VeChain co-founder and CEO Sunny Lu says blockchain is the "perfect solution" to AI agent trust. He said so on Cointelegraph's Chain Reaction show while promoting VeWorld.ai, an agent-building platform that runs on VeChain. The take is worth reading. The pitch is also the product, and that matters when you weigh it.
What Lu actually said
According to Cointelegraph's write-up, Lu's argument runs like this:
- ▸Autonomous agents that operate around the clock will need a "new accountability system" before the public trusts them.
- ▸His answer: "The perfect solution is a blockchain."
- ▸On VeWorld.ai, agents would carry digital "passports" and sit inside a "decentralized credibility system" that tracks the reputations of both agents and their creators.
- ▸Smart contracts could settle transactions between agents, or between agents and people.
Cointelegraph describes VeWorld.ai only as an agent-building platform that uses VeChain's blockchain. The article gives no details on its features, users, or results.
The other half: agents as small businesses
Lu also pitched agents as a way for non-developers to package expertise into services. He compared it to mobile app stores, which let individual developers distribute and monetize their work. His examples were personal: an educational agent for his daughter, an agent supporting his mother's daily routine and recovery, and a "virtual employee" handling purchasing, inventory, sales, and operations for an acquaintance's shop in Milan.
That is a marketplace thesis: people with specialist or local knowledge build an agent, and others use it. It is a reasonable thesis. It is also a claim, not a shipped result. The article presents it as Lu's vision, without independent verification.
Where we agree, and where we want a receipt
Agree: an agent that holds a wallet and acts without a human in the loop needs an identity other parties can look up, and a record of what it has done. If an agent can move funds, "trust me" is not a security model. Lu is pointing at the right problem.
Receipt, please: "blockchain" is a substrate, not a trust system. Writing an agent's passport onchain tells you the record exists and who wrote it. It does not tell you the agent behaves well. A reputation system is only as good as what feeds it: who can write to it, what it measures, and whether a creator can walk away from a bad record and start fresh. Lu's description, as reported, does not cover those questions. That is not a knock on the team. The article is an interview, not a spec.
Also worth noting: the main source for this story is the CEO of the company promoting the platform. Nothing here has been checked against a live product. Self-reported is not verified.
What a builder can take from this
Ignore the "perfect solution" framing and keep the checklist. If you build an agent that touches money, you need answers to four questions:
1. Identity. Can another party look up who or what this agent is, without asking you? 2. Reputation. What record exists, who writes it, and can it be gamed or reset? 3. Settlement. How does it pay and get paid? Which rails, which wallet controls, which limits? 4. Evidence. What can an outsider check today, versus what is promised on a landing page?
None of these require a particular chain to answer. They require that you pick parts of the stack on purpose. That is the job Sato Hub is built for: map the skills, MCPs, wallets, and payment rails an agent needs, and show the evidence for each, such as whether a repo is open and active and whether an install reproduces. The Sato Score is a transparency and liveness signal. It is not a safety or quality grade, and it does not say an agent is good. It says how much you can check.
What to watch
- ▸Specifics from VeWorld.ai. Which registry contracts, which reputation inputs, and what a creator's record looks like over time. Docs and code would settle most of this.
- ▸Who writes the reputation. Self-attestation, counterparties, or an independent reviewer. These give very different signals.
- ▸The app-store comparison. Lu's analogy is about distribution and monetization for individual builders. Watch whether any agent marketplace, VeWorld.ai or another, publishes how listings are reviewed before they sell anything.
- ▸Always-on agents and control. In a separate October Cointelegraph interview, Animoca Brands chairman Yat Siu said traders could use AI agents around the clock, but warned that users risk building strategies around services they do not control. Control and accountability are two sides of the same question.
The category needs more people asking Lu's question. It needs fewer people declaring it solved. Building your own agent? Start from the stack at satohub.ai/build: describe the agent, and Sato maps what builds it.