Did the Ethereum account abstraction fail?

Written by: Haotian

Last time I mentioned that the x402 protocol extends the Lightning Network, and recently while having dinner with a group of programmer friends, I was once again “challenged”: isn't x402 just the previous AA account abstraction?

The subtext is that Ethereum has been working on Account Abstraction for many years, investing a lot of resources into ERC-4337, Paymaster, and various grants and wallet service providers, but the actual results are evident to everyone, as many people criticize it for making a lot of noise but having little impact.

Although I don't think AA has declared failure, what is the crux of the issue?

1. Paymaster shifts the user's gas consumption onto the project party, which sounds great, but the project party's ability to cover costs is weak, the ROI is unclear, and it undoubtedly leads to a dead end in the business model. How can it survive relying solely on external funding without the ability to generate its own revenue?

2. AA account abstraction is limited to the EVM ecosystem, such as ERC4337, Paymaster, and EntryPoint contracts, which are all exclusive to Ethereum. If you want to achieve cross-EVM ecosystem usage that includes Solana, BTC, etc., you will need to continue adding intermediary layer services to realize this functionality. However, the problem is that these intermediary layer services introduce additional fees to be divided, thus posing a greater challenge to the ROI of the business model!

There are still many complex technical issues, and I won't elaborate on them, but let me say something that everyone can understand. AA is essentially a product of “technology for the sake of technology” and is a work stemming from the pure research trend of Ethereum in the past.

In contrast, what does the x402 protocol play? What are the differences? Some criticize how the ancient HTTP 402 status code, which has been around for 30 years, is being brought out, along with the game of gilding gold.

But don't forget, the HTTP 402 status code — this is the underlying protocol of the internet, the common language of Web2 and Web3.

AA requires smart contracts, on-chain state, and EVM virtual machine execution, while x402 only needs an HTTP request header, making it compatible with any system that supports HTTP—Web2 APIs, Web3 RPCs, and even traditional payment gateways.

This is not an optimization plan for technological stacking, but rather a “dimensionality reduction strike” that simplifies the protocol layer. Instead of fussing over various compatibility adaptations and trust methods at the application layer, it is better to first unify the standards at the top-level protocol layer.

The key is that x402 is naturally a very good cross-chain interoperability standard. As long as the Agent can send HTTP requests, can handle 402 responses, and can complete EIP-3009 authorization (or the equivalent standard of other chains), whether you are Base, Monad, Solana, Avalanche, or BSC, the protocol-level cross-chain is seamless and only manifests in the single point issue of settlement payments, making the cross-chain cost much lower in comparison.

Facilitator can serve multiple chains simultaneously, and users' payment history data can be indexed uniformly, allowing developers to connect once to “link” the entire ecosystem.

Overall, I feel that AA is a refined engineering product born from the mindset of researchers, while the x402 protocol is a pragmatism driven by market demand.

The question arises, will ERC-8004 follow the old path of AA?

Theoretically speaking, ERC-8004 is very similar to AA 2.0, still exclusive to EVM, requiring the deployment of a three-layer registry (Identity/Reputation/Validation). Early incentives also rely heavily on external subsidies or staking, which are pitfalls that AA has encountered before. For other chains to be compatible, an additional layer of trust cost must be added.

But the difference is that under the x402 framework, ERC-8004 is just a tool, not a leading standard. Other chains need to be compatible with the x402 protocol, not ERC8004.

This positioning difference is very important. What was the problem with AA back then? It wanted to become “the only standard for Ethereum payment experience,” requiring the entire ecosystem to revolve around it: wallets needed to adapt, applications needed to integrate, and users needed to change their habits. This kind of “top-down” strong push, without a killer application and clear ROI, naturally cannot be promoted.

Unlike ERC-8004, it does not need to be the main character, because x402 has already solved the most fundamental issue: payment. ERC-8004 simply provides an “optional” trust layer on this already functioning payment network.

Moreover, ERC-8004 rides on the coattails of x402, eliminating the need to build an ecosystem from scratch. x402 already has a clear business closed loop (Provider traffic generation, Facilitator charging), a complete technical stack (HTTP protocol + EIP-3009), and an active project ecosystem, so ERC-8004 only needs to be “plug and play.”