AIR Secures $50M in Funding for Autonomous AI Agent Security

The AI safety startup AIR has successfully closed a $50 million funding round to advance its security platform for autonomous AI agents. According to recent reports, the capital was raised across two distinct stages: an initial $10 million round led by Sequoia Capital, followed by a $40 million round spearheaded by Greenoaks. Additional participation came from Swish Ventures, Netz, and several high-profile angel investors from the cybersecurity and artificial intelligence sectors.

Addressing Vulnerabilities in the AI Agent Supply Chain

Founded by Yair Saban and Niv Hoffman, veterans of Israel’s elite Unit 8200 intelligence corps, AIR focuses on the burgeoning ecosystem of autonomous AI agents. Unlike traditional software, these agents often operate with significant independence, navigating the web, executing code, and retrieving data. AIR’s platform acts as an “inline firewall,” designed to monitor and control these agents in real-time.

• The system discovers all AI agents active within an enterprise network.
• It continuously audits the skills and tools utilized by these agents.
• The platform prevents interaction with non-compliant software or untrusted external sources.

This technology addresses a critical gap in enterprise security, as autonomous agents increasingly interact with internal databases and external APIs without human oversight.

Mitigating Risks of Malicious Plugins and Skills

Research conducted by AIR has identified significant risks within the current AI landscape, uncovering over 17,800 public AI add-ons linked to untrusted sources, totaling approximately 6.7 million installations. Some of these "skills" were found to be impersonating reputable brands like OpenAI and Anthropic to bypass security reviews. To counter this, AIR provides a marketplace of audited AI agent plugins, ensuring that enterprises can expand their AI capabilities without compromising their security posture.

Every enterprise has a firewall protecting its network. Now they need one protecting their AI agents, stated Yair Saban, CEO of AIR.

The startup, which currently employs roughly 40 specialists, intends to use the fresh capital to expand its operations and accelerate commercial activity. As the adoption of Model Context Protocol (MCP) servers and autonomous agents grows, the demand for specialized security layers like AIR's is expected to rise significantly within the enterprise sector.