Sequoia Capital has led a $10 million seed investment in AIR Security, a startup addressing a critical gap in enterprise AI safety as autonomous agents proliferate across organizations. The New York-based company formally launched this week with $50 million in total funding to build what it calls an "inline firewall" for AI agents, protecting them from malicious or compromised tools and plugins that agents independently select and execute.

Why Are AI Agents Becoming a Security Headache for Enterprises?

As coding agents spread through large enterprises, they autonomously reach for third-party skills, plugins, and Model Context Protocol (MCP) servers without human oversight. Security teams rarely see what gets picked up or executed. AIR Security's research uncovered a troubling landscape: more than 17,800 public AI add-ons, accounting for 6.7 million installations, drew instructions from outside sources that nobody had verified. The company also discovered AI skills impersonating Anthropic and OpenAI that could run arbitrary code once installed. Roughly 27% of the add-ons and skills AIR finds online get filtered out as potentially dangerous.

This vulnerability represents a fundamental shift in how enterprises deploy AI. Unlike traditional software, where administrators control what gets installed, agents operate at runtime with components from sources security teams have never reviewed. The problem isn't just scanning for known threats; it's continuous re-verification, since the components an agent uses can change underneath it at any moment.

How Does AIR Security Protect AI Agents?

• Discovery and Mapping: The platform maps all agents already running across endpoints, cloud accounts, and software-as-a-service applications, then continuously re-checks the components those agents depend on.
• Runtime Blocking: Untrusted tools and injected instructions are blocked before the agent acts on them, preventing malicious code execution even if a skill was previously trusted.
• Verified Marketplace: AIR runs a marketplace of add-ons it has already cleared, giving security teams a curated source of safe components for their agents.

The company's founders, Yair Saban and Niv Hoffman, both spent years in Unit 8200, the Israeli military intelligence corps, doing offensive cybersecurity work. They brought on Ryan Knisley as chief strategy officer; Knisley previously held the chief information security officer role at both Disney and Costco.

"Every enterprise has a firewall protecting its network. Now they need one protecting their AI agents," said Yair Saban, Chief Executive at AIR Security.

Yair Saban, Chief Executive at AIR Security

What Does Sequoia's Investment Signal About AI Security Priorities?

Sequoia Capital's $10 million seed investment, followed by Greenoaks Capital Partners' $40 million Series A, reflects venture capital's growing focus on AI infrastructure and safety. Bogomil Balkansky, a partner at Sequoia Capital, emphasized that the problem AIR addresses is "not a scanning problem" but one of continuous re-verification, since the components an agent uses keep changing. This distinction matters: traditional security tools scan for known vulnerabilities, but AI agents introduce a new class of risk where components evolve in real time.

The funding round also attracted high-profile angel investors, including Wiz co-founder Yinon Costica, Clay co-founder Varun Anand, Cognition President Zach Frankel, and Anne Neuberger, the former White House deputy national security adviser for cyber and emerging technology. This mix of venture capital and government cybersecurity expertise signals that AI agent security is moving from a niche concern to a boardroom priority.

More than 20 companies already use AIR's platform, with about a quarter of them being large enterprises. The strongest demand so far has come from financial services and pharmaceutical firms, two sectors where security breaches carry the highest stakes. AIR plans to spend the new funding on hiring researchers and expanding sales in the United States and Europe.

The timing of AIR's launch reflects a broader shift in venture capital strategy. While earlier AI funding focused on building larger and more capable models, investors are now recognizing that deployment safety and governance are equally critical. Sequoia's investment in AIR Security, combined with its backing of other AI infrastructure startups, suggests that the next wave of AI venture returns will come not from the models themselves, but from the systems that make those models safe and trustworthy for enterprise use.