Optimistically Trust: EigenLayer AVS will lead the Web3 privacy computing track with "cheap security."

Author: @Web3Mario

Introduction: EigenLayer AVS has been online for a while now. Aside from its long-promoted official guides like EigenDA and Layer2, I have discovered a very interesting phenomenon: EigenLayer AVS seems to be highly attractive to projects in the privacy computing sector. Among the nine AVS that have already launched, three belong to this sector, including two ZK co-processor projects, Brevis and Lagrange, and a trusted execution environment project, Automata. Therefore, I decided to conduct a detailed investigation to explore the significance of EigenLayer AVS for related products and the future development trends.

The Attraction of "Cheap Security" is Key to the Success or Failure of the EigenLayer AVS Ecosystem

With the TVL officially exceeding $15 billion, EigenLayer has had a very dreamy start. Of course, I believe that most of the funds are aimed at obtaining potential airdrop benefits, but this undoubtedly lays a solid foundation for EigenLayer to enter the next stage. The key to the next stage lies in the success or failure of the AVS ecosystem, as the scale of AVS's fee income determines the timing of EigenLayer's transition from the subsidy phase to the maturity phase.

Many articles have already introduced the technical details related to EigenLayer, so I will not elaborate further. In simple terms, EigenLayer creates a cheap consensus layer protocol by reusing Ethereum's PoS consensus capability, known as Restaking. First, I would like to discuss the core value of EigenLayer, which I believe has three main aspects:

* Decoupling the consensus layer from the execution layer allows for better handling of large-scale or high-cost data processing and consensus: Generally, mainstream blockchain protocols are considered solutions with high execution costs but low execution efficiency. The high execution cost is due to "competition for block space," a trendy term. We know that blockchain-based execution environments typically use market mechanisms to adjust the allocation of node computing resources, meaning that those who bid higher get priority for execution. The executors are in a competitive relationship, and when demand rises, fair prices will continue to climb, leading to higher execution costs. The low execution efficiency stems from the fact that blockchain technology was originally designed to be a settlement system for electronic currency, making the processing of transaction data time-sensitive. Therefore, a serial approach was adopted for designing the execution layer, which makes it inefficient in handling most time-insensitive scenarios, such as social networks and AI training.

Decoupling the consensus layer from the execution layer allows application developers to design dedicated execution environments, often referred to as application chains or Layer3, freeing users from competition with other application users and reducing usage costs. On the other hand, it enables developers to create more suitable execution layers based on different application scenarios, improving execution efficiency.

* Consensus as a service, fully exploring the potential market demand by productizing or resourceizing consensus: I think that anyone who has experienced the era of a hundred flowers blooming in Layer1 will have a common sentiment: the establishment of a consensus layer is usually expensive and difficult. Each party must maintain its own consensus security, which may involve computing power or staked funds, and they remain in a subsidy phase before generating sufficient profitability, with costs that cannot be considered low. Typically, the subsidized object is the token rewards obtained from mining. Only a few successful protocols can transition to relying on their own revenue capabilities, namely fee income, to maintain sufficient consensus capability, such as the transformation of Ethereum's economic model. This high startup cost deters many innovative applications, as establishing a suitable execution environment for their applications or building an application chain incurs excessive costs and faces significant risks. This has made the Matthew effect in the Web3 industry very evident, as the evolution of current Web3 technical solutions has been largely dictated by Ethereum's technical route.

By productizing or resourceizing consensus, innovative applications have another option: to purchase consensus services based on demand. For example, for an innovative application, if the total amount of funds managed by the application in the early stage is $1 million, it means that as long as it purchases more than $1 million in PoS consensus, it can ensure the security of its execution environment, as the economic cost of malicious behavior is negative. As the application develops, it can flexibly purchase consensus services in a quantitative manner. This reduces the startup costs for innovative applications, lowers their risks, and fully explores market potential.

* Cheap sources of consensus: Finally, EigenLayer's consensus source utilizes the reuse of Ethereum's PoS funds, which means that for PoS stakers who could originally only capture one layer of rewards, participating in EigenLayer allows for an additional layer of rewards. This cleverly transforms the relationship between EigenLayer and the industry leader Ethereum from competition to symbiosis, reducing the cost of attracting consensus funds. This also gives it an advantage in pricing, such as the consensus purchase fees of the AVS protocol, making it more attractive to innovative applications. It must be said that this is indeed a brilliant strategy.

These three points allow EigenLayer to provide "cheaper security" for Web3 applications compared to other Web3 execution environments, resulting in lower execution costs, better scalability, and more flexible business models. Therefore, I believe the key to the vibrancy of the EigenLayer AVS ecosystem lies in whether Web3 applications can be attracted by this cheap security and migrate in large numbers to this ecosystem.

The Cost of Use is the Fundamental Reason Restricting the Development of the Web3 Privacy Computing Sector

After discussing the core value of EigenLayer, let's take a look at the dilemmas faced by the Web3 privacy computing sector. I am not an expert in this field, so I focused on studying the current state of projects related to privacy computing among the launched AVS. This refers to the so-called ZK co-processors. I believe that most cryptographic products utilizing zero-knowledge proof algorithms face the same dilemma: high usage costs hinder the promotion of usage scenarios.

