Arcade AI raises $60M to secure authorization for AI agents

Authorization platform targets enterprise AI security

Arcade AI Inc. has closed a $60 million Series A round to expand its authorization platform designed specifically for artificial intelligence agents, according to details first reported by SiliconANGLE.

SYN Ventures led the investment, with participation from Morgan Stanley and publicly traded technology consulting firm Wipro Ltd. The round follows a $12 million seed raise in 2024, the year founders Alex Salazar and Sam Partee launched the company.

Why it matters

As enterprises deploy AI agents to automate workflows across business applications, authorization has emerged as a critical security challenge distinct from authentication. While authentication verifies an agent's identity, authorization determines which specific features and data that agent can access. Arcade addresses this gap with infrastructure that reduces custom development work while tightening security controls—a combination that becomes more valuable as agent deployments scale across organizations.

Solving the authorization problem

Traditionally, developers built custom authorization mechanisms for each AI agent project, a process that consumed engineering time and often introduced security vulnerabilities. Arcade provides prepackaged authorization features that integrate directly with companies' identity provider (IdP) systems—the databases that track application access permissions.

When permissions change in an organization's IdP, Arcade automatically updates AI agent access rights without manual intervention. The platform also addresses a common security weakness: most agent authorizations remain valid continuously, creating persistent attack vectors. Arcade instead authorizes specific agent actions rather than blanket access.

The system implements authorization using OAuth 2.0, an open-source protocol that manages application access through OAuth tokens. Arcade stores these tokens on behalf of developers, encrypting them before transmission to storage rather than after. The platform also employs salting mechanisms to protect against attacks that exploit credentials with identical plaintext values.

Platform capabilities and roadmap

Beyond authorization, Arcade enables companies to log agent actions for audit and compliance purposes. The platform provides access to more than 8,000 Model Context Protocol (MCP) tools that agents can use to automate business tasks.

"Agents don't fail in production because the model is wrong," Salazar said. "They fail because nobody can prove that for any given action by an agent, whether that agent on behalf of that user can perform that action on that resource. That's what we built."

Arcade plans to use the new capital to expand its agent governance features and grow its MCP tool catalog. Salazar previously held senior product roles at Oka Inc., while Partee came from Redis Inc.

The funding details were first reported by SiliconANGLE.