signal/raiseTrack rounds
Pre-seedCybersecurity · Berlin

Cygrid raises €2.66m pre-seed led by Atlantic to police what AI agents do inside companies

The Berlin startup, founded in January 2026 by two military intelligence veterans, checks every action an identity takes rather than only who can log in.

s/rFunding deskEurope · 2 min read

Cygrid, a Berlin cybersecurity startup, has come out of stealth with a €2.66m (about $3m) pre-seed round led by Atlantic, the pan-European pre-seed investor. The company announced the funding on 22 October 2026.

D11Z.Ventures also invested, alongside angel investors including Andreas Bermig, chief executive of bonify; Boris Gorin, vice president of product at Zscaler; and Vlad Tropko of AmberVale Capital, the company said.

No valuation was disclosed. EUVC dated its report on the round 21 October, a day before the company's release.

“Access control tells you who can log in. It says nothing about what they do next.”Dmitry Kharchenko, CEO and Founder, Cygrid

What Cygrid does

Most identity and access tools decide who may sign in and then step aside. Cygrid argues that this leaves companies exposed once a token or credential is valid, especially as staff connect AI agents and third-party apps to email, code and sales systems.

Its platform connects to a company's existing tools, starting in read-only mode, and maps employees, service accounts, API connections and AI agents into a single graph that records what each identity is allowed to do, what it actually does and on whose behalf. It combines fixed rules with AI-based behavioural analysis to decide what each identity may do. It then acts at three points: before access is granted, by correcting permissions that drift from policy, and in real time through a gateway that checks each AI agent tool call before it reaches the target system. The company says it can revoke a forgotten token in seconds and rewrite or block a risky agent command in milliseconds.

Cygrid said it had stopped more than 2,000 risky identity actions at over 20 companies before launch and that it runs in production on Google Workspace, Microsoft 365, GitHub, Slack and HubSpot. Customers named include Barfer's and Horizn Studios. Sources gave different partner counts: the company and Atlantic said more than 20 design partners, while EU-Startups reported 19, of which six were in production.

Who is behind it

Cygrid was founded in January 2026 by Dmitry Kharchenko, its chief executive, and Dekel Paz, its chief security architect. Both served in Israeli military intelligence. Kharchenko was previously chief technology officer at bonify, which SCHUFA acquired, and held senior technology roles at Friendsurance and Firelayers. Paz led security research teams at Zero Networks, Novartis and LightCyber, now part of Palo Alto Networks.

Marc-Olivier Lücke of Atlantic said AI agents are widening the identity attack surface faster than existing controls can keep up. The company's release and EU-Startups called him a general partner, while Atlantic's own post described him as managing partner.

What the money is for

Cygrid said the funding will go to building out its enforcement features, hiring and go-to-market work across Europe and beyond. EU-Startups reported that the company wants to become a European alternative to US and Israeli providers of AI identity security.

The company also said it won a grant under Germany's federal AI Nation programme and was selected for PwC Germany's Responsible AI programme. The founders plan to present the product at the it-sa security fair in Nuremberg from 27 to 29 October.

Company website: cygrid.io