Greek-founded Pallma AI has raised a $1.6M (€1.36M) pre-seed round led by Marathon Venture Capital to build the security layer for autonomous AI agents as enterprises rapidly adopt agentic AI systems.
👯 Founders: Pavlos Mitsoulis (CEO) & Dionysis Varelas (CTO)
📅 Founding year: 2025
🏭 Industry: AI Security / Enterprise Cybersecurity
💥 Problem: As AI agents move from pilots to core enterprise infrastructure, they introduce entirely new security risks that traditional cybersecurity tools can’t handle. Agentic AI systems can be manipulated through prompt injection, leak sensitive data, or execute harmful actions autonomously, often without human oversight, creating enterprise-wide risk.
📣 Solution: Pallma AI is an AI-native security platform built specifically to protect autonomous AI agents. It detects, monitors, and neutralizes threats in real time while continuously hardening AI systems against emerging attack vectors.
👥 Customers: Enterprises deploying AI-powered and agentic applications with deep access to internal systems and data
🌱 Stage: Pre-Seed
💰 Investment amount: €1.36 million
🚀 Funded by: Marathon Venture Capital, with participation from technology leaders at AWS, Meta, Google, Hack The Box, Workable, and others
👁️🗨️ Investor’s perspective:
“Pallma AI offers enterprise-grade agent security, via proactive discovery and runtime enforcement. Their solution has been received with enthusiasm by household name customers, and they are just getting started!” – George Tziralis, Marathon Venture Capital
💡 It will be spent on: Expanding the team across London and Athens, accelerating product development, and meeting rapidly growing enterprise demand.
💪 Their specialty: AI-native security for agentic systems, real-time threat detection, autonomous hardening, and AI red teaming & blue teaming.
🔑 Business model: Enterprise SaaS platform for AI security
👩🏫 Market: Enterprises adopting Large Language Models and autonomous AI agents as core infrastructure.
