Alex Svanevik said the firm has processed more than $500 million in trading volume as it shifts from onchain analytics to execution and sees strong demand for non-crypto perpetual markets.
Nansen is repositioning itself around the idea that AI trading agents could outnumber human traders within roughly two years, according to CEO Alex Svanevik. The onchain analytics firm has expanded beyond surfacing blockchain data into direct trading, and Svanevik said its rebuilt agent-based trading platform has now surpassed $500 million in volume. He described the company’s goal as enabling users to trade everything onchain with agents, a strategy that combines three changes at once: moving from analytics to execution, broadening from crypto to all asset classes, and shifting trade selection from humans to AI systems. Svanevik argued agents may outperform retail traders because humans tend to crowd into the same positions, while agents can draw on a wider range of inputs, models and tools. He also said that creates new vulnerabilities, including the risk that inference-based systems could be contaminated or poisoned. Nansen is therefore still backtesting and paper trading its autonomous agents rather than releasing immature products. Svanevik said one agent made $23 in profit while incurring $700 in inference costs, underscoring how early the economics remain. He said Nansen’s advantage comes from more than 500 million labeled blockchain addresses collected over six years, which he said helps its systems detect otherwise hard-to-see onchain activity. He added that demand is already moving beyond crypto, saying about 10 of the top 15 perpetuals by volume on Nansen track non-crypto markets including SpaceX, the S&P 500, gold and crude oil. While Svanevik expects a broader shift to agent-led trading inside two years, he said the real performance edge is still months away.