Cybersecurity Startup AIR Secures $50M From Sequoia and Greenoaks to Launch Inline Firewall for Autonomous AI Agents

The Rise of Autonomous Agents and a New Attack Surface

The enterprise technology landscape has shifted dramatically over the past year. Organizations are no longer content with static conversational models; instead, they are deploying autonomous AI agents capable of browsing the web, reading internal communications, executing file operations, and independently utilizing third-party tools. While this evolution unlocks unprecedented levels of business automation, it fundamentally breaks traditional perimeter security frameworks.

Unlike deterministic software that executes hardcoded logic, AI agents make fluid decisions based on the contextual data and instructions they encounter during execution. This dynamic nature exposes a sprawling, fragile supply chain of auxiliary software: plug-ins, Model Context Protocol (MCP) servers, subagents, and third-party skills. Because these components feed directly into an agent's operational memory, a compromised or malicious add-on can silently hijack the agent's behavior, leading to corporate data exfiltration, unauthorized system access, or internal fraud.

AIR Emerges From Stealth With Heavyweight Backing

Recognizing this critical vulnerability, cybersecurity startup AIR has officially broken out of stealth mode, announcing a massive $50 million funding round. The financing was co-led by premier venture capital firms Sequoia Capital and Greenoaks, with additional participation from Swish Ventures, Netz Capital, and high-profile angel investors including executives from Wiz and Cognition. The velocity and size of the funding round underscore the urgent venture capital appetite for foundational AI security infrastructure.

AIR was founded by cybersecurity veterans with deep roots in offensive security research and enterprise architecture. To fortify its market position, the company has also brought on seasoned enterprise security leaders, including former Disney and Costco CISO Ryan Knisley, to steer its strategic trajectory. The freshly injected capital will primarily fund an advanced research laboratory dedicated to mapping behavioral anomalies and threat vectors native to autonomous AI agents.

The Inline Firewall: Protecting Context Over Code

Traditional firewalls and API gateways monitor network traffic or access tokens, but they remain fundamentally blind to semantic intent and contextual manipulation. AIR addresses this gap by introducing what it defines as an inline firewall for AI agents. Rather than merely restricting what an agent is permitted to execute post-facto, AIR's technology intercepts and secures the information entering the agent's context window in real time.

• Discovery and Inventory: Automatically scans corporate environments to map every active AI agent and track its dependencies.
• Continuous Supply Chain Vetting: Evaluates public and private add-ons, skills, and MCP servers both prior to deployment and continuously during runtime.
• Targeted Remediation: When a vulnerability or malicious payload is detected, security teams can instantly trace dependent workflows and revoke access organization-wide.

Uncovering Severe Supply Chain Vulnerabilities

The necessity for proactive agent defense was underscored by baseline research published alongside AIR's market launch. In one comprehensive evaluation, the company discovered more than 17,800 public AI add-ons—accounting for approximately 6.7 million cumulative installations—that relied heavily on untrusted external instruction feeds.

Even more alarmingly, researchers identified malicious AI skills actively masquerading as trusted industry brands, including Anthropic and OpenAI, engineered specifically to bypass standard platform security reviews. Certain rogue skills possessed the capability to execute arbitrary code directly within enterprise host systems. These findings highlight a chilling reality: as agents gain elevated privileges inside corporate networks, bad actors are aggressively weaponizing the agent ecosystem's software supply chain.

Market Implications and What Investors Should Watch

For enterprise executives, the emergence of dedicated agent-security platforms signals the end of the laissez-faire era in corporate AI adoption. As regulatory scrutiny tightens and boardrooms demand measurable accountability over AI workflows, security budgets will inevitably pivot toward context-level defense mechanisms.

Investors and market analysts tracking the enterprise software sector should monitor several key milestones over the coming quarters:

• Enterprise Adoption Metrics: Track how quickly Fortune 500 companies mandate inline context firewalls for internal agent deployments.
• Ecosystem Partnerships: Look for strategic integrations between AIR and major cloud hyperscalers, LLM providers, and enterprise orchestration platforms.
• Regulatory Precedents: Evaluate how emerging global AI compliance frameworks treat liabilities stemming from compromised third-party agent skills and plug-ins.

Ultimately, AIR’s rapid ascent demonstrates that securing the future of artificial intelligence requires looking far beyond model training and data privacy—it demands absolute mastery over the autonomous tools and inputs shaping machine decisions.