Obsidian Raises $85M for AI Agent Security
Obsidian Security raised an $85M Series D as enterprise security shifts toward non-human identities and AI agents.
The next cybersecurity problem is not only human users. It is software identities, automations and AI agents acting across company systems.
What happened
Obsidian Security raised an $85 million Series D led by Crescent Cove Advisors, with participation from existing investors including Greylock and Menlo.
The company secures non-human identities and AI agents across third-party applications. That includes understanding what automated systems can access, what permissions they hold and whether their behaviour creates risk.
The round lands in the same broader category as several recent agent-security financings, showing that this is becoming a real enterprise security budget line.
Why it matters
AI agents are useful only if they can take actions: read data, trigger workflows, update systems and interact with tools. But every permission given to an agent also creates a new security surface.
Traditional identity tools were built mainly around employees and devices. Companies now need to manage software actors that can move quickly across applications.
Obsidian is targeting that gap.
The bigger picture
Agent security is turning into one of the clearest second-order markets created by generative AI.
The first wave of enterprise AI asked whether companies could build useful agents. The next wave asks whether those agents can be governed safely. Startups that solve permissions, auditability and non-human identity could become core infrastructure for AI adoption.