The Pulse
AIR Security Raises $50M to Protect AI Agent Supply Chains
AIR is coming out of stealth with $50 million in seed funding and a platform intended to vet the tools, add-ons and external content used by enterprise AI agents.
AI.info Team ·
AIR Security is coming out of stealth with $50 million raised across two seed rounds and a product designed to inspect the software, add-ons and external content that enterprise AI agents use.
The startup, founded by Chief Executive Yair Saban and Chief Technology Officer Niv Hoffman, offers a platform intended to discover agents operating inside companies, continuously vet the skills, tools and components they use, and block interactions with software or outside sources that do not meet security criteria, according to TechCrunch.
In its September 1 announcement, AIR described its mission as protecting the context in which AI agents operate. The company says it is building a firewall that filters untrusted inputs before they reach an agent’s context, with coverage across endpoint, cloud and software-as-a-service environments.
Supply-chain controls for agent tools
AIR’s premise is that agents can be influenced through more than their underlying model or application code. Skills, plug-ins, Model Context Protocol servers and other add-ons can connect agents to repositories, websites, packages and enterprise data. A component that was previously approved can become risky if a dependency changes or a developer account is compromised.
The platform combines visibility into active agents with an enforcement layer that can intercept actions such as loading a skill or retrieving content from the internet. It also evaluates tools and add-ons against a whitelist maintained by the company, TechCrunch reported. AIR offers a marketplace of vetted add-ons and skills for AI agents.
The company’s own announcement names Sequoia, Greenoaks, Swish, Netz and TTP among its investors.
Two seed rounds closed within weeks
According to TechCrunch, AIR raised the capital in two seed rounds that closed within weeks of one another. Sequoia led the first, a $10 million round, and Greenoaks led the second, a $40 million round. Swish and Netz participated alongside a group of angel investors that included Cognition President Zach Frankel, Wiz co-founder Yinon Costica, Eon co-founder Ofir Ehrlich, former White House deputy national security adviser Anne Neuberger, Omer Adam and Clay co-founder Varun Anand.
AIR has about 40 employees, according to TechCrunch. Saban said the funding will primarily support research hiring and expansion of the company’s go-to-market operations in the United States and Europe.
TechCrunch reported that AIR has more than 20 customers and that roughly one-quarter are large enterprises. Saban said the strongest demand so far has come from heavily regulated sectors, particularly financial services and pharmaceutical companies.
A crowded market for agent security
AIR is entering a field that includes Noma Security, Zenity, Astrix Security and Operant AI, which offer products spanning agent discovery, access controls, governance, runtime monitoring and MCP protections. AIR is positioning its product around repeated checks of the wider ecosystem of components an agent can use.
“This is not a scanning problem, it is a continuous re-verification problem. Inspecting every skill, plugin, MCP server and sub-agent an enterprise’s agents touch, re-inspecting each one every time it changes, in real time and across an entire company’s agent fleet, is an infrastructure problem long before it is a security problem. Air has spent the last year building that pipeline. You do not catch up to it by writing a better scanner.”
Bogomil Balkansky, partner at Sequoia, speaking to TechCrunch
Balkansky described the task as continuously inspecting skills, plug-ins, MCP servers and sub-agents as they change across an enterprise fleet. Saban has similarly argued that companies will seek independent controls that work across different AI providers and agent tools.
AIR’s launch centers on that proposition: enterprises adopting agents need a way to assess the changing external tools and content those systems consume, rather than treating approval as a one-time security decision.