Enterprises are handing more of their work to AI agents. Onyx Security just raised $113m to make sure a human can still say no.
The Israeli startup’s Series B, led by Bessemer Venture Partners, values it at about $640m, Calcalist reported. That is a several-fold jump, and it lands just four months after Onyx came out of stealth. Its total funding now stands at $153m.
What Onyx sells #
Onyx calls its product a secure AI control plane. In plain terms, it sits between a company’s AI agents and the systems they touch. It finds the agents running in an environment and watches each step of their reasoning. Then it inspects every action before it takes effect. The bad ones get blocked at runtime.
The scale is already real. Onyx says it secures more than 1.1 million agents across its customers, chief executive Maxim Bar Kogan wrote. It inspects more than 66 million AI sessions in real time. Its customers include Fortune 500 firms in banking, energy, healthcare and insurance. Revolut is one of them.
Empowered but not accountable #
The bet is about timing. “In 2025, less than one percent of actions were made by agents,” Bar Kogan said. “These new autonomous actors will soon be taking the overwhelming majority of actions, which begs the question, how do we guarantee these actions are legitimate?”
That is the gap Onyx is selling into. Agents get access to critical systems and complex jobs. But they are, in the company’s words, empowered without being accountable. In energy, finance or healthcare, an agent’s wrong move at machine speed is not a bad headline. It is an outage, a market event, or a patient at risk.
A crowded, fast-moving category #
Onyx is not alone. Securing AI agents has become one of the hottest corners of cybersecurity, from stealthy Israeli rivals to a billion-dollar acquisition. Investors are betting the category produces very large companies.
Onyx points to traction as its edge. Revenue has quadrupled since it left stealth. In June, Anthropic integrated Onyx to help its enterprise customers adopt AI safely. The urgency has a backdrop. This month one of OpenAI’s own models escaped a test sandbox and hacked a rival.
The founders, Bar Kogan and Gil Elbaz, are Israeli intelligence and Air Force alumni. The company now runs to more than 80 staff across Israel, the US and Canada. The new money goes to training Onyx’s own models and expanding in the US. As AI agents take on more of the work, Onyx is betting the harder problem is not building them. It is keeping them on a leash.
Get the TNW newsletter #
Get the most important tech news in your inbox each week.