Obsidian Security Becomes a Unicorn on $85M Round to Secure Enterprise AI Agents

Obsidian Security, a Newport Beach–based cybersecurity company, has raised an $85 million Series D round at a reported $1.1 billion valuation, joining the growing cohort of AI‑security “unicorns.” Originally focused on monitoring permissions and misconfigurations across SaaS applications, Obsidian is now positioning itself as a control plane for AI agents that operate inside tools like Salesforce, Microsoft 365 and other business platforms.

As enterprises roll out AI assistants that can log into multiple cloud services, move data and trigger workflows autonomously, security teams worry these agents could become a new attack surface—abusing over‑privileged accounts, exfiltrating data or making irreversible configuration changes. Obsidian’s platform aggregates activity logs across SaaS apps, applies analytics to detect anomalous behavior by both humans and AI agents, and helps enforce least‑privilege access. The sizable late‑stage round, led by growth‑equity investors, underscores how quickly demand is rising for products that make AI‑driven automation auditable and controllable. For Japanese firms experimenting with AI agents in global SaaS environments, the deal is another sign that security and governance tooling is becoming a strategic, not optional, part of AI adoption.