Onyx Raises $113M to Govern Enterprise AI Agents

A Fast Funding Jump for a Young Security Startup

Onyx Security, an Israeli cybersecurity startup, has closed a $113 million Series B round, pushing its valuation to roughly $640 million just four months after it emerged from stealth 1. The round was led by Bessemer Venture Partners, with participation from Cyberstarts, TCV, Conviction, FirstMark, Vintage Investment Partners, QuantumLight and G Squared 2. The speed and size of the raise underscore how urgently investors and enterprises alike are treating the challenge of controlling autonomous AI agents now embedded across corporate systems.

What Onyx Actually Does

Onyx's core product is what it calls a "Secure AI Control Plane," a layer that sits between a company's AI agents and the internal systems those agents are permitted to touch 12. Rather than trusting agents to act independently once deployed, Onyx's platform inspects every action an agent attempts to take and blocks those deemed risky or unauthorized 1. The goal is to give security teams a governance checkpoint for the growing population of AI agents now operating inside enterprise environments, ensuring humans retain oversight and control even as agents take on more autonomous, decision-making roles 2.

Why This Matters for Enterprise AI Adoption

The funding arrives at a moment when businesses are rapidly moving from experimenting with generative AI chatbots to deploying autonomous agents capable of executing multi-step tasks across internal software, databases and customer-facing systems. That shift raises the stakes considerably: an agent with broad system access that misfires, is manipulated, or acts on flawed reasoning could cause damage far beyond what a simple chatbot ever could. Onyx's pitch — and the reason investors moved so quickly to back it — is that enterprises need a dedicated control and inspection layer purpose-built for agentic behavior, not just traditional identity and access management tools retrofitted for AI.

Context and Outlook

Both accounts of the raise agree on the fundamentals: the $113 million Series B, the roughly $640 million valuation, and Bessemer's lead role, with one source detailing the fuller syndicate of backers including Cyberstarts, TCV, Conviction, FirstMark, Vintage Investment Partners, QuantumLight and G Squared 2. That a company barely out of stealth could attract this level of capital and valuation signals strong investor conviction that AI agent governance will become a standard, if not mandatory, layer of enterprise security infrastructure. As more organizations hand off operational tasks to autonomous agents, tools that inspect, constrain and audit agent behavior are likely to become as essential as firewalls or identity management systems were in earlier computing eras — positioning startups like Onyx at the center of a rapidly forming new security category.