Major financial platforms are preparing infrastructure for an emerging class of machine-to-machine transactions driven by AI agents. Companies including Visa and Coinbase are developing payment systems designed to support automated micro-payments generated by AI software.
Industry leaders argue that AI agents will soon execute far more transactions than humans as they purchase computing power, access APIs, and coordinate digital services. Brian Armstrong has predicted that AI agents could outnumber humans conducting internet transactions, while Changpeng Zhao suggested machine-driven payments could eventually exceed human activity by orders of magnitude.
Crypto advocates point to blockchain wallets and stablecoins as ideal tools for agent-driven payments because they can operate without traditional identity verification required by banking systems. Protocols such as x402 embed payments directly into internet requests, enabling software agents to pay for data, compute resources, or digital services automatically.
At the same time, traditional financial networks are building their own solutions. Visa has introduced its Trusted Agent Protocol to support AI-initiated payments on card rails, while other financial institutions are testing similar frameworks within regulated banking systems. The parallel development highlights a potential split between regulated consumer commerce and high-frequency machine-to-machine transactions conducted by AI agents.
