The Pulse
Opsin Raises $21M to Govern Enterprise AI Agents
Opsin Security announced a $21 million funding round led by SYN Ventures and Race Capital on October 15, 2026. The company plans to use the money to expand its AI-agent security platform, which tracks permissions and behavior across enterpr
AI.info Team · Written with AI ·
Opsin Security announced a $21 million funding round on October 15, led by SYN Ventures and Race Capital, to expand its tools for monitoring and controlling enterprise AI agents. The company says it will use the investment to serve more businesses, particularly in regulated industries, and to add to its agent-security platform. The announcement does not identify a round stage or say whether the amount represents total funding to date.
From data exposure to agent behavior
Opsin began by addressing the risk that generative AI systems could expose company information to people who should not see it. In a separate announcement in April 2025, the company said it had raised a $7 million seed round led by Race Capital, with participation from CapitalX, Lockstep Ventures and angel investors. Its new funding announcement describes a broader focus: tracking the identities, purposes, permissions, connectors and tool calls of agents.
The shift reflects a specific security problem in Opsin’s account. Enterprise agents can run on servers, connect directly to company data and take actions, while conventional controls often focus on endpoints, networks or prompt-and-response exchanges. Opsin argues those controls may not show whether an agent’s actions fit its assigned purpose.
How Opsin’s Risk Graph works
The company says its Risk Graph starts with a common model for agents across Microsoft 365 Copilot, Copilot Studio, Claude Enterprise, ChatGPT Enterprise and Google Gemini. It uses information about agent setup—including connectors, tool calls and sharing scope—alongside runtime activity to build context for assessing risk.
Opsin says the system looks for behavior such as an agent accessing information outside its role or returning sensitive data to users whose roles do not match. When it detects a policy boundary being crossed, the platform can quarantine the agent, block the action or alert the agent’s owner. The announcement describes those functions as product capabilities; it provides no independent performance measures or comparative test results.
Funding targets expansion, not a new round label
Opsin says the $21 million will support faster product development, service for more enterprises in regulated sectors, continued expansion of its risk-protection platform and hiring. The company names healthcare, financial services and manufacturing among the industries where it says customers are using Opsin.
The company also says it connects through APIs and requires no special training, with an initial assessment of agents across approved enterprise AI platforms delivered within 24 hours. Those are Opsin’s stated deployment terms; the funding announcement gives no pricing, customer count or breakdown of how much of the round is reserved for product work versus sales or hiring.
A wider remit than the 2025 seed
The 2025 seed announcement centered on identifying sensitive data that tools such as Copilot might expose and helping companies remediate oversharing. Opsin now presents its product as a way to govern agents themselves, including their permissions and actions across several enterprise platforms. The company says its new capital will fund that expansion and additional support for regulated-industry customers.