Ent.AI, a Santa Clara, California-based cybersecurity startup, emerged from stealth this week with $100 million in seed funding to build what it calls an “intent-aware” security platform for enterprises deploying AI agents, according to Crunchbase News.
The round was led by Decibel Partners. Craft Ventures, Crosspoint Capital Partners, Felicis, IQT (the CIA’s venture arm), Sequoia Capital, and Shield Capital also participated, per the Crunchbase report. The investor lineup signals how seriously the national security and enterprise VC communities are taking agent-specific threats.
What Ent.AI Is Building
The company was founded by former RiskIQ executives and members of Microsoft’s Security Copilot team. Its website describes the platform as providing “intent-aware security for the enterprise.”
According to Crunchbase News, Ent.AI’s platform analyzes both user and AI-agent behavior in real time to proactively prevent cyber threats. The thesis: as autonomous agents (OpenClaw, Claude Code, and others) run on enterprise machines with shell access, file system permissions, and API credentials, traditional security models designed for human-only workflows miss agent-specific anomalies.
Why $100M at Seed
A $100 million seed round is outsized by any standard. For context, that figure ties Ent.AI with Hydra Host (GPU infrastructure) and Twenty Technologies (cyber warfare) in Crunchbase’s weekly top funding list, which was led by Odyssey’s $310 million Series B.
The size reflects a bet that agent security is a category, not a feature. Enterprises deploying agents at scale need behavioral analysis systems purpose-built for autonomous workflows. Insider threat detection, SIEM tools, and endpoint protection were designed to model human behavior patterns. Agents exhibit fundamentally different activity: sustained API calls, automated file access, shell command execution, and tool invocation sequences that look anomalous under human-centric models.
The Competitive Landscape
Ent.AI enters a category that barely existed six months ago but is attracting rapid investment. NVIDIA launched SkillSpector, an open-source static security scanner for agent skills, earlier this month. Arcade AI raised $60 million in Series A funding for agent authorization infrastructure. Dream, an Israeli cybersecurity firm, closed a $260 million round and announced Hero, an autonomous vulnerability agent.
The pattern is consistent: enterprise buyers recognize that agent governance and security are blocking issues for production deployment. Ent.AI’s approach, focusing on real-time behavioral analysis rather than static scanning or access control, targets a different layer of the security stack.
The Signal
IQT’s participation is notable. The intelligence community’s venture arm investing in agent-specific security infrastructure suggests that government and defense agencies are deploying or planning to deploy autonomous agents and need security tooling that understands how agents behave differently from human operators.
Ent.AI has not disclosed its founding date, team size, or product availability timeline beyond emerging from stealth.