Kevin Mandia’s Armadin raised $190M in March to automate red-teaming with autonomous AI agents trained to find zero-days. The Register reported on the threat case from former CISA and NSA officials: every API and tool granted to an agent expands attack surface. Surveys show 48% of security professionals rank agentic AI as their top 2026 threat.
The concern is grounded. But Armadin’s founder describes the exact problem his company was built to solve. That alignment is obvious. Evidence is thin: a survey number, vague “machine-speed” references, an unverified Taiwan incident, and zero published research.
Agent access control matters. The risk is real. But the loudest voices have the most to gain from making it urgent. This is how emerging security categories develop: real risk and sales velocity.
Sources: If you’re not using AI to attack your own systems, your adversaries will • Armadin Secures Record-Breaking $189.9M in Seed and Series A Funding
Coverage: Agentic AI Attack Surface: Why It’s the #1 Cyber Threat of 2026 and How to Secure It
Related on this blog: [QT] The AI Agent Security Threat: Serious Concern, Thin Evidence • [AUTO] AI Phishing Defenses Outpaced by AI Attack Volume • [AUTO] Newer Models Escalate Faster