Okta has agreed to acquire AI identity security startup Permiso Security in a deal valued at just under $200 million, structured as an almost all-cash transaction, according to a source with knowledge of the terms. Okta did not dispute the $200 million figure when asked, though it declined to confirm specifics. The purchase is Okta's clearest bet yet that securing AI agents and other machine identities will become a durable line of business as enterprises push autonomous software deeper into production.
The price marks a sharp step-up for Permiso. The startup had raised about $29 million to date, including an $18.5 million Series A led by Altimeter Capital in April 2024 that valued the Palo Alto company at roughly $80 million post-money. At just under $200 million, Okta is paying more than 2x that Series A mark roughly 15 months later — a premium that reflects how quickly the AI agent security category has repriced.
Okta expects the deal to close in the third quarter of its fiscal 2027, subject to customary closing conditions. CEO Todd McKinnon has been pushing Okta beyond its core role of verifying human users at login and toward continuous monitoring of what users, applications, and AI agents do once inside a network. Permiso slots directly into that expansion.
Key facts
- 01Okta is acquiring Permiso Security for just under $200 million in an almost all-cash deal, per a source with knowledge of the terms.
- 02Permiso had raised about $29 million to date, including an $18.5 million Series A led by Altimeter Capital in April 2024.
- 03That Series A valued the Palo Alto startup at roughly $80 million post-money — the exit prices Permiso at more than 2x that mark.
- 04The deal is expected to close in the third quarter of Okta's fiscal 2027, subject to customary closing conditions.
- 05Permiso's SandyClaw platform, introduced in April, sandboxes AI agent skills to flag malicious behavior before deployment.
Permiso emerged from stealth in 2022, co-founded by former FireEye executives Paul Nguyen and Jason Martin. Its software helps security teams spot suspicious behavior in cloud environments after credentials have been granted, with a specialty in detecting attacks that use stolen or compromised identities to move laterally through cloud infrastructure. More recently it extended coverage to AI agents and other non-human identities — the piece Okta is most interested in.
In April, Permiso introduced SandyClaw, a platform that runs AI agent skills inside a sandbox to analyze behavior and identify malicious patterns before the agents are deployed to production. That pre-deployment posture is a different sell from the runtime monitoring most identity vendors are pitching, and it gives Okta a technical wedge alongside its existing identity governance products.
The rationale sits in Okta's broader repositioning. Identity vendors are racing to expand from login verification to full-lifecycle monitoring as enterprises deploy AI agents that act on their own credentials, spin up their own sessions, and touch data far beyond a single authenticated user. Every agent is effectively a new identity to govern, and the number of non-human identities inside a typical enterprise now dwarfs the number of humans.
Competition in the segment has intensified. CrowdStrike, SentinelOne, and cloud-native identity providers have all rolled out agent-identity features in the last year, and standalone startups continue to raise. Okta's response — buying rather than building — signals it wants product parity fast rather than waiting on internal roadmaps to catch up with the market's pace.
For Permiso investors, the outcome is a clean cash exit at roughly 2.5x the last private mark inside two years. For Altimeter, which led the Series A, it is a fast return in a security segment where many startups have struggled to reach meaningful revenue against entrenched incumbents. The all-cash structure also means no lingering equity conversion mechanics tied to Okta's stock.
The open question is integration. Okta's identity fabric is broad but not historically strong in runtime threat detection, and merging Permiso's cloud-detection engine and SandyClaw sandbox into Okta's existing platform without slowing either product will take time. Customers already running Permiso standalone will want clarity on pricing and roadmap continuity, and Okta has not detailed either.
The deal frames the next phase of the identity market. As AI agents become first-class actors inside enterprise systems, the vendors that own the identity layer are the ones best positioned to charge for governing them — and Okta is now paying up to make sure that layer includes AI-native detection. Expect more consolidation in the space through 2026 as the larger identity and security vendors buy their way into agent coverage rather than build it from scratch.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
