{
  "id": 4892026,
  "title": "Air Security, which builds a security service for extensions and other tools installed on AI agents, emerges from stealth with $50M led by Sequoia and Greenoaks (Meir Orbach/CTech)",
  "url": "https://urgent.news/2026/09/01/air-security-which-builds-a-security-service-for-extensions-and-other",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-01T15:55:03.000Z",
  "source": {
    "name": "Techmeme",
    "slug": "techmeme",
    "url": "https://www.calcalistech.com/ctechnews/article/r13apdnugg"
  },
  "original_language": "en",
  "account": "Israeli cybersecurity firm AIR Security has unveiled a groundbreaking $50 million funding round, led by Sequoia Capital and Greenoaks, and joined by other prominent investors. The funding marks the company's emergence from stealth mode, following the development of a security platform tailored for AI agents. AIR Security's co-founders Yair Saban and Niv Hoffman, with backgrounds in offensive cybersecurity and enterprise infrastructure, aim to address the unique security challenges posed by AI agents as they increasingly perform important tasks on behalf of enterprises. The company's research indicates that malicious content, compromised tools, and untrusted information encountered by agents can lead to data theft, fraud, and unauthorized access. AIR Security's solution involves an \"inline firewall\" designed to protect the context of AI agents, filtering malicious instructions, untrusted information, and compromised tools before they can influence the agent's behavior. This real-time firewall continuously evaluates and discovers add-ons and plugins across an organization's AI-agent supply chain, enabling security teams to revoke the use of malicious or unapproved components. With a focus on prevention rather than detection, AIR Security aims to provide enterprises with a new layer of AI agent security as they expand the capabilities of these systems.",
  "summary": "The Israeli startup emerged from stealth with research showing thousands of AI add-ons rely on untrusted sources …",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}