AIR, an AI security startup, has emerged from stealth mode after securing $50 million in funding through two seed rounds led by Sequoia and Greenoaks. Founded by Yair Saban and Niv Hoffman, both former members of Israel’s Unit 8200 intelligence corps, AIR’s platform focuses on protecting companies by monitoring AI agents operating across their systems. As AI agents employ numerous skills, plug-ins, and other add-ons to interact with company environments and the internet, AIR provides continuous vetting of these components to prevent security risks and unauthorized actions.

The company’s technology detects AI agents running within an organization, identifies the skills and tools they use, and enforces security policies by blocking harmful or unapproved interactions. AIR maintains a constantly updated whitelist of skills and add-ons, thoroughly screening them for changes or malicious intent, with the platform filtering out around 27% of potential components. Besides security, AIR also offers a marketplace of vetted skills and add-ons for AI agents, enabling safer integrations within enterprises.

AIR has attracted over 20 customers, including large enterprises in highly regulated sectors like financial services and pharmaceuticals, where demand for AI oversight is most pressing. The startup operates in a competitive landscape with companies like Noma Security, Zenity, Astrix Security, and Operant AI offering related solutions to monitor, secure, and govern AI agents and their associated components. However, AIR’s continuous vetting approach distinguishes it by tackling the challenge of ongoing re-verification of AI agent add-ons in real-time across complex environments.

With approximately 40 employees, AIR plans to use its new capital primarily for expanding research capabilities and accelerating market growth in the U.S. and Europe. CEO Yair Saban emphasizes that while AI labs may implement internal security measures over time, enterprises will still require an independent solution that works across multiple AI vendors and environments to ensure robust and comprehensive protection against evolving threats in AI agent ecosystems.