Wintermute: crypto's next growth wave will come from AI and the 'machine economy'

In its research on the future of the crypto market, market maker Wintermute describes an emerging trend it frames as the industry's next growth wave after the speculative cycles of past years: crypto becoming the settlement layer for a new "machine economy," where deals are made not by people, but by autonomous programs.

What the "Machine Economy" Actually Means

The idea centers on AI agents — self-directed programs capable of perceiving their environment, making decisions, and acting autonomously toward a goal, without constant human prompting. Unlike a chatbot that simply answers a request, an agent can continuously browse the internet for information, execute smart contracts, transfer funds, negotiate with other agents, and run entire workflows on its own. Wintermute frames the broader idea this way:

“Crypto is becoming the clearing and settlement layer that the internet economy has always needed — one that operates continuously, transparently, and without requiring permission from centralised gatekeepers.”

— Wintermute Ventures, "Digital Assets in 2026" research note

Quote source: Wintermute

Agentic Payments as the First Real Application

Analysts see agentic payments as the idea's most mature application right now: micropayments AI agents make to access a service, pay for an API call, buy data, or rent compute within a single workflow made up of dozens of such transactions. The key reason crypto rails specifically are needed, rather than traditional payment rails: stablecoins let an agent settle in seconds without human involvement, while bank transfers simply aren't built for that volume of micropayments. The x402 protocol (the Machine Payments Protocol), already running on Solana and Base for more than a year, lets agents programmatically discover, access, and pay for services in stablecoins without any human in the loop.

Why This Is the Next Wave, Not the Current Trend

There's an important nuance here: over the past year, AI investment, by Wintermute's own estimates, has actually been pulling capital away from crypto that might otherwise have flowed into the market, capping its upside. In other words, until now, AI and crypto have largely been competing for the same institutional capital. The "machine economy" thesis is fundamentally different: it's not two sectors competing for the same investor, but a technical convergence where AI systems themselves become users of crypto infrastructure rather than an alternative to it. That's exactly why Wintermute frames this as a wave still to fully unfold, rather than something already realized.

What This Means in Practice

If the thesis holds, value will accrue less to individual "AI tokens" and more to the infrastructure agents actually use to settle with each other — stablecoins, payment protocols, and networks with low fees and fast confirmation. For anyone tracking the market, that means watching not just trading volume for specific tokens, but the growth in the number and volume of actual machine-to-machine transactions moving through protocols like x402.

This material is for informational purposes only and is not investment advice.