Franklin Templeton and Circle make the case for blockchain as AI agent infrastructure

Franklin Templeton‘s head of digital assets and innovation, Sandy Kaul, argues that autonomous AI agents represent a structural demand driver for blockchain networks rather than a speculative angle, positioning public chains as the natural payment rail for machine-to-machine commerce. Her thesis holds that agentic AI, systems built to complete tasks with minimal human input, will generate vast volumes of micropayments that traditional card and bank networks cannot handle economically, because fees routinely exceed the value of the transaction itself.

Kaul contends that public blockchains offer three properties that legacy rails lack at the relevant scale: programmable transactions, cryptographic identity, and near-instant settlement. As AI agents pay for API calls, computing time, or data access, they would hold native digital assets and settle peer-to-peer, bypassing intermediaries. Rising transaction volumes would in turn increase demand for native tokens used to pay network fees, while also expanding revenue available for developer incentives, security, and decentralised applications.

Circle chief executive Jeremy Allaire frames the same dynamic as a single convergent shift rather than two parallel trends. In a recent paper he argued that AI is compressing the cost of knowledge work toward zero while programmable digital money is doing the same for payments and coordination; the implication is that AI agents become economic actors in their own right, hiring other agents and acquiring services autonomously over blockchain infrastructure. Robinhood’s May launch of AI-powered investing tools, which allow agents to trade stocks and make purchases on behalf of users, illustrates that the agentic layer is already entering production, adding weight to the infrastructure argument both Kaul and Allaire are advancing.