AI Security Firm AIR Security Raises $50 Million Seed Funding

AIR Security, a cybersecurity firm developing a supply chain security platform for AI agents, has successfully raised $50 million in a Seed funding round led by Sequoia Capital and Greenoaks Capital. This initial funding round also saw participation from Swish and Netz, alongside notable private investors including Ynon Kreiz, co-founder of Wiz; Zack Frankel, president of Cognition; and Israeli singer Omer Adam. The company was founded in February 2026 by Yair Saban and Niv Hoffman, who leverage their decade-long military experience in offensive cyber, enterprise infrastructure, and AI security research. AIR Security currently employs 40 individuals in Israel, with a primary focus on establishing a robust research lab dedicated to studying AI agent behavior. Ryan Keanelly, former CISO of Walt Disney and Costco, has joined as Chief Strategy Officer.

AIR Security's platform aims to secure the AI supply chain by ensuring that AI agents only interact with safe information, preventing malicious data from compromising their operations. "Organizations are building trust in AI agents and giving them more and more permissions. The agent becomes dangerous when it is exposed to malicious information. We ensure that the information it encounters is only safe information," explained CEO Yair Saban in an interview with Calcalist. Saban also revealed that the company has already secured major clients and begun hiring in the US, emphasizing their focus on prevention as a key differentiator from competitors.

The company has conducted security research highlighting risks in the AI agent supply chain. One study identified over 17,800 public AI plugins, with approximately 6.7 million installations, relying on untrustworthy external sources. Another research effort uncovered AI Skills impersonating well-known brands like Anthropic and OpenAI to bypass platform security checks, with one even capable of executing arbitrary code on enterprise systems. To counter these threats, AIR Security has developed a real-time firewall for AI agents that filters incoming data, blocking malicious instructions and untrusted information before they can influence AI decisions or actions. The platform enables security teams to continuously discover, examine, and manage all AI plugins, Skills, and agents within an organization, both before and after deployment. If a plugin is identified as malicious or vulnerable, the system can immediately locate all dependent agents and processes and revoke its usage across the enterprise.