Onyx Security raises $113 million Series B to police enterprise AI agents

Onyx Security raised a $113 million Series B led by Bessemer Venture Partners on July 29, 2026, money it will spend policing the AI agents companies are turning loose inside their systems.

The round brings the two-year-old startup’s total funding to $153 million and values it at roughly $640 million post-money, per figures tied to the deal. Cyberstarts, TCV, Conviction, FirstMark, Vintage Investment Partners, QuantumLight and G Squared also took part.

Onyx sells what it calls a Secure AI Control Plane: software that discovers AI agents running across browsers, endpoints, SaaS apps and cloud accounts, then inspects and can intervene on their actions in real time. The pitch is that autonomous agents now act with employee-level access and need employee-level oversight.

The company says it secures more than 1.1 million agents and has inspected more than 66 million AI sessions, and that revenue quadrupled in the four months since it left stealth. Those figures are Onyx’s own and are not independently verified. It says customers include Fortune 500 firms in energy, financial services and healthcare, and Anthropic partnered with it in June 2026 to help secure enterprise deployments of its models.

The raise is a bet that “agent security” becomes its own enterprise category rather than a feature folded into existing tools. That is unproven: incumbents in identity and cloud security are moving toward the same problem, and much depends on whether agent adoption grows fast enough to sustain a standalone market.