AIR secures fifty million pounds for AI agent security

Artificial intelligence security firm AIR has publicly launched with fifty million dollars in seed financing. The capital will support the company’s mission to help businesses monitor and secure the software supply chain associated with autonomous AI agents. The funding was confirmed in a report by TechCrunch, which detailed the startup’s debut on September 1. The company aims to address emerging vulnerabilities in the rapidly evolving ecosystem of AI-driven tools used by enterprises.

The startup was founded by chief executive Yair Saban and chief technology officer Niv Hoffman. Both founders are veterans of Israel’s Unit 8200 intelligence corps. AIR secured the total funding across two seed rounds that closed within weeks of each other. Sequoia Capital led the initial ten million dollar round. This was followed by a forty million dollar investment led by Greenoaks Capital. The later round also saw participation from enterprise security founders and angel investors. This combined approach allowed the company to raise substantial capital quickly to establish its market position.

Corporate AI agents frequently deploy third-party plug-ins and Model Context Protocol servers to connect with internet resources. They also utilise digital skills to autonomously execute tasks across internal databases. As companies grant these agents broader operational access, unvetted software components present significant supply-chain vulnerabilities. Malicious actors can potentially poison the data or code that an agent consumes. AIR’s platform seeks to mitigate these risks by discovering active agents across a network. It flags unauthorized employee tools and continuously checks external components for malicious alterations. The system also blocks non-compliant software. The startup reported that its system currently filters out approximately twenty-seven percent of the agent add-ons and skills evaluated online.

Customer adoption has been strongest in heavily regulated sectors, particularly financial services and pharmaceuticals. AIR currently serves more than twenty corporate clients. Large enterprise customers represent about a quarter of that total. Saban emphasised the current lack of governance across AI agent extensions. He drew a parallel to historical enterprise software challenges, noting that drivers now require signatures to load code into the kernel. He stated that skills and plug-ins lack this same verification mechanism, despite being the same lesson. The funding arrives as financial institutions confront a growing governance gap surrounding autonomous systems. Traditional cybersecurity infrastructure was engineered for static applications rather than agents that interpret natural language prompts. Because financial services firms rely on agents to process transactions and draft compliance filings, unmonitored activity creates severe operational and regulatory risks. These threats highlight the need for dedicated platforms that enforce audit trails and strict access controls before rogue code reaches core banking networks.