AI security funding surged in early August 2026, with 12 cybersecurity companies raising $1.09 billion in disclosed rounds between July 15 and August 4. Nearly 57% of capital ($620M) went to three companies: Horizon3.ai ($250M at $2B+ valuation), ThreatLocker, and Glow. Notably, 7 of the 12 companies (58%) directly protect AI agents, AI-enabled applications, or non-human identities. This signals that agentic AI security has moved from R&D concern to core investment category as enterprises deploy autonomous systems at scale.
Key deals included Zenity raising $125M in Series C (led by Norwest) for AI-agent security governance, and Obsidian Security securing growth capital for SaaS identity protection targeting AI-agent data exposure. Horizon3.ai, providing autonomous penetration testing via its NodeZero platform, raised $250M at a $2B+ valuation with 120% reported ARR growth and 7,000+ customers. The categories winning capital: AI-agent security, identity governance, endpoint protection, and automated security validation. Enterprise urgency is driving the round velocity—Fortune 500 adoption of Claude, ChatGPT, and other tools has created visibility and compliance liability gaps.
For infrastructure architects and security practitioners, this marks a structural shift: AI agent governance is no longer optional, and vendors claiming to handle autonomous system risk are well-capitalized to build at scale. The deployment risk has moved from model risk to post-deployment attack surface. Enterprises will need guardrails spanning prompt injection, data exfiltration via agents, unauthorized agent-to-SaaS access, and third-party supply-chain compromise from AI agents. Buyers should expect M&A consolidation as mega-cap cloud providers (Microsoft, Google, AWS) acquire niche AI security startups to integrate agent governance into base platforms.