Sapiom Raises $15.75 Million for Machine-Native Financial Infrastructure

SAN FRANCISCO, Calif. - Sapiom has raised $15.75 million in seed financing to build a financial layer for artificial intelligence agents, addressing one of the most significant technical hurdles in the push for autonomous software. The funding round was led by Accel, with participation from a list of high-profile institutional backers including Gradient Ventures, Okta Ventures, Menlo Ventures, and Array Ventures. The San Francisco-based company was founded by Ilan Zerbib, a former engineering leader at Shopify, who is leveraging his background in complex commerce systems to build what the company describes as machine-native financial infrastructure.

The core problem Sapiom aims to solve is the inherent friction in how software interacts with the digital economy. Currently, when an artificial intelligence agent requires a specific resource—such as a proprietary dataset for training or an API call to perform a customer service task—the developer responsible for that agent must manually arrange vendor relationships, enter payment credentials, and manage identity permissions. This manual overhead limits the scale and speed at which autonomous systems can operate, keeping them largely confined to pre-integrated, siloed environments.

By combining access credentials, spending policies, and settlement for machine-to-machine purchases into a single platform, Sapiom centralizes the operational requirements for these transactions. Under this model, an agent could autonomously acquire an API call, a dataset, or computing capacity without a human developer intervening to separately authorize every individual vendor relationship. The system acts as a specialized intermediary that translates the requirements of an autonomous agent into the structured legal and financial formats required by service providers.

The emergence of Sapiom arrives at a time when the tech industry is shifting its focus from simple generative chat interfaces to 'agentic' workflows. These agents are designed not just to suggest text or code, but to execute actions across the web. However, for an agent to be truly autonomous, it requires its own wallet and authorization keys. Sapiom's platform is built on the premise that the next generation of software will act as a direct economic participant, requiring its own layer of the financial stack that is distinctly different from user-facing products designed for humans.

Industry analysts have noted that the current infrastructure for identity and billing was built for a world of human-driven clicks and credit card entries. When these systems are applied to machines, they often lack the granularity and speed required for high-frequency micro-transactions. Sapiom’s purpose-built layer is designed to make autonomous software easier to deploy across many suppliers, providing a standardized protocol for how an agent identifies itself and pays for a service in real-time.

A critical component of the Sapiom proposition is the retention of human oversight. While the agents operate autonomously, the business using the agent remains responsible for its actions. Sapiom allows these companies to set strict limits on spending and resource consumption while maintaining a comprehensive, auditable record of every transaction. This governance layer is intended to mitigate the risks of runaway costs or unauthorized data access, which have become primary concerns for enterprises experimenting with agentic AI.

The seed round lands as venture capital interest in AI infrastructure continues to intensify, even as investors debate the timeline for mass adoption. Sapiom is effectively betting that agents will become a new class of economic participant while companies remain responsible for governing what those agents can do. This bet mirrors previous shifts in internet architecture, such as the transition to cloud services or the rise of the API economy, both of which required new layers of middleware to handle authentication and monetization.

However, significant execution risks remain for the startup. The market for autonomous agents may develop more slowly than the significant seed funding suggests. Currently, most agent activity is still concentrated in controlled, highly specific workflows rather than the open-ended digital economy Sapiom envisions. If the adoption of autonomous agents stalls due to security concerns or lack of reliability, the demand for a specialized machine-to-machine payment layer could reflect that stagnation.

Competitive pressures are also significant. Major cloud providers, identity management platforms, and established payment processors like Stripe or Adyen can extend their existing platforms to accommodate machine customers. Sapiom must prove that its combined product—which tightly integrates credentials with payments—is inherently safer and simpler than using separate, established tools for identity and financial settlement. The company will need to demonstrate that a unified approach offers performance advantages that larger incumbents cannot easily replicate.

Compliance presents another hurdle. Financial transactions, especially those involving automated entities, are subject to rigorous regulatory frameworks including anti-money laundering (AML) and know-your-customer (KYC) requirements. Sapiom will need to meet these demanding compliance standards while ensuring its platform remains frictionless for developers. As a new entrant in the financial technology space, the company will be under scrutiny to ensure that its machine-native infrastructure does not become a vector for unauthorized financial activity.

With the $15.75 million in new capital, Sapiom plans to expand its list of integrations and work closely with early builders of AI agents. The startup’s success will likely be measured by the volume of recurring, authorized purchases occurring across multiple suppliers through its rails. The goal is to move beyond one-off proof-of-concepts toward a state where agents can reliably navigate the web’s service layer to solve complex business problems without human hand-holding.

As the industry watches for the next stage of AI development, the focus is shifting from what models can say to what they can do. If Sapiom can make machine transactions dependable and auditable, it may become a foundational piece of infrastructure beneath a new generation of applications. In this future, the end-user may never see the payment process, as the underlying infrastructure handles the rapid exchange of value between machines in the background.