Ex-Microsoft Security Copilot founders just raised $100M to secure the endpoint AI agents are taking over

The team that built Microsoft Security Copilot just left and raised one of the largest seed rounds in cybersecurity history.

$100 million. Led by Decibel, Sequoia, Crosspoint Capital, Craft Ventures, Shield Capital, Felicis, and In-Q-Tel.

Their diagnosis: every EDR tool on the market — CrowdStrike, SentinelOne, Defender — was built to catch malware and suspicious code. None of them can tell you whether the AI agent accessing your production database should be doing what it's doing.

The company is called Ent.

Their approach: "intent-aware" prevention. A lightweight agent monitors activity across apps, browsers, data transfers, and local AI systems. Before an action completes, it checks whether that action aligns with company policy. If a coding agent starts pulling files from a database it's never touched, Ent intervenes at the moment of decision — not after the data is gone.

This is not theoretical. Ent says Global 2000 companies in hospitality, financial services, and defense are already running it.

The advisory board includes former CISOs from Google, Aetna, and MassMutual, the former NSA director, and the former VP of Microsoft Azure Cloud Security.

Here's what your CISO needs to hear: you now have AI agents with production credentials, API tokens, and database access. They act exactly like developers doing real work. Your current security stack sees nothing wrong. Ent's entire thesis is that the question has shifted from "is this malicious?" to "should this agent be doing this at all?"

Audit your agent access policies today. Map every credential your AI agents hold. If your security team can't answer which agents have access to what — you are already behind.

SOURCE: https://techfundingnews.com/ent-100m-seed-ex-microsoft-security-copilot-ai-threats/
VERIFIED: TechFundingNews (June 16, 2026), SiliconANGLE (Entro/SailPoint coverage confirms market context), Fortune Business Insights (EDR market data)
SIGNAL: The $100M seed validates that enterprise security must evolve beyond identity-based access for humans to intent-based governance for autonomous agents. CISOs who haven't mapped their agent identity surface are operating blind.