Crypto Long & Short: What will the AI agents run on?

You're reading Crypto Long & Short, our weekly newsletter featuring insights, news and analysis for the professional investor. Sign up here to get it in your inbox every Wednesday.

Note: The views expressed in this column are those of the author and do not necessarily reflect those of CoinDesk, Inc., CoinDesk Indices or its owners and affiliates.

This is your institutional newsletter, Crypto Long & Short. This week:

By Utkarsh Ahuja, founder and managing partner, Moon Pursuit Capital

The conversation around artificial intelligence has become dominated by a familiar question: Which companies will win? Investors are trying to identify the next dominant model, application or AI platform, and enormous amounts of capital are being deployed accordingly. But I think there is another investment question that deserves considerably more attention. If AI becomes as economically important as markets expect, what financial infrastructure will that economy actually run on?

Increasingly, I believe the answer will involve blockchain and digital assets. That does not mean every AI company needs a token, nor does it mean attaching “AI” to a crypto project suddenly creates value. The more interesting opportunity sits underneath both sectors, where autonomous software begins interacting with financial systems and needs infrastructure capable of moving value as quickly and programmatically as it moves information.

We are moving toward a world in which AI agents will not simply generate text or analyze data. They will negotiate with other agents, purchase computing resources, pay for data, execute transactions and manage financial decisions within parameters established by humans. Traditional payment infrastructure was built around people and institutions initiating transactions. It was not designed for potentially millions of autonomous software agents conducting low-value transactions continuously across borders.

Blockchain infrastructure is unusually well suited to that environment because money can become programmable. An AI agent can interact with a wallet, execute a smart contract or transfer a stablecoin without requiring the same layers of manual intervention associated with traditional financial infrastructure. Stablecoins are particularly important here because they provide a bridge between blockchain's programmability and a familiar unit of account.

This is where I think investors should broaden their view of the AI trade. The opportunity is not limited to owning companies building models or applications. If AI creates significantly more machine-to-machine economic activity, there will also be demand for the infrastructure that allows those machines to transact, prove identity, establish ownership and exchange value.

Identity is a good example. As autonomous agents become more capable, markets will need ways to determine who or what is behind a transaction and what an agent is authorized to do. The same applies to provenance. When AI systems consume data, create intellectual property or execute transactions, establishing where information came from and who owns what becomes increasingly important. Blockchains will not solve every aspect of these problems, but their ability to provide verifiable, shared records makes them a natural part of the infrastructure stack.

There is also a capital-markets angle that I think gets overlooked. Tokenization is already bringing traditional assets onto blockchain rails, while stablecoins have demonstrated that meaningful financial activity can operate onchain. At the same time, AI is making financial decision-making increasingly automated. Those trends are currently developing separately, but their convergence could be significant. An AI agent that can analyze markets but cannot efficiently hold, exchange or settle assets is inherently limited. Put programmable intelligence together with programmable assets, and the possibilities become much more interesting.

For investors, this requires some discipline because technological convergence does not mean every project exposed to AI and crypto deserves a premium valuation. We saw during previous crypto cycles how quickly narratives could outperform the fundamentals, and the same risk exists around AI at present. I would rather look for infrastructure solving a genuine problem created by greater automation: payments, settlement, identity, cybersecurity, custody and the rails connecting traditional and digital markets.