Chainfeeds Summary:
The infrastructure phase is nearing completion, and the application phase has begun. The transition from protocol-ready to product-ready will occur within the next two to three months.
Article source:
https://www.odaily.news/zh-CN/post/5209193
Article Author:
EH
Opinion:
EH: In January, the proxy economy achieved three key breakthroughs almost simultaneously in a very short period, marking the first time that the underlying infrastructure had formed a closed loop. OpenClaw's GitHub stars surpassed 100,000, attracting over 2 million developers within a week, providing proxies with a truly usable task execution and browser control environment; Moltbook, as the first "AI-only social network," officially launched, gathering 1.2 million proxy identities in its first week; and ERC-8004 went live on the Ethereum mainnet on January 29, with contributors covering wallets, foundations, cloud services, and trading platform members, completing the on-chain trust layer for proxies. The framework, social, and trust layers were completed simultaneously, enabling proxies to have the basic conditions for independent operation for the first time. At the same time, x402 found a sustainable pricing equilibrium range, with nearly 90% of services currently setting prices between $0.01 and $0.10, a sweet spot where stablecoin settlement costs are significantly lower than credit card fees. As the market converges towards a micropayment economy, the average price of x402 dropped from $0.81 to $0.29 within a month, with over 20 million transactions completed. The payment track, requiring no API key and natively supporting HTTP, has been successfully implemented, marking the first time that the agent business has been established both technically and economically. The key innovation of ERC-8004 lies in decomposing "trust" into composable on-chain modules. The protocol consists of three registries: an identity registry based on ERC-721, providing portable, censorship-resistant agent identities; a reputation registry recording feedback from each interaction; and a verification registry supporting multiple trust models, from staking to zero-knowledge proofs. Currently, over 30,000 agents have registered on the mainnet, and the trust infrastructure already exists. The real uncertainty lies in the speed of its large-scale adoption. Superficial data shows a significant decline in both the number of transactions and transaction volume recently, but deeper analysis reveals a typical structural cleansing. On-chain analysis shows that nearly half of the early transaction volume came from non-natural wash trading; after removing this, the decline in genuine demand significantly narrowed. Meanwhile, the ratio of buyers to sellers has improved significantly, the number of service providers has decreased, but individual services are serving more real users. This is not a decline, but a process of "quality replacing quantity": speculation and incentives drive exits, real demand begins to emerge, laying a more solid foundation for subsequent application layer construction. The biggest opportunity in the current agent economy is not on the supply side, but on the demand side and the connection layer. There are already more than 1,500 independent service sources on the supply side, while the demand side has gathered about 1.2 million active agents. There are three obvious gaps between the two: the lack of a unified cross-platform search and discovery mechanism; the lack of objective capability benchmarks to prove whether agents "can do things well"; and the lack of a trust gating mechanism that connects on-chain trust verification and payment execution. The protocol is in place, but the product layer has not yet appeared. In reality, agents still need to query multiple platforms separately to find services, resulting in fragmented interfaces and inefficient distribution. The real opportunity lies in building a unified index layer to achieve cross-platform search, real-time availability, and price comparison, forming an "agent version of the App Store." At the same time, trust also needs to complete the other half: not only verifying whether payments are reliable, but also verifying whether it has the ability to complete complex tasks. Predictive markets and verifiable scenarios are becoming ideal platforms for capability benchmarking. On the payment side, approximately 20 million transactions per month are executed without any trust verification, creating significant leverage potential for "trust-gated payment middleware." The infrastructure phase is ending, the application phase has begun, and the window of opportunity is clearly open to builders.
Content source 
