Security startup Oak emerged from stealth this week with $60 million in seed funding to build what it calls an AI-native Identity Operating System — a unified control plane meant to govern every identity in an enterprise, whether it belongs to a human employee, a machine service account or an autonomous AI agent. The round was co-led by Accel, Greylock Partners and CRV, with participation from Hetz Ventures, AlphaDrive Ventures and a group of strategic angel investors.

Replacing a fragmented identity stack

Oak was founded by Shai Morag, whose previous cloud identity and security startup, Ermetic, was acquired by public cyber company Tenable for $265 million in 2023; Morag stayed on at Tenable as chief product officer before starting Oak with co-founder Tal Marom. The pitch is that legacy identity and access management tools were built for a world of human logins and static service accounts, and are now buckling under the explosion of non-human identities — API keys, service accounts and, increasingly, autonomous AI agents that can create sub-agents, call tools and act on data with far less oversight than a human user.

Oak's platform uses an AI connector framework that reaches into on-prem, cloud, SaaS and homegrown applications to build what the company describes as a continuously updated, live identity graph from raw evidence across every identity type in an organization — replacing what is typically a patchwork of separate identity governance, privileged access and non-human-identity tools with a single system of record. The product is already generally available and running in production at enterprise customers, according to the company.

Why it matters

Oak's launch is the latest sign that "agent identity" is becoming its own well-funded security category, as enterprises that have started deploying autonomous agents run into a governance gap: existing identity systems were never designed to track what an AI agent is allowed to touch, how its permissions should expire, or how to audit its actions after the fact. A $60 million seed round — a large check for a company just leaving stealth — signals investors expect agent-identity management to become a standard line item in enterprise security budgets as agentic deployments scale beyond pilots.