Okta Acquires Permiso Security for ~$200M, Expanding AI Agent Protection
Identity platform Okta buys AI security startup Permiso in nearly all-cash deal to bolster machine-identity threat detection as enterprises deploy autonomous software.
Last verified:
Okta Expands Machine-Identity Security with Permiso Acquisition
Okta is acquiring Permiso Security, a cloud-threat detection specialist, for roughly $200 million in a primarily cash transaction, according to TechCrunch reporting. According to TechCrunch, Okta Chief Product Officer Ely Kahn did not challenge the valuation when contacted, though the company declined to disclose formal transaction particulars. The deal is slated to finalize during Okta’s third quarter of fiscal 2027, pending standard regulatory approvals.
The transaction reflects a strategic pivot within identity-management software: enterprises increasingly require continuous surveillance of AI agents and applications after they’ve obtained network access, not merely verification at the login gate. This shift has intensified vendor competition to protect machine identities—the digital credentials and behavioral signatures of autonomous systems—as companies integrate AI deeper into daily operations.
Permiso’s Expansion into AI Agent Monitoring
Permiso emerged from stealth in 2022 and focuses on identifying attacks that exploit compromised identities to traverse cloud infrastructure. The startup, headquartered in Palo Alto and co-founded by former FireEye security researchers Paul Nguyen and Jason Martin, recently broadened its scope. In April 2026, Permiso unveiled SandyClaw, a platform that analyzes AI agent capabilities within an isolated sandbox to flag malicious behavior prior to production deployment.
According to TechCrunch, Permiso has accumulated $29 million in total funding. A Series A round in April 2024, spearheaded by Altimeter Capital and valuing the company at around $80 million post-money, represented the largest single injection at $18.5 million.
Why This Matters
Okta’s move signals that enterprises view AI-agent governance as a critical security frontier. Traditional identity layers—username, password, multi-factor authentication—were built for human users; they do not capture the behavioral anomalies that autonomous systems introduce. By absorbing Permiso’s threat-detection infrastructure and research team, Okta positions its identity fabric to serve both organic human authentication and the emerging class of machine identities.
For security teams already managing Okta’s core platform, the integration offers native visibility into AI-agent activity without integration overhead. For Permiso’s customers and investors, the $200 million exit—on an $80 million Series A valuation two years prior—underscores rapid value creation in the AI-security layer, likely to trigger follow-on funding rounds in the non-human identity space through 2027.
Frequently Asked Questions
What does Permiso Security do?
Permiso detects suspicious activity in cloud environments after identities gain access, and recently expanded to monitor AI agents for malicious behavior before deployment via its SandyClaw platform.
Why is Okta buying a company focused on AI agent security?
As enterprises deploy autonomous software, traditional identity verification at login is insufficient—organizations need continuous monitoring of what AI agents and applications do post-access, a capability Permiso provides.
What was Permiso's funding history?
Permiso raised $29 million total, including an $18.5 million Series A in April 2024 led by Altimeter Capital that valued the startup at approximately $80 million post-money.