Onyx Security, whose platform enables enterprises to deploy AI agents securely, raised a $113M Series B led by Bessemer, taking its total funding to $153M
Just four months after emerging from stealth, the Israeli startup is betting that securing AI agents will become one of cybersecurity's biggest markets.
Context & Ripple Effects
Onyx’s latest round follows its $35M Series A for managing AI-agent operational risks in March, marking a rapid escalation in resources for the same enterprise-security thesis.
The financing arrives as nearby category peers are also raising: Hush recently secured a $30M Series A to control enterprise AI agents. Together, the coverage suggests buyers and investors are treating agent governance as a distinct security control point rather than a feature of general AI tooling.
First-order effects
- Onyx gains capital to expand its platform and compete for enterprise deployments of secure AI agents, after emerging from stealth only four months ago.
- Bessemer’s lead investment and the jump to $153M in total funding give Onyx a substantially stronger funding position in the AI-agent security market.
Second-order effects
- Other vendors focused on controlling or securing enterprise agents, including Hush, face a better-funded rival and a clearer imperative to differentiate their governance and operational-risk capabilities.
- Enterprise security teams evaluating agent rollouts are likely to encounter a more crowded set of specialized suppliers, increasing the importance of evaluating how controls fit existing security operations.
Third-order effects
- If comparable funding and customer adoption persist, AI-agent governance could consolidate into a durable security category spanning permissions, oversight and operational-risk controls rather than remain an experimental add-on.
- The category’s development will depend on whether enterprises deploy agents broadly enough to create repeatable demand; funding alone does not establish a lasting platform market.
The trend: AI-agent adoption is creating a new security layer focused on governing what autonomous software can access, do and change inside enterprises.