing new under the sun, just the difference between a priori and a posteriori.

A priori, represented by Solana and Sui, requires transactions to explicitly declare which parts of the chain state they modify, allowing for pre-checking for state conflicts (e.g., accessing the same AMM pool) before packaging blocks. If conflicts are found, those conflicting transactions are discarded.

A posteriori, also known as optimistic parallelism, is represented by Aptos BlockSTM, which assumes that there are no conflicts and collects transactions first, then checks after execution. If conflicting transactions are found, they are declared invalid, the results are refreshed, and the process is repeated until all transactions in the block are executed. Sei, Monad, MegaETH, and Canto use similar solutions.

We have also seen solutions for parallelization in scenarios of state conflicts (like accessing the same AMM pool mentioned above) in the primary market, but they seem to be relatively complex in engineering, and their commercial viability is still under evaluation.

2) From the degree of emphasis on parallel EVM - it can also be divided into two schools.

One is represented by Monad and Sei, which regard how to parallelize transactions as the main scalability approach, making parallelism the primary narrative. For example, Monad not only employs optimistic parallel processing but also has developed MonadDB, with asynchronous I/O specifically designed to complement parallel processing.

The other is the approach of Fantom, Solana, and MegaETH, where parallelism is one of the scalability solutions, but merely one among others. Parallelism serves as a supporting narrative, with performance improvements relying more on other technical solutions.

For instance, Fantom's Sonic upgrade focuses on the FVM virtual machine + an optimized Lachesis consensus mechanism. Solana's next phase emphasizes the modular architecture of the Firedancer new client, optimizing network communication mechanisms and signature verification, among other things.

MegaETH aims to achieve Realtime Blockchain. It is built on Paradigm's newly developed Reth high-performance client, with further optimizations in the full node state synchronization mechanism (synchronizing only state differences rather than all data), hardware design of the Sequencer (using a large amount of high-performance RAM with storage capabilities for state access, avoiding slow disk I/O), and improvements in the Merkle Trie data structure. This represents a comprehensive enhancement across software, hardware, data structures, disk I/O, network communication, and transaction ordering and parallel processing, pushing the performance ceiling of EVM to its limits, approaching "Realtime Blockchain."

DA Layer

The DA layer has not seen significant technological iterations, so this track is not as competitive as the execution layer, with only a few main players.

ETH's CallData has been upgraded to Blob, leading to a significant decrease in fees across various Layer2s. Now, ETH has become a "less expensive" DA.

Celestia's greater role is that after its launch, it became the first project to propose the concept of the DA layer, pulling the DA track's ceiling from a $2 billion FDV to $20 billion, thus opening up the landscape and imagination. Many new Layer2 Appchains naturally prefer Celestia for DA. (FDV stands for "Fully Diluted Valuation," a valuation metric derived from Token price * total supply.)

Avail has spun off from Polygon and is technically more like an "enhanced version of Celestia," using the Polkadot consensus mechanism of Grandpa+BABE, which theoretically supports more decentralized nodes compared to Celestia's Tendermint. It also supports Validity Proofs that Celestia does not support, etc. However, the technical differences are far less important than the ecological aspects, and Avail still needs to catch up in the ecological layer.

EigenDA was also launched a few days ago alongside the EigenLayer mainnet. As one of the strongest narratives in this round + a project skilled in commercial collaboration, I personally feel that EigenDA's adoption rate will not be low. Theoretically, as long as it "feels safe and is inexpensive," not many projects genuinely care whether you use Validity Proof or Fraud Proof, or whether DAS is supported, etc.

Three DA projects worth mentioning are:

1) Near DA

Near is a magical public chain that was originally designed for sharding and is still doing so, but it is also working on DA. It is cheaper than Celestia and supports rapid settlement for Layer2;

Chain abstraction - Recently, Near launched chain signatures, allowing users to request signatures for transactions on any chain through a single NEAR account;

AI - Its founder, Illia, is one of the eight creators of Transformer and was patted on the shoulder by Jensen Huang at the Nvidia conference. They are currently planning to hire AI engineers, with an official announcement to be made next month… Hexagonal warriors, I also threw myself into the DA track.