The concept of ZK co-processors seems to have become less important. As the name suggests, the original intention of products in this sector is to use zero-knowledge proof algorithms to provide co-processor services for current mainstream blockchain systems, allowing them to offload complex and expensive computational operations to be executed off-chain while ensuring the correctness of execution results through zero-knowledge proofs. The most classic example of this modular thinking is the relationship between CPUs and GPUs. By delegating parallel computing operations, such as image processing and AI training, which CPUs are not good at, to another independent module, the GPU, execution efficiency is improved.

A classic ZK co-processor project's technical architecture is roughly as follows. This is a simplified technical architecture of Axiom, one of the leaders in this sector. In simple terms, when a user has a demand for complex computation, they can use Axiom's off-chain service to compute results and generate related ZK Proofs. Axiom will then call the on-chain verification contract with the results and proofs as parameters. This contract verifies the correctness of the results using three pieces of data: the execution results, the execution proofs, and the key information of the entire chain provided by Axiom to the on-chain, such as the transaction merkle root (the process of maintaining the key information of the entire chain is also trustless). After verification, the results will be notified to the target contract through a callback function to trigger subsequent operations.

It is generally believed that the proof generation process is compute-intensive, while the verification of proofs is relatively light. According to Axiom's documentation, a single on-chain ZK Proof verification operation roughly requires a verification gas fee of 420,000. This means that assuming the gas price is 10 Gwei, the user needs to pay 0.0042 ETH for verification costs. Assuming the market price of ETH is $3000, the cost would be around $12. Such costs are still too high for ordinary C-end users, greatly limiting the potential use scenarios for this product.

Referring to a frequently promoted use case of a ZK co-processor project, the Uniswap VIP program, Uniswap can use ZK co-processors to set up a loyalty program similar to CEX for its traders. When a trader's cumulative trading volume reaches a certain level over a period, the protocol offers rebates or reductions on fees for that trader. Considering that calculating cumulative trading volume is a complex operation, Uniswap can adopt the ZK co-processor solution to offload the computation off-chain, reducing computational costs while avoiding large-scale modifications to the on-chain protocol.

Let's do a simple calculation. Suppose Uniswap sets a VIP event where users can enjoy zero fees if they can prove that their cumulative trading volume exceeds $1,000,000 in the past month. If a trader chooses to trade in a Uniswap pool with a 0.01% fee and their single transaction volume is $100,000, the fee would be $10. However, the verification cost would be $12, which discourages the user from participating in this service, raising the threshold for participation in the event, ultimately benefiting only the whales.

Similar cases should not be hard to find in related pure ZK architecture products. The use cases and technical architectures are excellent, but I believe that the cost of use is the core constraint hindering the expansion of usage scenarios for related products.

The Impact of EigenLayer's "Cheap Security" on Related Products as Seen from Brevis's Transformation

Now, let’s take a look at how Brevis, one of the first AVS launched, has been influenced by EigenLayer. I hope to illustrate that EigenLayer's "cheap security" has a significant attraction for related cryptographic products.

The core team of Brevis comes from an old star project, Celer Network, consisting of a group of talented Chinese technologists. After some struggles, they launched Brevis in early 2023, positioning it as a ZK full-chain data computation and verification platform. Of course, this is essentially no different from a ZK co-processor, but the latter sounds cooler. For a long time, Brevis operated using the aforementioned so-called "Pure-ZK" solution. This made it challenging to promote usage scenarios. However, on April 11, they announced a partnership with EigenLayer and a new "crypto-economics + ZK proof" solution called Brevis coChain. In this solution, the verification layer has been moved from the Ethereum mainnet to a coChain maintained by AVS.

When users generate computational demands, they compute results through client-side circuits and generate related ZK Proofs, then send computation requests to Brevis coChain via on-chain smart contracts. Upon receiving the request, AVS verifies the correctness of the computation and, once verified, packages the relevant data for some compression processing before sending it to the Ethereum mainnet, asserting the correctness of the results. For a period afterward, similar to other "optimistic verification" solutions, it will enter a challenge period, during which challengers can submit corresponding ZK fraud proofs to dispute a result and seek penalties for the wrongdoer. After the penalty period, AVS will complete subsequent operations using the callback of the target contract through on-chain contracts. Considering that most privacy computing topics focus on how to achieve trustlessness through mathematics, I would like to refer to this solution as "optimistic trustlessness."

Similarly, Lagrange and Automata must have gone through a similar mental journey before introducing optimistic trustless solutions utilizing AVS. The benefit of this solution is that it significantly reduces verification costs. Because the costly on-chain verification computation is no longer needed to obtain correct results, it instead relies on the optimistic trust in the processing results of EigenLayer's consensus layer and the security provided by ZK fraud proofs. Of course, transitioning from trust in mathematics to trust in human nature will face some challenges in the Web3 field. However, I believe this is an acceptable outcome compared to the practicality it brings. Moreover, this solution will effectively break the constraints of verification costs on the promotion of usage scenarios. I believe that many interesting products will be launched soon. Additionally, this solution will create a demonstration effect for other products in the privacy computing sector. Considering that this sector is still in a blue ocean stage, it should be more conducive to the promotion of new paradigms compared to the fiercely competitive rollup-related sectors. I believe the AVS ecosystem will be the first to witness an explosion in the privacy computing sector. As I am not an expert in the relevant cryptographic field, there may be some inaccuracies in my writing, and I hope experts will provide corrections.