Exclusive: Hadrian raises $40m as AI cyberattacks accelerate

I have seen a lot of cybersecurity pitches lately. Most of them follow a predictable script: AI is coming for your data, so buy our AI to protect it. It is a reactive arms race that usually results in more dashboards and more fatigue for the people actually doing the work. But Hadrian just closed a $40 million Series A round that shifts the focus from building taller walls to understanding how those walls actually get kicked down.

Led by Index Ventures, this round signals that the market is finally admitting a hard truth. The old way of doing security—running a scan once a quarter or hiring a firm for a two-week manual audit—is dead. If you are a builder in this space, you know that the interval between a vulnerability being discovered and it being exploited is shrinking to almost zero. Automated attackers do not wait for your scheduled maintenance window.

The Offensive Shift

Hadrian is not building a firewall. They are building a hacker. Their platform mimics the behavior of a malicious actor, constantly probing the perimeter of a company’s digital infrastructure. It is what the industry calls offensive security, but automated at a scale that humans cannot match. For founders, this is the first logical use case for LLMs and autonomous agents that actually makes financial sense.

Think about the typical enterprise attack surface. You have cloud instances, forgotten subdomains, APIs that were supposed to be decommissioned, and third-party integrations. A human penetration tester might find three or four ways in over a week. An automated system that understands context can find those same entry points in seconds and keep looking 24/7. This is the difference between a snapshot and a movie.

Why $40 Million Matters Now

The timing of this funding reflects a specific anxiety in the C-suite. We are seeing a massive surge in AI-generated phishing and automated exploit discovery. Attackers are using large language models to write polymorphic code that bypasses traditional signature-based detection. If the bad guys are using automation to find holes, the good guys cannot rely on manual labor to patch them.

Hadrian’s approach is to treat security like a data problem rather than a perimeter problem. By mapping out the entire attack surface and testing it continuously, they provide a real-time risk score. This is a far cry from the static reports that usually sit in a CTO’s inbox gathering dust. For builders, the takeaway here is clear: security is becoming a continuous integration problem, just like code deployment.

The Problem with Modern Defense

Most security tools are noisy. They generate thousands of alerts, most of which are false positives or low-priority issues. This leads to what I call alert blindness. When everything is a priority, nothing is. Hadrian claims to solve this by focusing on reachability. It is not just about finding a vulnerability; it is about proving that the vulnerability can actually be exploited to reach sensitive data.

The value is not in the list of bugs; it is in the evidence of the path an attacker would take.

If you are building in AI or crypto, you are already a high-value target. The complexity of these stacks makes them prone to configuration errors. In the crypto world, we have seen millions lost not because the cryptography was weak, but because an API key was exposed or a cloud bucket was left open. Automation is the only way to catch these trivial but devastating mistakes before someone else does.

The Founder’s Perspective

From a founder’s perspective, I am skeptical of any tool that promises to be a silver bullet. Cybersecurity is a process, not a product. However, Hadrian is leaning into a reality that many vendors try to ignore: talent is scarce. There are not enough high-level security researchers to go around. By automating the grunt work of reconnaissance and initial exploitation, they allow the human teams to focus on the high-level architecture and remediation.

This round of funding also highlights a geographic shift. While Silicon Valley usually dominates these headlines, Hadrian is part of a growing European ecosystem that is taking a more pragmatic, engineering-heavy approach to AI. They are not chasing the hype of general intelligence; they are building specific tools for specific, high-stakes problems.

What This Means for the Build

If you are developing software today, you need to stop thinking of security as the final step before launch. With tools like Hadrian hitting the mainstream, the expectation is that your infrastructure is being hardened in real-time. This changes the requirements for your DevOps and SRE teams. They need to be able to ingest this offensive data and act on it immediately.

• Continuous Recon: You should assume your public-facing assets are being scanned every hour.
• Reachability Matters: Do not waste time patching vulnerabilities that are behind three layers of authentication and zero-trust architecture unless they are actually reachable.
• Automation Parity: If your defense is slower than the attacker’s automation, you have already lost.

The $40 million infusion will likely go toward refining their AI models to reduce those false positives even further. As these models get better at understanding complex business logic, the line between a human hacker and an autonomous agent will continue to blur. We are moving toward a world where your security stack is essentially a digital twin of your enemy, constantly fighting against your own code to make it better.

The Bottom Line

Hadrian’s rise is a symptom of a larger trend: the professionalization of the autonomous attack. We are entering an era where manual security is a liability. For builders, the message is simple: build as if you are already being breached, because somewhere out there, an agent like Hadrian—or a much less friendly one—is already looking for the way in. The goal is no longer to be unhackable; it is to be faster at finding your own flaws than the person trying to exploit them.

Read the original at Sifted →