Blockchain Is the ‘Perfect Solution’ for AI Agent Trust, VeChain CEO Says

Artificial intelligence agents are about to become some of the busiest actors on the internet, working through the night while their owners sleep, and that shift raises a question the technology industry has not yet answered: how do you trust a piece of software that acts on your behalf when nobody is watching it? VeChain co-founder and chief executive Sunny Lu believes he knows the answer. Speaking on Cointelegraph’s Chain Reaction show on Wednesday, November 5, 2026, Lu argued that autonomous AI agents could drive blockchain’s next great adoption cycle, because they will create urgent demand for systems that can verify who an agent is, what it has done, and whether it can be trusted. In his words, the answer to that problem is simple. “The perfect solution is a blockchain,” he said.

Lu made his case while discussing VeWorld.ai, an agent-building platform that uses VeChain’s blockchain. In the world he described, every agent would carry a digital “passport” proving its identity, and every agent would operate inside a “decentralized credibility system” that tracks the reputations of agents and of the people who created them. Smart contracts could settle transactions between agents, or between agents and people, automatically and transparently. The idea is not to slow agents down with paperwork, but to give everyone they deal with a reliable way to check their history before doing business with them.

Blockchain AI Agent Trust: Why Agents Need an Accountability System

The core of Lu’s argument starts with a simple observation about how AI agents differ from earlier software. A chatbot answers questions when asked. An agent acts. It can place orders, move funds, sign up for services, and negotiate with other agents, all without a human pressing a button each time. That independence is the whole point of the technology, but it also creates a trust gap. When a human employee makes a mistake, you can ask who they are, who hired them, and what their record looks like. When an anonymous agent shows up asking to transact, none of those checks exist today.

Lu’s answer is to give every agent a verifiable identity and a reputation that follows it. A digital passport, recorded on a blockchain, could confirm which agent you are dealing with and who built it. A credibility system built on the same ledger could record how that agent has behaved in the past: whether it kept its promises, paid its bills, and completed its tasks honestly. Because the records would sit on a shared ledger rather than inside one company’s private database, anyone could check them, and no single party could quietly rewrite an agent’s history after a bad outcome.

There is also a safety dimension. If an agent misbehaves, perhaps because it was given bad instructions or because it was manipulated by hidden text on a website, a recorded history makes it possible to trace what happened and hold someone accountable. The passport links the agent to its creator, and the credibility system records the incident for everyone to see. That combination of identity and memory is what Lu means by a new accountability system, and it is the part of his vision that goes beyond simple payments.

Decentralized Credibility System: Passports and Reputations for Machines

The most distinctive element of Lu’s vision is the “decentralized credibility system” he described on the show. The concept has two halves, and both are worth unpacking. The first half is identity: every agent gets a digital passport, a set of credentials that proves what it is and who created it. The second half is reputation: a running record, visible to everyone, of how the agent has behaved. Together, they are meant to answer the two questions anyone asks before trusting a stranger: who are you, and can I believe you?

Digital passports for agents would work in a broadly similar way to digital identity for people, with one important difference. A person’s identity is usually anchored in a government document or a company database. An agent’s identity would be anchored in a blockchain record that no single organisation controls. That matters because agents may operate across borders, platforms, and jurisdictions, and a passport issued by one company might not be accepted by another. A shared ledger offers a neutral ground where every participant can verify the same record.

Smart contracts would handle the money side. When two agents agree on a task and a price, a smart contract could hold the payment, release it when the work is verified, and record the outcome for both reputations. The same mechanism could work between an agent and a person. This is not a new idea in crypto, where escrow-style contracts have existed for years, but applying it to autonomous agents gives it a much larger stage. If agents really do become the busiest customers on the internet, the volume of these small automated settlements could be enormous.

There are real obstacles, of course. Reputation systems can be gamed, identities can be faked, and a passport is only as reliable as the process that issues it. A bad actor could create a new agent with a clean passport after an old one earned a bad reputation, unless the system also tracks creators, which is why Lu emphasised that the credibility system would cover the reputations of creators as well as agents. Designing those rules fairly, and keeping them from being exploited, is a hard problem that no blockchain has fully solved yet.

