Onyx Security Funding: $113M Series B to Govern AI Agents

A big round for a young company

Onyx Security, a startup with offices in Tel Aviv and New York, has closed a $113 million Series B round. The company says it is building tools to control and govern AI agents operating inside enterprises. The round closed on July 29, and the company is only about two years old 1. The deal reportedly values Onyx at $640 million 1. With this round, the company has raised $153 million in total 2.

Those two figures suggest that Onyx raised roughly $40 million before this round. The Series B therefore makes up most of the capital the company has collected so far. That is a steep climb for a firm this young. It also shows how quickly investors are trying to stake positions in what could become a new category of security software.

What Onyx is selling

Both reports describe the company's focus in similar terms. One says Onyx helps organizations control AI agents 1. The other says it governs them 2. The difference in wording is small, but it points to two sides of the same problem.

• Control is about stopping autonomous software from doing things it should not do.
• Governance is about the wider framework of policy, oversight, and accountability that enterprises need once agents start acting on their behalf.

AI agents differ from chatbots in an important way. They do more than generate text. They can call tools, reach into internal systems, move data, and carry out multistep tasks with little human involvement. As companies wire these agents into workflows, the attack surface and the operational risk grow with them.

Traditional identity and access management was built around human users and fixed service accounts. It was not designed for software that decides what to do next. That is the gap startups like Onyx are trying to fill.

Overlap and divergence in the coverage

The two accounts agree on the core facts: the $113 million figure, the Series B label, and the enterprise AI agent focus 12. Where they differ is in emphasis.

SecurityWeek's coverage centers on the cumulative funding total. That framing places the round within the company's broader fundraising history 2.

Crypto Briefing takes a more pointed angle. It highlights the $640 million valuation as evidence of a "massive bet" on AI governance 1. It also argues that crypto builders should be paying close attention 1. That argument makes some sense. Autonomous agents that hold keys, sign transactions, or interact with smart contracts face a sharper version of the same control problem enterprises face. A mistaken action in an on-chain setting can be irreversible.

Neither account names the investors behind the round, and neither gives revenue or customer figures. So the valuation is better read as a signal of investor appetite than as a verdict on Onyx's commercial traction.

Why it matters

The broader story here is less about one company and more about where AI security dollars are going.

The first wave of AI security spending focused largely on models. It covered prompt injection, data leakage through chat interfaces, and the safety of model outputs. The rise of agentic AI moves the concern toward behavior. The key questions become:

• What is the agent allowed to touch?
• Who authorized it?
• How can its actions be audited or reversed?

A $640 million valuation for a two-year-old company suggests investors believe agent governance will become a budget line in its own right, rather than a feature folded into existing security platforms 1. That belief is not guaranteed to hold. Large security vendors and cloud providers have every incentive to build similar controls into products enterprises already pay for. Fast-growing standalone categories in security have often ended in consolidation, with startups acquired by incumbents seeking to fill gaps.

The reading

Onyx's round looks like a conviction bet placed early. Investors appear to be paying for position in a market that is still forming. The company's Israeli and New York footprint fits a well-worn pattern in cybersecurity: technical talent in Tel Aviv paired with go-to-market reach in the U.S.

Whether Onyx becomes a durable independent platform or an attractive acquisition target, the funding confirms one thing. Securing what AI agents actually do, not just what they say, has become one of the more heavily capitalized questions in enterprise technology.