Davide Crapis, the AI lead at the EF, argues that the motivation is as philosophical as it is technical.
As artificial intelligence reshapes everything from finance to cybersecurity, the Ethereum Foundation (EF) is carving out a strategy for how the world’s second-largest blockchain fits into that future.
Instead of trying to fuse blockchains and AI at the level of raw computation — something Ethereum was never designed to handle — the EF sees the network playing a different role: acting as a coordination and verification layer in an increasingly AI-mediated world.
Davide Crapis, the AI lead at the EF, argues that the motivation is as philosophical as it is technical. More and more digital activity is being handled by AI systems, whether it’s answering questions, executing trades, screening applications or writing software. If those systems are controlled by centralized entities, the values that underpin much of the crypto movement — decentralization, self-sovereignty, censorship resistance and privacy — could erode.
“If AI doesn’t have the properties we care about — self-sovereignty, censorship resistance, privacy — and then we use AI for everything, basically no one has those properties anymore,” he said to CoinDesk in an interview at NEARCON 2026.
In that sense, Ethereum’s AI push is less about competing with OpenAI or Google on model size and more about ensuring that as AI becomes the interface to the internet, it doesn’t quietly recentralize power.
The EF’s strategy rests on two broad fronts. The first is what Crapis calls decentralized AI coordination. As autonomous AI agents — software programs capable of carrying out tasks on their own — become more common, they will need ways to identify themselves, build trust and exchange payments. Ethereum, he argues, is well-suited to provide that infrastructure.
“Ethereum functions as a public, governance-less verification layer for AI,” he said.
In practical terms, that means the heavy computing work of AI remains off-chain, on traditional servers. But Ethereum can help agents discover one another through public registries, assess reputation through transparent histories, route payments and anchor cryptographic proofs that verify outcomes. Crapis likens it to a decentralized version of Google Reviews combined with payment rails.
The EF has been involved in developing standards to formalize this ecosystem, including a protocol for agent identity and trust, known as ERC-8004. According to Crapis, these standards are gaining traction beyond Ethereum, signaling that the coordination layer for AI agents may become blockchain-based even if the AI itself is not.
The second focus area centers on bringing Ethereum’s core principles — such as privacy, openness, censorship resistance, and security — into the world of AI. Crapis refers to this effort internally as “Props AI,” shorthand for the values the Ethereum ecosystem has historically prioritized.
Privacy is a major part of that conversation. Interacting with centralized AI services can gradually generate detailed user profiles based on queries, usage patterns and behavior.
From Ethereum’s perspective, the challenge is to design AI systems that allow users to retain greater control over their data and identity. One approach is to encourage more AI processing to occur locally on users’ devices whenever possible, reducing the amount of information that needs to be sent to centralized servers.
The broader goal is to ensure that as AI becomes embedded in everyday digital interactions, individuals still retain meaningful control over their data and how it is used, rather than handing that power entirely to large platforms.
“We want to create a world where users retain as much data and power as possible,” Crapis said. “We just don’t give it to operators.”
Security concerns also underpin the strategy. As AI systems grow more capable, they are likely to automate and scale cyberattacks in ways that strain existing defenses. Crapis predicts a near future in which AI systems can convincingly impersonate humans, undermining traditional authentication methods.
“We will probably see hacks orchestrated by AI,” he said. “The old security models break when AI can impersonate a human.”
In that environment, cryptographic keys may become more important. Control of a private key is mathematically verifiable and does not depend on human judgment. Crapis frames Ethereum’s long-term role in stark terms.
“In a world where AI is in the wild, we want Ethereum to be the place with the big lock,” he said. “If I have the keys, I still have power.”
Crapis described the AI initiative that the EF is doing as one of several major priorities rather than the dominant one. Still, the move reflects a growing recognition within the crypto industry that AI will shape the next phase of the internet. If that future is mediated by intelligent agents rather than human clicks, the question becomes who controls the rails those agents run on.
Ethereum’s bet is that even if it doesn’t power the brains of AI, it can help govern the environment in which those brains operate, anchoring identity, coordinating payments and preserving user control.
