Onyx Security
A fast-scaling, well-capitalized leader in the emerging AI agent control/governance category, with real revenue growth and a marquee integration partner (Anthropic).
Visit Website ↗ + Add to CompareInnovation Matrix Assessment
Quadrupled revenue in four months post-stealth and shipped an Anthropic integration, indicating rapid product iteration and adoption.
Still early in enterprise deployment maturity despite fast growth; operational track record is short.
Raised a $113M Series B in July 2026 at a $640M valuation led by Bessemer, just two years after founding, with an Anthropic partnership announced in June 2026.
AI agent permissioning and control is a newly emerging category directly created by the shift to agentic enterprise software.
Efficacy claims are largely self-reported growth metrics; no independent red-team or benchmark data was found.
Controlling autonomous AI agent behavior is becoming a top-tier enterprise security priority as agentic AI adoption scales.
Why CISOs Should Care
Onyx gives CISOs a control plane to manage permissions, monitor activity, and prevent risk as enterprises deploy autonomous AI agents at scale -- an increasingly urgent need as agentic AI adoption accelerates.
What Makes It Different
Onyx focuses specifically on the AI control problem -- governing what autonomous agents are permitted to do and monitoring their behavior in production -- rather than general cloud or identity security, and has already integrated with Anthropic for enterprise AI adoption.
The Matrix Verdict
62/100 — INCREMENTAL INNOVATOR
A fast-scaling, well-capitalized leader in the emerging AI agent control/governance category, with real revenue growth and a marquee integration partner (Anthropic).
Editorial Note: Claims vs. Verified Findings
Company-reported quadrupled revenue since a stealth launch four months prior is a self-disclosed growth figure.
Sources
Alternatives to Onyx Security
Adaptive Security
AI-driven platform that simulates deepfake, voice, and multichannel social-engineering attacks to train and test organizations against next-generation phishing.
Quilr
Early-stage agentic AI security startup building a 'Service-as-Software' platform to guard against human-related breaches and secure AI agent…
Tenzai
An agentic AI penetration testing startup building autonomous 'AI hackers' to find and validate exploitable vulnerabilities at a…
Zenity
Governance and security platform for AI agents and low-code/no-code development, securing agent identity, permissions and behavior across the…
Charm Security
Agentic AI workforce that investigates and intervenes on scams and fraud in real time, reading manipulation and intent…
Alice (formerly ActiveFence)
Israeli AI security company (rebranded from ActiveFence in January 2026) offering a lifecycle platform to test, guard, and…