The payments layer of the internet was built for humans — humans who click "buy," enter card details, and confirm purchases. Natural, a startup founded just one year ago, believes that layer needs to be rebuilt from scratch for a world where AI agents are increasingly the ones making purchasing decisions.

The company has raised $30 million to pursue that thesis, positioning itself as foundational infrastructure for what the industry is calling the agentic economy.

Why Existing Payment Rails Fall Short

Today's payment systems — including those offered by Stripe, the dominant player in developer-facing payments — were architected around user sessions, browser flows, and human authentication. When an AI agent needs to autonomously purchase a service, book a resource, or execute a transaction on a user's behalf, it's essentially trying to fit a square peg into a round hole.

The core problems are structural:

  • Authentication flows assume a human is present to complete verification steps
  • Fraud detection models are tuned to human behavioral patterns, not agent-initiated traffic
  • Spend controls and permissions lack the granularity needed for delegated, policy-driven transactions
  • Reconciliation and audit trails aren't designed for multi-agent pipelines

Natural is building financial architecture that treats agents as first-class principals — with programmable spend limits, fine-grained permissioning, and transaction flows that don't assume a human is in the loop at every step.

Taking Aim at Stripe

The framing of this as a direct challenge to Stripe is provocative but not unreasonable. Stripe's dominance comes from owning the developer relationship in payments, and Natural appears to be pursuing the same vector — just one layer ahead in the technology cycle.

Stripe has been investing in AI tooling, including agent-friendly APIs and updated authentication mechanisms. But incumbents optimizing existing infrastructure for new paradigms often leave room for purpose-built challengers. Natural is betting that the architectural compromises required to bolt agent support onto human-facing systems will be a meaningful competitive disadvantage.

What This Means for Founders and Builders

For startup founders building AI-native products — particularly those involving autonomous workflows, AI assistants with spending capabilities, or multi-agent systems — the payments layer is increasingly a product decision, not just a vendor selection.

Key implications:

  • Agentic products need payment infrastructure that can enforce policy programmatically — not just rate-limit API calls, but actually govern what an agent is allowed to spend, on what, and under what conditions
  • Liability and compliance questions around agent-initiated transactions are still being worked out; your payment provider's architecture will shape your exposure
  • Developer experience matters — the startup that makes it trivially easy to embed compliant, auditable agent payments into a product will capture the next generation of builders the way Stripe captured the last

Natural's pitch is essentially that they're building the Stripe for machines. Whether that's a standalone category or a feature Stripe eventually absorbs is the central strategic question hanging over the space.

Early Days, High Stakes

At one year old and $30M raised, Natural is early. The agentic economy it's betting on is still largely theoretical at commercial scale — most enterprise AI deployments still have humans approving significant transactions. But the infrastructure bets that matter most are placed before the demand is obvious.

The broader market is moving fast. Competitors in the agent infrastructure space are staking out adjacent territory across memory, orchestration, and identity. Payments is a logical and high-value layer to own — whoever controls how AI agents spend money controls a significant lever in the agentic stack.