Cygrid raises €2.66M to police what AI agents do inside companies
What's the deal? CygridDealroom has a profile for this one. Try Dealroom →, a Berlin-based cybersecurity startup, has exited stealth with a €2.66 million pre-seed round for its identity-enforcement platform that controls access to confidential company data. AtlanticDealroom has a profile for this one. Try Dealroom → led the round, with participation from D11Z, the venture arm of the Schwarz GroupDealroom has a profile for this one. Try Dealroom →, plus angel investors Boris Gorin, Andreas BermigDealroom has a profile for this one. Try Dealroom →, and Vlad TropkoDealroom has a profile for this one. Try Dealroom →.
What's the endgame? Founded in January 2026 by Dmitry Kharchenko and Dekel Paz, both former members of Israeli Military Intelligence, Cygrid wants to control what AI agents actually do inside enterprises. Unlike conventional identity tools, which grant access at sign-in and stop there, it checks each action an agent takes and blocks or narrows it in real time.
The platform maps human, non-human, and AI agent identities into a single graph, correlating what each is entitled to, what it does, and on whose behalf it acts. The company says this lets it revoke a forgotten token in seconds and stop a risky agent command before it executes.
Why now? Attacks on trusted identities are rising. A 2026 CrowdStrike report found 82% of cyber detections involve no malware; attackers simply exploit leaked tokens or compromised agent providers. Palo Alto Networks says machine identities now outnumber human ones by 109 to 1.
By the numbers: Nine months after incorporation, Cygrid serves 19 design partners across fintech, e-commerce, and SaaS, with six in production. In one deployment, a Berlin D2C company found AI agents and third-party connections its own security team did not know about, then used Cygrid to fix more than 100 risky findings.
The capital will go toward expanding enforcement capabilities, growing the team, and scaling go-to-market across Europe. Cygrid aims to become the European alternative to US and Israeli AI identity-security providers.
What they're saying: "Access control tells you who can log in. It says nothing about what they do next," said Kharchenko, chief executive officer and co-founder. "Cygrid is the platform that does: it decides what each identity may do and enforces it before anything executes."
The signal: With more than 80% of Fortune 500 companies deploying AI agents and nearly a third of employees using connected agents without approval, per Microsoft's Cyber Pulse report 2026, the identity attack surface is widening faster than existing controls can manage. Cygrid's bet is that continuous, action-level enforcement — not one-time sign-in checks — becomes the default as enterprises scale agent use.
Read more: EU-Startups
Image credit: Idaho National Laboratory
