Why did Onyx Security raise $35M?
Funding to secure agent-driven systems
Onyx Security, an Israeli startup focused on the operational risks posed by AI agents, closed a $35 million Series A led by Conviction after a $5 million seed round in 2024. The company builds tools designed to detect, contain, and manage the unique failure modes that arise when autonomous AI agents are given access to corporate systems and data.
The raise reflects investor concern about a fast-growing problem: as companies deploy agents to automate workflows—accessing email, file systems, scheduling, and transactional systems—those agents introduce new attack surfaces and reliability hazards. Recent research and reporting have shown that multi-step agent behaviors can leak sensitive information, be manipulated into unsafe actions, or simply make catastrophic mistakes when they interact with real-world systems.
What Onyx focuses on
- Agent governance: Controls for who can deploy agents, what tools they can call, and what data they may access.
- Runtime monitoring: Detecting anomalous agent behavior in real time and shutting down or quarantining suspect processes.
- Policy enforcement: Applying enterprise safety rules to automated workflows so that agents can’t exfiltrate data or execute risky operations without oversight.
Why investors care
- Rising demand: Organizations moving from pilots to production need operational guardrails that are different from traditional app security.
- Cost of failure: Misbehaving agents can create business, legal, and regulatory exposure that scales quickly.
- Market timing: With many startups and incumbents building agent products, a specialized security layer is becoming essential infrastructure.
Onyx’s new capital will help the company scale engineering and sales to meet enterprise demand as agent deployments multiply across industries.