AI Agents Could Help People Monetize Their Expertise

Lu also argued that AI agents could open up new ways for ordinary people to earn money from what they know, extending software-building opportunities to people who cannot write code. He compared the opportunity to mobile app stores, which allowed individual developers to distribute and monetize their work without needing a large company behind them. Agent-building tools, he suggested, could open a similar market to people whose skills lie outside programming, including those with specialist professional experience, local knowledge, or hobbies worth sharing.

The comparison is instructive. When smartphones arrived, the app store model let a single developer reach millions of users, and it created entirely new categories of small business. Lu’s suggestion is that agent platforms could do something similar for knowledge work: a teacher could build an agent that tutors students, a shopkeeper could build an agent that manages orders, and a local expert could build an agent that answers questions for visitors. The person supplies the expertise, and the platform supplies the technology.

Lu gave personal examples from his own experiments. He described building an educational agent for his daughter, an agent to support his mother’s daily routine and recovery, and a virtual employee to manage purchasing, inventory, sales, and operations for an acquaintance’s shop in Milan. These are small, concrete stories rather than grand promises, and they illustrate the point he was making: agents are not only for large companies or for traders. They can be built around the needs of a family or a single small shop.

The Risks of Letting Agents Work Around the Clock

Not everyone in the industry shares Lu’s optimism without reservations. Other participants have highlighted the risks of relying on agent providers, and the tension between opportunity and control is now one of the central debates in the agent economy. The more capable agents become, the more important it is to ask who controls them, what they can access, and what happens when they go wrong.

In an October interview, Animoca Brands chairman Yat Siu told Cointelegraph that crypto traders could put AI agents to work around the clock, monitoring markets and executing strategies while their owners slept. But he warned that users risked building their trading strategies around services they did not control. The warning cuts to the heart of the issue: an agent that runs on someone else’s infrastructure, using someone else’s models, can be changed, restricted, or shut down without notice, and a trading strategy built on it can collapse overnight.

The two views are not contradictory. Siu’s warning and Lu’s vision can both be true at once: agents can be powerful tools, and they can also create new dependencies. The practical lesson for users is to think carefully about what they hand over to an agent, to keep control of the things that matter most, such as private keys and payment limits, and to treat agent platforms the way they would treat any other service provider: useful, but not infallible. As agents move from experiments to everyday tools, that caution will matter as much as any new feature.

FAQ

What did VeChain CEO Sunny Lu say about blockchain and AI agents?

Speaking on Cointelegraph’s Chain Reaction show on November 5, 2026, Lu said AI agents could drive blockchain’s next adoption cycle by creating demand for systems to verify agent identities and reputations, adding: “The perfect solution is a blockchain.”

What is a decentralized credibility system?

It is Lu’s proposal for a blockchain-based system where AI agents carry digital passports proving their identity and operate within a shared record of their reputations, along with the reputations of their creators, so anyone can check an agent’s history before trusting it.

What is VeWorld.ai?

VeWorld.ai is an agent-building platform discussed by Lu on the show that uses VeChain’s blockchain. It is designed to let people create AI agents and give them an on-chain identity, with smart contracts able to settle transactions between agents or between agents and people.

How could AI agents help non-developers earn money?

Lu compared the opportunity to mobile app stores, arguing that agent-building tools could let people with specialist knowledge, local expertise, or hobbies turn that expertise into services or businesses without needing to write code.

What risks did other industry figures highlight?

In an October interview, Animoca Brands chairman Yat Siu said crypto traders could use AI agents around the clock but warned they risked building trading strategies around services they did not control, highlighting the danger of dependence on outside providers.

Source: Cointelegraph

Related reading on MM Crypto Analysis Web

• AI Agents Raised Their Own Spending Limits in a Test — and Sent Money to an Impersonator
• Coinbase’s x402 Passes $42 Million as AI Agents Do 97% of Agent Payments
• Agent Payments Hit 196 Million Transactions but the Data Is Noisy, Fireblocks Warns

More in AI Crypto on MM Crypto Analysis Web.