Cathie Wood Says Investors Should ‘Follow the Agents’ as AI Starts Handling Money
ARK Invest CEO Cathie Wood says investors may increasingly need to track where AI agents direct spending as the technology moves beyond generating answers and begins carrying out transactions.
Wood made the comments during a panel at Robinhood’s Summit in Houston on Wednesday, According to CoinDesk, arguing that investors may need to shift from simply watching technology developers to monitoring the behavior of AI agents themselves.
“We’re probably going to be talking more and more about ‘follow the agents,’” Wood said.
AI agents are software systems designed to perform tasks on behalf of users, rather than simply responding to prompts or producing text. As these systems become capable of selecting software, services and other digital resources, their spending decisions could provide a new signal of where technology demand is developing.
AI Agents Could Become a New Demand Signal
Wood’s remarks came during a broader discussion covering artificial intelligence, private markets and technology investing. Her comments focused on how the growing use of autonomous software could change the way investors identify emerging technology trends.
Developers have traditionally been an important indicator of technological adoption because engineers tend to use and build around tools they find useful. If large numbers of AI agents eventually make their own decisions about which applications, services and networks to access, their activity could provide another source of insight into where demand is moving.
That shift, however, raises a practical question: how will AI agents pay for the services they use?
The issue becomes more significant as agents gain greater authority to act on behalf of their users and potentially conduct transactions without requiring approval for every individual purchase.
Agentic Finance Raises Questions About Financial Control
Joseph Chalom, co-CEO of SharpLink and former head of digital assets at BlackRock, has argued that the financial infrastructure supporting AI agents should not become concentrated among a small group of banks or technology companies.
“A world full of intelligent agents means nothing if a handful of companies decide where your money can go,” Chalom wrote in the final installment of a three-part series on agentic finance published last month.
Chalom's argument extends beyond whether an AI agent can technically spend money. He focused on the degree of authority users grant to agents and the structure of the financial systems through which those transactions are processed.
For example, a user could authorize an AI agent to spend up to $500 when booking a hotel without giving the software unrestricted access to the user's bank account. Under such a model, the user would also retain the ability to revoke the agent's authority and review a record of its actions.
Portability Could Become a Key Issue
Chalom also argued that users should be able to move their AI agents between financial providers rather than becoming dependent on a single company's infrastructure.
Under that model, an agent could retain its identity, financial information and permissions when switching providers. Chalom compared the concept with mobile-number portability, where customers can move a phone number from one carrier to another without losing the number itself.
The discussion highlights a broader issue emerging alongside the development of autonomous AI: as agents take on more financial responsibilities, the systems governing their permissions, payments and provider relationships could become increasingly important.
Writer: Marcus RenfieldCrypto Market Analyst & Onchain WriterMarcus Renfield covers cryptocurrency markets with a focus on onchain data, Bitcoin price action, and emerging market narratives. His writing examines how capital flows, network activity, and broader market structure influence short- and medium-term trends.He aims to provide clear, data-informed analysis for readers seeking a deeper understanding of crypto market dynamics.