Cyber Security

Okta agrees to buy Permiso Security to expand identity threat detection for AI agents

Okta agreed to acquire Permiso Security, adding identity threat detection tools for cloud, machine identities, and AI agents.

#AI agents #Cloud Security #Identity Security #Okta #Permiso Security

CyberOGZ Team • August 01, 2026 • 4 views • 3 min read

Format News Brief

Read Time 3 min

Category Cyber Security

Updated Aug 01, 2026

Okta has signed a definitive agreement to acquire Permiso Security, a cloud-native identity security company focused on detecting and responding to threats across human users, machine identities, and AI agents. The deal gives Okta a clearer path from login-centered identity management into ongoing identity threat detection and response, an area that is becoming more important as companies put autonomous software agents inside production workflows.

The company announced the agreement on July 30 and said Permiso's capabilities will be folded into the Okta Platform after the transaction closes. Okta did not disclose financial terms, but TechCrunch reported that the acquisition is valued at just under $200 million and is structured as an almost all-cash deal, citing a source familiar with the transaction. TechCrunch also reported that Okta expects the transaction to close in the third quarter of its fiscal 2027, subject to customary conditions.

Why it matters

Identity systems have traditionally concentrated on authenticating a user or service at the moment of access. The threat model is changing. Cloud workloads, service accounts, and AI agents can keep acting after they have been granted permission, which means defenders need to spot risky behavior after authentication, not only block bad logins at the front door.

Okta says Permiso brings behavioral analytics, identity risk signals, and real-time threat detections that can help customers surface excessive privileges, unusual agent behavior, policy violations, and high-impact access patterns across multi-cloud environments. The company also highlighted SandyClaw, Permiso's sandboxing platform for analyzing AI agent skills and prompts before deployment, as part of the technology it plans to add to its identity security fabric.

The acquisition also shows how quickly enterprise security vendors are repositioning around non-human identities. AI agents, scripts, service accounts, and application-to-application connections can carry broad privileges while producing behavior that is less predictable than a normal employee login. If compromised or misconfigured, those identities can become a route for lateral movement through cloud and SaaS systems.

Okta frames the deal as an expansion of identity threat detection and response across the full identity lifecycle.
Permiso's P0 Labs research team is expected to add post-authentication cloud and AI threat intelligence to Okta's detection work.
The transaction follows a broader industry push to treat AI agents as identities that need governance, monitoring, and containment.

For customers, the practical promise is a more unified view of what authenticated identities are doing after access is granted. For the security market, the deal is another sign that agentic AI is pulling identity management, cloud detection, and security operations closer together.