In the current overheated environment of AI, let’s take a moment to look at the changes that have occurred in the infrastructure layer, especially in the public chain infrastructure, in recent months. There are some new things worth mentioning.
(Background information:
Forbes names XRP, ADA, LTC, and other top 20 public chains as “crypto zombies” with little use besides speculation.
)
Table of Contents:
ETH, or the further deconstruction of a single chain
Execution layer
1) From the perspective of how transactions can be executed in parallel – there is nothing new under the sun, just a distinction between a priori and a posteriori.
2) From the emphasis on parallel EVM – it can also be divided into two schools of thought.
DA layer
1) Near DA
2) BTC&CKB
3) New DA
Settlement layer
Conclusion
Recently, in the primary market, the hottest track is undoubtedly AI, followed by BTC. 80% of the projects discussed every day are concentrated in these two tracks. At one point, I could discuss 5 or 6 AI projects in a day.
It can be foreseen that the AI bubble will reach its peak in the coming years. With hundreds of new AI projects going online, the market value of the AI track will reach its peak. When the bubble bursts and everything falls apart, true unicorns that have found the intersection of AI and Crypto will emerge, pushing this track and the entire industry forward.
So, in the current overheated environment of AI, let’s take a moment to look at the changes that have occurred in the infrastructure layer, especially in the public chain infrastructure. Some new things are still worth mentioning.
When Celestia first proposed the concept of modularity and the concept of the DA layer, the market actually took a lot of time to digest and understand. Now, this concept has become well-known, and various RaaS infrastructures have flooded the market, with the number of infrastructures surpassing the number of applications and users. (RaaS stands for Rollup-as-a-Service, which refers to providing ready-to-use Rollup products and services to help application developers quickly launch Rollups.)
The execution layer, DA layer, and settlement layer have made some different technological advancements in the past few months, with each layer spawning 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.
The most popular concept in the execution layer is undoubtedly parallel EVM, represented by Monad, Sei, MegaETH, and supported by projects like FTM and Canto. However, just as not all ZK projects protect privacy, projects labeled as parallel EVM have different technical roadmaps and end goals.
Taking Sei’s diagram as an intuitive display, it is clear that in an optimistic scenario, the performance improvement from sequential processing to parallel processing is significant.
Parallel EVM can be divided into several different technical approaches:
1) Pre-verification, represented by Solana and Sui, requires transactions to explicitly declare which parts of the chain state they modify, allowing for pre-detection of conflicts (such as access to the same AMM pool) before block packing. Conflicting transactions causing these conflicts are then discarded.
2) Post-verification, also known as optimistic parallelism, represented by Aptos BlockSTM, assumes that there are no conflicts and includes transactions before conducting checks. If conflicts are detected, the transaction is declared invalid, the result is refreshed, and the process is repeated until all transactions in the block are executed. Sei, Monad, MegaETH, and Canto employ similar solutions.
In the primary market, we have also seen parallel solutions for situations involving state conflicts (such as accessing the same AMM pool). However, these solutions appear to be more complex and it is uncertain whether they are commercially viable, and are still under evaluation.
One approach is represented by Monad and Sei, which consider parallelism as the main scalability approach. In addition to optimistic parallel processing, Monad also has MonadDB, specially developed for parallel processing, and asynchronous I/O to complement parallel processing.
The other approach is taken by Fantom, Solana, and MegaETH, where parallelism is one of the scalability solutions but not the only one. Parallelism is a supplementary narrative, with performance improvements relying more on other technical solutions.
For example, Fantom’s Sonic upgrade focuses on the FVM virtual machine and the optimized Lachesis consensus mechanism. Solana’s next stage focuses on the modular architecture of the Firedancer new client, optimized network communication mechanisms, signature verification, and more.
MegaETH aims to achieve a Realtime Blockchain. Based on the efficient client Paradigm’s Reth, it further optimizes and improves the full-node state synchronization mechanism (synchronizing only state differences rather than all data), the hardware design of the sequencer (using large amounts of high-performance RAM with storage capabilities for state access, avoiding slow disk I/O), and the improvement of the Merkle Trie data structure. It comprehensively enhances software, hardware, data structure, disk I/O, network communication, transaction ordering, and parallel processing, pushing the performance ceiling of EVM to the limit, approaching “Realtime Blockchain.”
The DA layer has not seen significant technological iteration, so the level of hype in this track is far less than that of the execution layer. There are only a few main players in this field.
ETH’s CallData upgrade to Blob has significantly reduced the fees for various Layer2 solutions, and now ETH is considered a “not so expensive” DA.
Celestia’s biggest contribution is being the first project to propose the concept of the DA layer, raising the ceiling of the DA track from a $2 billion FDV to $20 billion, opening up new possibilities and imagination. Many new Layer2 Appchains naturally prefer Celestia as their DA. (FDV stands for “Fully Diluted Valuation,” an evaluation index derived from token price multiplied by total supply.)
