Agentic Security: A New Era of Cybersecurity

Oct 7, 2025

By delivering agentic-to-agentic protection, agentic security empowers us a safer future to defend against AI-driven scams and AI agent-powered cyber attacks.

Data Lens

  • In 2025, global cybersecurity threats reached 34.6 billion events, with approximately 40% being AI-driven.
  • By 2026, total threats are projected to exceed 50 billion, fueled by a 47% year-over-year growth rate in AI-driven attacks.
  • By 2027, 17% of all cyberattacks are projected to employ generative AI, with AI-enabled fraud potentially reaching $40 billion in the United States.
  • By 2030, the market value is projected to reach $7.84B at 33.83% CAGR, and documented software vulnerabilities are expected to exceed 1 million.

The Opportunity

AI agents with reasoning, planning, and tool-using capabilities create larger attack surfaces than traditional software. Their ability to remember and recall information opens new types of vulnerabilities that didn't exist before. We're looking for solutions in the following areas.

  • Application-Layer Security: As AI speeds up development cycles, applications change constantly, creating security gaps that established companies struggle to address. This leaves room for startups to provide specialized protection.

  • SOC Automation: Traditional security tools aren't sufficient against AI-powered attacks. Security operations centers need automated systems that can detect, analyze, and respond to threats without human intervention.

  • Agentic-to-Agentic Security: When AI systems interact with each other, they need security measures that work across different network layers. These systems can independently plan and adapt their approaches, requiring security that can match their flexibility.

Sources

Nami Venture Partners