AIR Security has emerged from stealth with $50M to build an inline firewall for AI agents. The New York startup closed two rounds in its first six months: a $10M seed led by Sequoia Capital and a $40M round led by Greenoaks Capital Partners, with Swish Ventures and Netz Capital also participating.
Founder Saban started AIR with chief technology officer Niv Hoffman; both spent years doing offensive cybersecurity work in Unit 8200, the Israeli military intelligence corps. Roughly 40 people now work there, joined by a chief strategy officer who previously held the top security job at Disney and Costco.
The problem AIR targets: coding agents in large enterprises fetch third-party skills, plugins and Model Context Protocol servers on their own, often bypassing the security team. AIR sits in that path, mapping the agents running across endpoints and cloud accounts, then screening the instructions, tools and data that flow into each agent’s context before it acts.
Its research found more than 17,800 public AI add-ons accounting for 6.7 million installations that drew instructions from unverified outside sources, including add-ons impersonating Anthropic and OpenAI that could run arbitrary code. Roughly 27% of add-ons the company scans get filtered out. More than 20 companies already use the platform.
Angel backers include Wiz co-founder Yinon Costica, Clay co-founder Varun Anand, Eon co-founder Ofir Ehrlich and Cognition president Zach Frankel. Sequoia partner Bogomil Balkansky describes the challenge as continuous re-verification rather than one-time scanning, since agent components change underneath users.
