AI x Web3 Right Now: Five Signals Founders Are Watching

Frequently Asked Questions

01.Is ZKML production-ready for on-chain use cases in 2026?

ZKML, or zero-knowledge machine learning, allows a protocol to verify that a machine learning model produced a specific output without revealing the model weights or input data. As of April 2026, projects like EZKL and Modulus Labs have demonstrated verifiable inference on smaller models, but latency and proof generation cost remain barriers for real-time on-chain settlement. The technology is production-adjacent, not yet production-standard for most DeFi use cases.

02.Does the EU AI Act apply to Web3 AI agents?

The EU AI Act entered phased enforcement starting February 2025, with full high-risk AI system obligations applying from August 2026. Autonomous AI agents making financial decisions in DeFi protocols could fall under high-risk or limited-risk classifications depending on their decision scope. MiCA does not explicitly cover autonomous agent behavior, creating a regulatory gap that compliance teams at Web3 companies need to monitor actively.

03.What is ERC-8004 and why does it matter for AI agents on Ethereum?

ERC-8004 is a proposed Ethereum standard that defines a permission framework for trustless AI agent wallets, allowing autonomous agents to hold and spend on-chain assets within defined capability scopes. It matters because it creates a verifiable, audit-friendly boundary for what an AI agent can do without human approval, reducing the risk of unconstrained autonomous spending on Ethereum mainnet.