Milan Startup Beelzebub Raises $3.4 Million to Build AI Traps for Hackers

Key points

• Beelzebub, founded in Milan in 2025, closed a €3 million (roughly $3.4 million) seed funding round led by venture capital firm United Ventures.
• Total funding raised by the company now stands at $3.8 million.
• The platform uses AI-powered decoy systems to detect intruders who have already broken into a network and contain them without human intervention.
• Beelzebub plans to open offices in Rome and San Francisco and grow its European client base.
• The company's CEO says human defenders alone can no longer keep pace with AI-assisted criminal attacks.

A Milan-based startup called Beelzebub has secured €3 million (about $3.4 million) in seed funding to expand a cybersecurity platform built around a simple, unsettling idea: the hackers are probably already inside.

United Ventures, an Italian venture capital firm, led the round. Combined with earlier investment, Beelzebub has now raised a total of $3.8 million since the company was founded earlier this year.

What does the platform actually do?

Beelzebub's software starts from the assumption that a criminal has already broken into a company's network and is quietly waiting for the best moment to strike. Rather than simply trying to keep attackers out, the platform spreads digital traps across the network to catch them when they move.

Those traps are powered by large language models (LLMs), the same class of artificial intelligence technology that underlies tools like ChatGPT. When an intruder interacts with a fake file, a dummy login page, or a convincing but hollow database, the platform flags it immediately, isolates the affected part of the network, and begins gathering intelligence on how the attack works.

Think of it as a roach motel wired to a security camera and an alarm system, all running without anyone needing to press a button.

The system also runs continuous tests against its own network, simulating the moves a criminal might make to expose weak points before a real attacker finds them. An AI analyst then reads any malicious software it catches and produces a full incident report.

For organisations in sectors such as banking, healthcare, or critical infrastructure that cannot send sensitive data to outside servers, the whole platform can run on the company's own hardware rather than in the cloud.

Who is Beelzebub targeting as customers?

The company is primarily chasing organisations covered by NIS2, the European Union's updated Network and Information Security directive, which is a set of legal rules that force certain companies and public bodies across Europe to meet minimum cybersecurity standards or face fines.

Beelzebub's platform is designed to satisfy several of those requirements out of the box. Live threat intelligence, meaning real-time information about active criminal tactics, feeds into the system from more than 60 independent security researchers worldwide.

"Cybersecurity is a nonstop battle, and one that humans can no longer fight alone," said founder and CEO Mario Candela. "AI-powered attackers are too powerful, too competent, too fast, and too numerous, meaning the only way to fight back is at the same pace and intensity."

The fresh capital will fund a larger research team, new offices in Rome and San Francisco, and a push to sign up clients across Europe.

Beelzebub's raise was first reported by SecurityWeek.