Avail has separated from Polygon and is technically more like an “enhanced version of Celestia.” For example, it uses the Grandpa+BABE consensus mechanism from Polkadot, which theoretically supports more decentralized nodes than Celestia’s Tendermint. It also supports Validity Proof, which Celestia does not. However, the technical differences are not as important as the ecosystem, and Avail still needs to catch up in terms of the ecosystem.
EigenDA went online a couple of days ago along with the EigenLayer mainnet. EigenLayer, as one of the strongest narratives in this round and a project that excels in business cooperation, I personally feel that the adoption rate of EigenDA will not be low. In theory, as long as it is “perceived as secure and cheap,” not many projects really care whether you use Validity Proof or Fraud Proof, or whether DAS is supported.
What is worth mentioning are the following three DAs:
1) Near is a magical public chain that was originally designed for sharding, and it still is, but it has also developed DA. It is cheaper than Celestia and supports fast settlement on Layer2.
2) Chain Abstraction – Near recently introduced Chain Signature, allowing users to request transaction signatures for any chain through a single NEAR account.
3) AI – Founder illia, who is part of the Transformer Octet, was tapped on the shoulder by Huang Renxun at NVIDIA’s conference. Now, they are planning to hire AI engineers, and relevant announcements will be released on their official website next month… Hex Warrior, I also threw it into the DA race.
Because BTC’s Layer1 does not support smart contracts and cannot settle directly, almost all BTC EVM Layer2 solutions are used as DAs. The only difference lies in whether ZK Proofs are directly applied to BTC or whether the hash of ZK Proofs is applied. It’s as if you can’t call yourself a “BTC Layer2” without doing this.
Recently, I actually encountered a new project saying, “I won’t install it. I am an ETH Layer2, and DA settlements are done on ETH, but I serve the BTC ecosystem!” It’s quite amusing… The only alternative to the expansion solution is RGB++ introduced by CKB, where CKB becomes a DA-like entity due to the UTXO-based binding technology, making BTC a de facto settlement layer for RGB++.
Two new DA approaches are worth mentioning, without mentioning the project names. One is the combination of DA and AI, where it serves as a high-performance DA and also acts as a storage layer for AI models, training data, and training trajectories. The other is an improved fault-tolerant mechanism at the underlying layer of Celestia and other DAs, providing more robust network states in an unstable environment with randomly dropped nodes in each round.
Originally, this layer was almost monopolized by ETH, with Celestia competing in the DA field, and the execution layer having a plethora of Layer2 solutions. The only exception was the settlement layer, where ETH was the only option.
However, this situation is about to change. I have already seen several new projects moving in the direction mentioned at the beginning of the article. Some old projects have also started pivoting towards this direction, namely the further deconstruction of ETH through ZK verification/settlement layers (competing for ETH’s business).
Why did this concept emerge? The reason is that running contracts on ETH Layer1 to verify ZK Proofs is not necessarily the optimal choice:
1) Technically, in order to verify the correctness of ZK Proofs, developers need to write verification contracts based on the ZK projects and the chosen ZK Proof System, using Solidity. This requires reliance on various cryptographic algorithms, such as supporting different elliptic curves.
These cryptographic algorithms are usually complex, and the EVM-Solidity architecture is not the optimal platform for implementing these complex cryptographic algorithms. For some ZK projects, the cost of developing and verifying these verification contracts is also high. This to some extent hinders the native integration of some ZK ecosystems into the EVM ecosystem. Therefore, ZK-friendly languages like Cario, Noir, Leo, and Lurk can currently only verify on their own Layer1. At the same time, updating or upgrading these things on ETH is always “difficult for a big ship to turn around.”
2) In terms of cost, although the majority of fees on Layer2 are accounted for by DAs, verifying ZK contracts also incurs gas fees, which are definitely not a cheap option on Ethereum. Moreover, the cost of verification is also greatly affected by the fluctuating ETH gas fees, which sometimes turn it into a “premium chain.”
As a result, the concept of new ZK verification/settlement layer projects has emerged. These new projects are still relatively early-stage, with Nebra as a representative. There are also old projects pivoting in this direction, such as Mina, and Zen, which recently passed a new proposal.
The overall approach of most projects in this track is as follows:
1) Support multiple ZK languages.
2) Support ZK aggregation proof for higher efficiency and lower cost.
3) Faster final confirmation time.
ZK settlement layers and decentralized Proof Markets are likely to be closely linked, as having technology also requires computing power. We may see some settlement layer projects collaborating with Proof Market projects, or settlement layer projects that control computing power directly creating their own Proof Market, or Proof Market projects that control technology taking on the role of a settlement layer. The market will ultimately determine the path.
Other areas of infrastructure, such as Oracle and MEV in the OEV field, and ZK lightweight client in the interoperability field, have been covered in various articles and will not be discussed further here. I will share more interesting things next time I come across them.