Onyx Security, a startup focused on securing enterprise use of artificial intelligence agents, said it has launched with $40 million in funding from venture firms Conviction and Cyberstarts.
The company said its platform is designed to help businesses manage the risks associated with increasingly autonomous AI systems that are being deployed across departments such as engineering and customer service.
AI agents are rapidly becoming a core part of how companies operate, but their unpredictable behavior can create operational and security risks. Errors in reasoning or so-called hallucinations can lead to unintended actions, while attackers may exploit AI systems using specially crafted prompts.
Onyx says traditional cybersecurity tools are not designed to manage systems that act autonomously within business workflows.
To address that challenge, the company is building what it calls a Secure AI Control Plane, a centralized platform designed to oversee AI agents and allow companies to adopt them safely across their organizations.
The platform is designed to continuously identify AI agents operating inside a company, monitor their reasoning processes and approve or correct their actions to ensure compliance with security and governance policies.
“Every enterprise is becoming an agent operator — whether they planned to or not,” said Maxim Bar Kogan, co-founder and chief executive of Onyx. “Agents are given access to the most critical systems in the enterprise, but what are our guarantees they will not make serious mistakes or get compromised?”
Bar Kogan said the goal is to ensure companies can define which systems AI agents are allowed to access and under what conditions.
The platform is intended to support multiple departments within companies. Security teams gain visibility into AI activity and protection against attacks, while governance teams can more easily meet regulatory requirements. Infrastructure teams can manage AI systems while balancing cost, performance and reliability.
The system uses proprietary AI models and supervisory agents designed to analyze how AI systems reason and intervene when necessary. The platform can block actions, require human approval or redirect an AI agent’s decision-making in real time.
Hila Zigman, a partner at Cyberstarts, said the rise of AI systems integrated into business operations creates a new security challenge.
“These are not just software tools to be protected — they are systems that make decisions, access sensitive information and are integrated into critical business processes,” Zigman said. “Onyx is building a solution designed for the AI era rather than trying to adapt legacy tools.”
The company was founded by Maxim Bar Kogan, a cybersecurity executive who served in Unit 8200 of the IDF, and Gil Elbaz, an AI researcher who previously worked under the chief technology officer of Nvidia and served in an IDF AI research unit.
Onyx said it already counts multiple Fortune 500 companies among its customers and has more than 70 employees across Israel, the United States and Canada.
The company said the new funding will be used to expand engineering and product teams, develop additional AI models and scale sales efforts as demand grows for systems that can oversee AI agents in large enterprises.