Obsidian Security has closed an $85 million Series D at a $1.1 billion valuation, joining the unicorn club as enterprises scramble to keep AI agents under control. Crescent Cove Advisors led the round, with Greylock Partners and Menlo Ventures also participating, and total funding now exceeds $200 million.
Demand signals are strong: more than 100 customers pay over $100,000 a year, at least 14 spend above $1 million, and 60 of the Fortune 500 use the platform across finance, social media and telecom. The new capital is earmarked for R&D and deeper penetration of the Fortune 500 and Global 2000.
The business case rests on a striking imbalance: Obsidian counts 144 non-human identities for every human one inside third-party applications, and frontier-model agents keep widening that gap. Its runtime layer blocks privilege escalation, excessive data access and policy violations while agents work, and it logs every MCP server an organization connects, mapped to the agents that call on it.
The company is also folding governance into the developer tools enterprises already use. New controls for Anthropic’s Claude Code and Cowork let security teams cap agent permissions around production data and refuse unsanctioned tool calls at runtime. Chief executive Hasan Imam argues the risk lives where the work does, since agents gravitate to third-party applications where the data sits. The round lands a day after agent security rival Zenity raised $125 million, two big bets on the same anxiety in as many days.
