Nansen CEO bets on AI agents to overtake human traders within two years

I have spent a significant amount of time looking at dashboards. If you have been in the crypto space for more than a minute, you probably have too. We have spent an entire cycle obsessed with 'on-chain analytics'—trying to read the tea leaves of whale wallets and exchange inflows to figure out where the money is moving. But here is the cold truth: by the time you see the data on a pretty chart, the opportunity is usually gone. The machines have already eaten the alpha.

Alex Svanevik, the CEO of Nansen, seems to be reaching the same conclusion. Nansen built its reputation on being the window into the blockchain, labeling wallets and making sense of the chaos. Now, they are signaling a massive pivot. They are moving away from just showing you what happened and moving toward building AI agents that do the work for you. Svanevik is putting a timeline on it, too: he expects AI agents to overtake human traders within the next two years.

The Death of the Dashboard

For a founder, this is a signal to stop building better mirrors and start building better hammers. A dashboard is passive. It requires a human to log in, interpret data, manage their emotions, and click a button. In the time it takes you to double-check a gas price, an algorithm has already executed ten trades. The gap between information and action is where humans lose.

Svanevik’s bet is that the future of crypto isn't 'user-facing' in the traditional sense. It is 'agent-facing.' We are moving toward a reality where the primary users of decentralized finance protocols won't be people with MetaMask extensions, but autonomous scripts with private keys. This isn't just a slight shift in tools; it is a total overhaul of how we design software.

Why Two Years?

Two years sounds aggressive, but in the context of the current AI crossover, it is actually quite conservative. We have already seen the rise of 'AI influencers' on X (formerly Twitter) launching their own tokens and managing their own treasuries. While many of these are currently glorified marketing stunts, the underlying plumbing is being laid down right now.

The reason agents will win so quickly is due to the unique nature of crypto. Unlike the traditional stock market, crypto is open 24/7, highly volatile, and—most importantly—permissionless. You don't need to ask permission from a bank to let a piece of Python code move $100,000. The infrastructure for AI to act as a financial entity only exists in Web3. You cannot give an AI agent a traditional bank account easily, but you can give it a Solana or Ethereum wallet in seconds.

What This Means for Builders

If you are building in this space, you have to look at your product and ask: 'If a human never looked at this, would it still work?' This is a complete departure from the 'user experience' obsession of the last five years. We have been trying to make crypto easy enough for your grandmother to use. Svanevik is suggesting that your grandmother will never use it—she will just hire (or subscribe to) an agent that does it for her.

This shifts the developers' focus to a few key areas:

• Machine-readable APIs: If your protocol documentation is bad and your API is buggy, agents will skip you. You aren't competing for eyeballs anymore; you are competing for integration.
• Verifiable Compute: As agents take over, we need to know that the AI is actually doing what it says it is doing. This is where the intersection of ZK-proofs and AI becomes a practical necessity rather than a whitepaper buzzword.
• Safety Rails: A human might realize a trade looks 'weird' and stop. An agent will follow its logic off a cliff. Building the autonomous 'kill switch' is going to be a massive sub-sector.

The Skeptic’s Corner

I wouldn't be doing my job if I didn't point out the risks here. We have seen what happens when high-frequency trading bots go dormant or glitch out in traditional markets—you get flash crashes. In an environment like crypto, where liquidity is often thin and fragmented, a world dominated by AI agents could lead to extreme volatility that makes our current 'bumpy' days look stable.

There is also the 'dead internet' theory applied to finance. If every trade is just an agent reacting to another agent, does the market actually reflect any real-world value? Or does it just become a giant feedback loop of mathematical noise? Svanevik and Nansen are banking on the idea that these agents will be smarter than the average retail trader, but 'smarter' doesn't always mean 'safer.'

The Founder Perspective

From where I sit, Nansen’s pivot is a survival move. Analytics is becoming a commodity. Every L2 has a block explorer; every niche has a specialized data tool. By moving into the 'agentic' space, Nansen is trying to own the execution layer. They don't just want to tell you that a smart money wallet is buying; they want to be the engine that allows you to mimic that wallet instantly.

For those of us building in AI and crypto, the takeaway is clear: stop building for the 'average user.' That person is tired, overwhelmed, and probably out of money. Start building for the autonomous agent. The 24-month countdown has started. If your tech stack doesn't support automated, programmatic interaction by 2026, you are building a ghost town.

The future of crypto trading isn't a better UI; it's no UI at all.

We are entering the era of the 'headless' financial system. It’s going to be fast, it’s going to be efficient, and for a lot of people, it’s going to be terrifying. But for the builders who get the infra right, it is the biggest opportunity since the inception of the smart contract.

Read the original at The Block →