Natural raises $30M to reinvent payments for AI agents — and take on Stripe

The autonomy ceiling just got a price tag. AI agents can source vendors, compare freight quotes, and message suppliers to arrange delivery. They cannot pay for the shipment. A human still has to authorize the transaction. That single friction point — the moment an agent needs a person to release funds — is where Natural plants its flag.

Kahlil Lalji watched this ceiling harden. His previous startup, Ivella, a Y Combinator-backed banking product for couples, sold to Earnin in 2023. He stayed on as an engineer for two years. The zero-interest-rate era that fueled fintech experimentation had evaporated. Lalji wanted out of finance. He told TechCrunch he felt burned by the sector. Then the agents arrived.

"I kept on coming back to it," he said. "It just feels obvious that agentic payments are going to be structurally the most important problem in the space." That phrase — structurally the most important problem — carries weight. It means the bottleneck sits below the application layer, in the rails themselves. Credit cards and ACH were architected for human intent. They require a person to authenticate, to approve, to bear liability. An agent that operates at machine speed and machine scale cannot wait for a thumbprint.

Natural, founded in 2025 with Ivella co-founder Eric Wang and former Nextdoor engineering manager Walt Leung, does not want to wrap Stripe. It wants to replace the substrate. The company calls itself an "agent orchestration layer" that lets agents move and store funds. Integrate Natural, and your agent can pay a vendor, collect from a customer, transact with another agent — no human in the loop. That is the claim. The proof arrives when the beta lifts.

Kirsten Green, founder and managing partner at Forerunner, led the $30 million Series A. Her firm focuses on consumer experience and the future of commerce. She did not back a checkout widget. She backed a rewrite of dispute handling, identity, and reversibility for a world where the counterparty may be code. That ambition separates Natural from the crowd of agent-tool startups raising on hype. The $40 million total capitalization before general availability is unusual. It suggests the investors see a platform play, not a feature play.

Stripe sees it too. The payments giant is racing to redesign its own rails for agents. The incumbent advantage is distribution, trust, and a decade of regulatory scar tissue. Natural's advantage is a blank page. No legacy ledger to migrate. No merchant base to protect from breaking changes. No quarterly earnings call that demands compatibility over correctness. Lalji says the team has made "enough critical architectural decisions" to give them a "good shot." That is founder language for: we chose the hard primitives early.

The primitives matter. How does an agent prove it represents the principal? How does a vendor verify the agent's authority without a human clicking "I agree"? What happens when the freight arrives damaged — who initiates the chargeback, on what evidence, settled on what timeline? Today's rails answer these questions with human institutions: card networks, banks, courts. Agent-native rails need algorithmic equivalents. Natural's bet is that those equivalents can be built as infrastructure, not negotiated case by case.

Skepticism is warranted. Pre-launch infrastructure companies accumulate technical debt in stealth mode. The "beta trial" label means real money has not yet moved at scale through Natural's pipes. Regulatory scrutiny will intensify the moment agents move meaningful volume without human oversight. Money transmission licenses, KYC/AML obligations, consumer protection statutes — these were written for people. Regulators have not blessed agent personhood. Natural will need lawyers as much as engineers.

Stripe's response will not be passive. They own the merchant relationships. They own the developer trust. They can ship agent APIs on top of existing rails and iterate toward native architecture while collecting revenue. Natural must convince developers to bet on a new network before the old one adapts. That is a distribution war, not just a technology war.

But the timing is not arbitrary. The model capability curve has bent sharply. Agents that could not reason through multi-step procurement six months ago now do it reliably. The payment bottleneck has shifted from theoretical to operational. Companies deploying agents in production are hitting the wall daily. They need a solution now, not when Stripe's roadmap delivers.

Lalji's return to finance after swearing it off is the strongest signal. Founders do not re-enter sectors they hate unless the opportunity structure has fundamentally changed. He sees a new asset class: agent-to-agent commerce volume that cannot exist on current rails. The $30 million buys time to build the rails before the volume arrives.

The editorial take: Natural is not a Stripe competitor in the way Plaid or Adyen were. It is a thesis that the payment stack must be rewritten for non-human actors. If the thesis holds, the company that defines the primitives — identity, authorization, dispute, settlement — becomes the Stripe of the agent economy. If it fails, the primitives get absorbed into Stripe's next major version. Either way, the $30 million ensures the experiment runs at sufficient scale to answer the question. The agents are waiting.