Natural, a fintech startup building payment infrastructure for autonomous AI agents, closed a $30 million Series A round led by Kirsten Green of Forerunner Ventures, pushing its total funding past $40 million. The raise signals a bet that software agents will soon need their own bank-grade wallets, ledgers, and settlement channels, separate from the human-facing payment systems that dominate today. But the company’s reliance on traditional bank partners to custody those funds raises a pointed question: can deposit-insurance rules designed for individual account holders scale to cover potentially millions of AI-controlled balances?
Why AI agent payment rails demand new financial plumbing
The core product Natural describes is a stack of services, including wallets, vaults, payments, ledgering, multi-bank settlement, fraud controls, and agent identity, all designed so that an AI agent can hold funds and execute transfers without a human approving each step. That ambition collides with how money actually moves in the United States. Natural is not a bank. Its disclosures page states plainly that banking services and FDIC pass-through insurance are provided through bank partners and sweep program network banks. In practice, that means every dollar an AI agent holds sits inside a regulated institution’s ledger, not on Natural’s own balance sheet.
The primary U.S. bank partner is Column National Association, confirmed as a federally chartered national bank by the FFIEC registry. Natural’s services agreement also names Bridge as a U.S. financial partner. This two-partner arrangement works at small scale, but it introduces a structural ceiling that grows more visible as agent adoption accelerates.
Standard FDIC deposit insurance covers up to $250,000 per depositor, per insured bank, per ownership category, according to the agency’s own single-account guidance. When a fintech uses sweep programs to spread customer funds across multiple network banks, each bank’s $250,000 limit applies separately, which extends coverage. Yet the question of how an AI agent qualifies as a “depositor” under existing FDIC rules has no published regulatory answer. If thousands of agents controlled by a single company each hold funds in the same sweep network, regulators could treat those balances as belonging to one beneficial owner, collapsing coverage back to a single $250,000 cap. No examiner statement or rulemaking currently addresses this scenario.
Column, Bridge, and the $40 million funding trail
Natural’s own Series A note puts the $30 million round in the context of a broader product vision: wallets and vaults for agent-held funds, real-time ledgering across multiple banks, compliance tooling, and an identity layer that ties each transaction to a specific agent. Kirsten Green of Forerunner led the round, and the company says total funding now exceeds $40 million, which includes earlier pre-seed and seed capital.
The company’s services agreement names Column and Bridge as the two U.S. financial partners responsible for money movement. Column handles wallet accounts and sweep program deposit placement. Bridge facilitates additional payment functions. Neither Column nor Bridge has released public statements detailing their operational role, risk controls, or capacity limits for agent-driven transaction volumes. All available detail comes from Natural’s own legal terms and disclosures, not from the banks themselves.
That gap matters because the entire value proposition depends on bank-grade reliability. If an AI agent initiates a payment that fails at the settlement layer, the downstream consequences, such as a missed vendor payment or a stalled procurement workflow, fall on the business that deployed the agent. Without published data on transaction volumes, settlement failure rates, or stress-test results from Column or Bridge, outside observers have no way to gauge how close the system is to its operational limits. For now, customers must treat the agent infrastructure as a black box whose resilience depends on counterparties they cannot directly evaluate.
Unresolved regulatory questions around agent deposit insurance
The biggest open issue is whether existing deposit-insurance rules can accommodate the model Natural is building. FDIC pass-through insurance works when a fintech’s end users are identifiable natural persons or legal entities with distinct ownership claims. An AI agent is neither. It operates on behalf of a company or individual, but the agent itself holds the wallet in Natural’s system design. If regulators decide that the beneficial owner is the company deploying the agent rather than the agent itself, sweep-program coverage could prove far less generous than the per-agent model implies.
No official regulatory filing, examiner letter, or rulemaking currently addresses how multi-bank settlement for AI agents fits within FDIC single-account limits. That silence is not neutral. It means Natural and its bank partners are operating in an interpretive gray zone where a single enforcement action or supervisory letter could force a rapid restructuring of how funds are held. A determination that all agent balances for a given business roll up to one ownership category at each bank would compress coverage and potentially require rerouting deposits or capping balances per client.
A second unresolved question involves transaction velocity. Autonomous agents could generate thousands of micro-transactions per day, each requiring real-time ledgering and compliance screening. Traditional bank core systems were not built for that cadence. Column is a technology-forward bank, but its public profile does not disclose throughput benchmarks for agent-scale workloads. If transaction volumes spike beyond what the bank’s systems can clear in real time, Natural would face a difficult choice: throttle agent behavior to match banking rails, or add new partners and complexity to keep pace with demand.
Compliance adds another layer of strain. Every agent-initiated transfer must still comply with existing regimes for anti–money-laundering, sanctions screening, and suspicious-activity reporting. The more autonomy agents have to originate payments, the more pressure falls on Natural and its banks to prove that automated controls are as effective as traditional, human-centric processes. Regulators have not yet published detailed expectations for AI-native payment flows, leaving institutions to interpret how far they can lean on algorithms for risk decisions.
What Natural’s raise signals for AI-native finance
Natural’s funding round is, in part, a wager that these regulatory and operational gaps will be resolved in favor of AI-native infrastructure rather than against it. By embedding directly with banks instead of trying to operate in a gray area outside the regulated perimeter, the company is positioning itself as a compliant conduit rather than a challenger to banking authority. That strategy may make supervisors more willing to experiment, but it does not eliminate the need for formal guidance on how agent accounts should be treated.
For businesses considering agent-based automation, the current environment demands careful risk assessment. They must understand that an “agent wallet” is ultimately a bank account held through intermediaries, subject to the same insurance caps and counterparty risks as any other deposit. They should ask how funds are distributed across partner banks, what happens if one bank in the sweep network fails, and how quickly they can reconfigure agent behavior if regulators change the rules.
Natural’s success or failure will likely influence how other fintechs design infrastructure for autonomous software. If the company can demonstrate that agent-driven payments can run at scale on existing banking rails without major compliance incidents or outages, it may set a template for the industry. If, instead, the model runs into binding limits on deposit insurance, throughput, or supervisory patience, it could push innovators toward alternative architectures, including on-chain settlement or non-deposit custodial structures that sit further from the FDIC safety net.
For now, the Series A round gives Natural the capital to expand its engineering team, deepen integrations with Column and Bridge, and court early adopters willing to pilot agent-based financial operations. The open question is whether the legal and regulatory scaffolding around those operations will evolve quickly enough to match the pace of AI deployment. Until regulators speak directly to how autonomous agents fit into deposit-insurance categories and transaction-monitoring frameworks, every dollar flowing through these new rails will carry a measure of uncertainty that no amount of venture funding can fully underwrite.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
