Natural, a fintech startup, has raised $30 million to solve a problem the payments industry has only begun to confront: how do AI agents buy things? The company is building payment rails designed for autonomous software, systems that can hold funds, make purchases, enforce budgets and prove who authorized a transaction, positioning itself squarely against Stripe and the incumbent card networks as machine-to-machine commerce accelerates.
The premise is that the next explosion in transaction volume will not come from humans tapping buttons but from agents booking travel, procuring supplies and subscribing to services on a person's behalf. Existing infrastructure assumes a cardholder present and consenting; agents break every one of those assumptions. Who is liable when an agent overorders? How does a merchant verify a purchasing agent has authority? How does a bank distinguish fraud from a legitimate swarm of autonomous buyers? Natural's answer is a purpose-built layer of identity, permissions and settlement for software with wallets.
The raise lands in the middle of an industry-wide scramble. Card networks have announced agent programs, startups are experimenting with stablecoin rails, and major commerce platforms are debating how to let AI assistants check out on a user's behalf. Whoever defines the standards for agent payments could occupy the same strategic ground Stripe claimed in the last era.
Natural plans to expand its engineering and go-to-market teams and pursue partnerships with merchants and financial institutions. The bet is simple and risky: that agentic commerce arrives fast enough that incumbents cannot retrofit, and slow enough that a startup can survive the wait.
