Sapiom is Building Infra to Give AI Agents Trusted Access to the API Economy

Sapiom, giving AI agents trusted access to the API economy, led by Ilan Zerbib, has raised a $15.75M Seed round, led by Accel, with participation from Gradient, Array Ventures, Okta Ventures, Menlo Ventures, Anthropic, Coinbase Ventures, Formus Capital, and Operator Collective.

For two decades, the internet economy has been built for humans; interfaces, authentication, payments, and risk models all assume a person is present to sign up, manage access, and approve spending.

That assumption is breaking.

Economic activity is shifting from human-to-business to machine-to-machine. AI agents will move beyond chat interfaces to take action, provisioning compute, purchasing data, calling paid APIs, and orchestrating complex workflows autonomously.

Over the last decade, billions have gone into scaling AI intelligence, and almost nothing into scaling access.

But the agentic era can’t advance until AI agents can autonomously access and operate real-world systems and services.

That creates a paradox: agents are brilliant, but powerless.

Today, AI can write code but can’t buy the infrastructure to run it. It can plan a marketing campaign, but can’t pay for the SMS API to launch it. It can design a landing page, but can’t fund image generation to fill it. And it can build an app with user accounts, but can’t provision the authentication needed to secure it.

Sapiom gives AI agents trusted access to the API economy

Money is the universal API key. If an agent can safely spend, it can access compute, data, inference, messaging, and thousands of specialized services, without pre-built integrations or vendor-by-vendor onboarding.

Sapiom makes that possible. We turn spending into a developer primitive: policy-driven, programmable, and safe, just like any other call in the stack. Identity (KYA), wallets, policy enforcement, risk controls, metering, billing, and settlement are all abstracted behind a single integration.

Every transaction generates verifiable context, who the agent is, what it’s allowed to do, who it paid, and what happened, creating a feedback loop that strengthens controls as systems scale. That’s what makes autonomous agents enterprise-ready.

Every shift in commerce requires new primitives.

Experience building at the edge of the financial stack, from founding Earny (acquired in 2021) to leading payments engineering at Shopify, scaling Shop Pay past $100B in GMV, reveals a consistent truth: infrastructure built for one era cannot serve the next.

The card-not-present stack was optimized for humans. As AI agents emerge, it’s clear that human-centric rails cannot support machine-centric commerce.

History is consistent:

• In-person commerce required card networks.

• Internet commerce required developer payment APIs.

• Machine commerce requires a way for software to spend across the API economy.

AI agents are becoming economic actors, provisioning infrastructure, negotiating with vendors, allocating capital, and operating continuously at a massive scale. That future cannot run on rails designed for humans clicking “Buy.”

It requires a new layer: trusted, programmable access to the API economy that makes machines spend safe, scalable, and native to software. That is the primitive being built.

Building the machine-native future

Machine commerce is already underway. Software will soon negotiate, procure, and settle value at a scale far beyond human commerce.

The equation is simple:

Reasoning × Access = Capability

Models are already powerful. What’s missing is safe, programmable access to the real economy.

Sapiom provides that access, turning paid services into a native software primitive and unlocking the full potential of existing AI systems.

You scale the reasoning. Sapiom is the access.