Ethereum Joins the Race to Become the AI Economy’s Base Layer

The Ethereum Foundation is formally entering the artificial intelligence race with the launch of a new unit, the “dAI Team,” that aims to make Ethereum the “preferred settlement and coordination layer” for AI agents and the machine economy. The move comes as other networks, like NEAR Protocol and Solana, push similar efforts to develop blockchain-based infrastructure for the AI economy.

In an X post on Monday, Sept. 15, dAI Team lead Davide Crapis, previously a researcher at the Ethereum Foundation, said that the initiative will focus in two directions: First, building tools that let AI agents and bots transact and follow rules without intermediaries; Second, developing a decentralized AI stack so that future AI development doesn’t depend on a small number of centralized companies.

Part of the effort centers on ERC-8004, a new Ethereum standard and so-called “trust layer” in the works that’s meant to give AI agents something like an on-chain reputation score.

“This standard extends the Agent‑to‑Agent (A2A) protocol with a trust layer that allows participants to discover, choose, and interact with agents across organizational boundaries without pre‑existing trust,” Crapis explained in a forum post describing the proposed standard.

With ERC-8004, instead of blindly trusting a bot, users would be able to check an agent’s identity, see how it’s behaved in the past, and confirm whether it meets certain requirements through registries stored on Ethereum. As Crapis added in the post, the Ethereum Foundation will work closely with the Linux Foundation and A2A ecosystem stakeholders to “refine and improve the specifications of this extension.”

Nick Johnson, lead developer at ENS, a protocol for blockchain-based domains, told The Defiant that the urgent challenge with AI is to develop “decentralised standards for agent identity and trust so autonomous systems remain accountable and interoperable.”

“In a future defined by AI agents transacting and collaborating on-chain, technologists must ask how signature protocols like those underpinning ENS can set the benchmarks for integrated identity and reputational transparency, rather than let centralised approaches dominate,” Johnson said in comments to The Defiant.

Joining the Pack

By creating a dedicated team to focus on decentralized AI and the future AI agent economy, the Ethereum Foundation joins other established players already working to become the go-to network for the decentralized AI economy.

NEAR Protocol has branded itself “the blockchain for AI,” developing frameworks like Shade Agents, which lets autonomous bots act across chains while keeping data private.

Alex Shevchenko, a former NEAR Protocol product manager and currently CEO of Aurora Labs, told The Defiant that there are “two intriguing use cases for AI on blockchain,” transferring value and research:

Solana is also equipping developers with an Agent Kit that allows them to plug AI models directly into DeFi and on-chain apps, enabling AI agents to autonomously manage trades, liquidity, and other DeFi and web3 activities, including NFT creation and lending.

Unlike its competitors, SingularityNET, a blockchain protocol also explicitly focused on decentralized AI, has taken the unification route, teaming up with Fetch.ai and Ocean Protocol to launch the Artificial Superintelligence Alliance. The Alliance is a shared AI playground where the three projects pool resources and give developers a single ecosystem to deploy AI without relying on centralized players.

Outside of crypto-native solutions, tech giant Google just yesterday revealed its own payments protocol for AI agents, AP2, from Google Cloud, promising a combination of TradFi payment methods and stablecoins. Google Cloud’s crypto partners in the venture include Coinbase, as well as MetaMask and the Ethereum Foundation itself.