2) BTC & CKB

Because BTC's Layer1 does not support smart contracts and cannot settle directly, dozens of BTC EVM Layer2s are essentially using BTC as DA. The only difference is whether they directly throw ZK Proof onto BTC or just throw the hash of ZK Proof, as if they cannot call themselves "BTC Layer2" without doing so.

Recently, I encountered a new project saying, "I’m not doing it anymore; I’m just ETH Layer2, and all DA settlements are on ETH, but I serve the BTC ecosystem!" Quite amusing… The only alternative scalability solution is RGB++ launched by CKB, where CKB becomes a quasi-DA existence, while BTC, due to the black technology of UTXO homomorphic binding, nearly becomes the settlement layer for RGB++.

3) New DA

I’ll mention two new DA ideas I’ve seen, without naming the projects. One combines DA with AI, serving not only as a high-performance DA but also as a storage layer for training data and training trajectories of AI large models; the other improves the error-correcting code mechanism underlying DA like Celestia, providing a more robust network state in a dynamic network (where a few nodes randomly drop out each round) under such unstable conditions.

Settlement Layer

Originally, this layer was almost exclusively occupied by ETH. DA has Celestia as competition, and execution has a host of Layer2s. Only the settlement layer remains, as other chains like Solana and Aptos do not yet have Layer2s, and BTC's Layer2 cannot use BTC for settlement. Currently, the only settlement layer you can think of is basically ETH.

However, this situation is about to change. Several new projects are moving in the direction mentioned at the beginning of the article, and some older projects are also beginning to pivot in this direction, that is, further deconstructing ETH through ZK verification/settlement layers (competing for ETH's business).

Why has this concept emerged? The reason is that running contracts on ETH Layer1 to verify ZK Proof is theoretically not an optimal choice:

From a technical perspective, to verify the correctness of ZK Proof, developers need to write verification contracts based on ZK projects and their chosen ZK Proof System using Solidity. This requires relying on several cryptographic algorithms, such as supporting different elliptic curves. These cryptographic algorithms are usually quite complex, and the EVM-Solidity architecture is not the optimal platform for implementing these complex cryptographic algorithms. For some ZK projects, the cost of writing and verifying these verification contracts is also very high. This somewhat hinders the native integration of some ZK ecosystems into the EVM ecosystem, so languages friendly to ZK like Cario, Noir, Leo, and Lurk can currently only verify on their own Layer1. Additionally, updating or upgrading these things on ETH is always "difficult to turn around."

From a cost perspective, although the "protection fee" paid on Layer2 is largely DA, ZK contract verification also requires gas fees, and verifying on Ethereum is certainly not a cheap option. Coupled with the fact that ETH gas fees occasionally soar, turning it into a "noble chain," verification costs will also be significantly affected.

Thus, new ZK verification/settlement layer concept projects have emerged, with new projects still relatively early, represented by Nebra. Older projects are also pivoting in this direction, such as Mina and Zen, which just passed a new proposal.

The overall thinking of most projects in this track is basically:

• Support for multiple ZK languages
• Support for ZK aggregated proofs, more efficient and cheaper
• Faster final confirmation times

The ZK settlement layer and decentralized Proof Market are likely to be tied together, as having technology requires computational power. We may see some settlement layer projects collaborating with Proof Market projects, or settlement layers with computational power directly creating their own Proof Market, or Proof Market projects with technology creating their own settlement layer. Ultimately, how this will unfold will be determined by the market.

Conclusion

In other areas of infrastructure, such as Oracle and MEV fields like OEV, and interoperability fields like ZK light clients, there should be plenty of articles online discussing them, so I won’t elaborate further. Next time I see some new and interesting developments, I will share them with everyone.