This Milan Startup's 'Hacker Traps' Just Caught $3.3M to Outsmart AI Cyberattacks

Jan de Vries · 2026-07-27

Cybersecurity is feeling the heat. Attackers are now using AI to break into systems faster than most human teams can react. One Milan-based startup, Beelzebub, thinks the only way to fight back is with traps that think for themselves. Beelzebub, an AI-native cybersecurity platform, just announced a $3.3 million Seed funding round (converted from €3 million) led by Italian DeepTech VC United Ventures. The company plans to use the cash to expand its research team, open new offices in Rome and San Francisco by the end of 2025, and go after clients in Europe—especially those regulated under NIS2. They’re also already building tech to protect AI agents directly. This round follows a $330,000 (converted from €300k) pre-Seed from strategic investors, bringing total funding to roughly $3.6 million. ### Why This Matters Mario Candela, CEO and founder of Beelzebub, puts it simply: “Cybersecurity is a nonstop battle, and one that humans can no longer fight alone. AI-powered attackers are too powerful, too competent, too fast, and too numerous.” The platform started as Candela’s research project back in 2021 and officially incorporated in July 2025. Its core idea is clever: instead of just building a wall around your network, Beelzebub assumes the bad guys are already inside. It sets up fake infrastructure—servers, databases, APIs, cloud systems—that look real. When an attacker touches one of these traps, the platform detects and blocks the attack before it can do any damage. ### How It Works Beelzebub’s platform has three main pieces: - **Arcangelo** (red team): Simulates targeted attacks to test your defenses. - **Beelzebub Managed** (blue team): Plants LLM-powered traps throughout your infrastructure. - **Caronte** (AI analyst): After a trap catches an attacker, Caronte reverse-engineers the malware and delivers a full incident report in a fraction of the time a human would need. These three work together in an automated offense-defense loop. No handoffs between tools or vendors. The traps update in real time to mimic new CVEs (Common Vulnerabilities and Exposures), which forces any dormant attackers to reveal themselves. The platform is available as SaaS or on-premises for sensitive environments. Over 60 independent researchers feed live threat intelligence into the system as new attack methods emerge. It’s NIS2-ready by design. ### What’s Next Massimiliano Magrini, co-founder and Managing Partner at United Ventures, explains the big picture: “The cybersecurity market is undergoing a structural reset. When offensive AI scales faster than any human team can respond, the only viable defense is one that operates at the same speed and intelligence.” For Beelzebub, this seed round is about scaling fast. They’re opening offices in Rome and San Francisco, hiring more researchers, and building tools to protect AI agents themselves. The goal is to bring modern cybersecurity to companies that can’t afford to compromise.