There is one thing I’ve noticed repeating fairly consistently in crypto: with each market cycle, a new main character is found to place expectations on. In the past it was DeFi, then NFTs, then modular, then restaking... and now it’s AI Agents. People talk about automated agents replacing humans in decision-making; people talk about a future where crypto wallets can trade on their own, manage portfolios on their own, and interact with applications without us touching a thing. But I’ve seen too many stories told with beautiful narratives before any evidence appears to show that users truly need it.
That’s also the most persistent problem in crypto—and probably the most boring thing to have to repeat. We’re very good at building infrastructure for needs that don’t exist yet. We build highways before there are cars on the road; we debate throughput, frameworks, and coordination layers while the number of actual users using the product every day is sometimes still very modest. With AI Agents, the question that always nags me isn’t whether the agent is intelligent or not. The question is whether users truly want to delegate action authority to an automated entity on the blockchain, because there’s a pretty big gap between watching a demo and allowing a system to manage assets on your behalf.
From that perspective, the comparison between the Newton Protocol and the Virtuals Protocol is quite interesting. Both are betting on AI Agents, but they seem to be looking at the problem from two different directions. Virtuals Protocol appears to be choosing to build an economy centered around agents. The agent becomes an entity with an identity, a community, the ability to generate revenue, and even be traded like a kind of asset. It has a very crypto color, very narrative-driven—and you have to admit that Virtuals knows how to get the market’s attention. They talk about agents like small startups existing on the blockchain; they craft a story compelling enough for the community to join, at least in the early stages.
Meanwhile, Newton Protocol makes me think of a less flashy approach. They don’t seem to focus too much on turning agents into a cultural product or a community. What Newton cares about more appears to be the action orchestration layer, the verification layer, and the ability for AI to carry out tasks on behalf of users in an environment that can be verified. It may not be as exciting as owning an AI influencer with its own token, but at least from my viewpoint, this is where the story becomes more realistic.
Because if AI Agents truly become part of the crypto ecosystem in the next few years, what matters probably won’t be how good the agent is at chatting. More important is: who is responsible when the agent does something wrong, who can verify its decisions, and who ensures it’s acting according to the user’s original intent. That’s something I see the market talk about less, because it doesn’t create a sense of excitement. It feels more like a problem of trust than a problem of technology.
Virtuals Protocol currently has the advantage of visibility within the community. They’ve managed to turn AI Agents into a highly viral narrative. Users find it easy to understand, easy to join, and also easy to speculate on—but that doesn’t necessarily mean they’re leading in terms of real-world application. The market’s attention sometimes only reflects better storytelling, not necessarily solving the hardest problem.
Newton Protocol, on the other hand, seems to be trying to tackle parts that fewer people want to discuss. The backend, orchestration, verification... The things that are often overlooked during the period when the market is excited by a new narrative, but it’s precisely those things that determine whether AI Agents can go beyond the scope of experimentation.
In the end, I think this race is still too early to name a clear leader, because in crypto, leading the narrative and leading in usage are two completely different concepts. A whitepaper can be very convincing, a demo can be very smooth—but if, after another year, most agents are still only interacting with each other rather than serving real users, then the whole story will just become a familiar loop of the market.
Virtuals Protocol is holding attention, while Newton Protocol seems to be pursuing sustainability in system design. One side may be winning the identity recognition battle; the other might be preparing for a longer-term fight. As for who is actually leading the AI Agents race, maybe no one knows yet. I’m still watching—this part needs time to answer.