Armadin receives $255.5 million to detect vulnerabilities through AI agent swarms

Cybersecurity startup Armadin Inc. announced today that it has raised $255.5 million in funding at a valuation of over $2.5 billion.

Andreessen Horowitz and Accel jointly led the Series B round. They were joined by Alphabet Inc.’s GV Startup Fund, Kleiner Perkins, Menlo Ventures and others.

Armadin was founded last year by well-known cybersecurity entrepreneur Kevin Mandia. He previously founded Mandiant Inc., a security breach investigation company that was acquired by Alphabet for $5.4 billion in 2022. Mandia is also co-founder of Ballistic Ventures, a cybersecurity-focused startup fund that supported Armadin’s newly closed funding round.

His new company offers a platform that uses swarms of artificial intelligence agents to find vulnerabilities in companies’ infrastructure. Armadin says its agents work in a sandbox that blocks potentially harmful activity. The company uses insights from its swarms to help customers improve their cybersecurity defenses.

Swarms of agents have carried out several high-profile security breaches in recent months. One of the reasons for their effectiveness is that they can parallelize cybersecurity tasks. A hundred agents can run a hundred vulnerability scans at once, which is significantly faster than running them one at a time. This acceleration leaves companies with less time to mitigate the threat.

AI agent swarms also have other novel abilities. When an agent discovers a vulnerability, he can share his findings with the other members of his swarm, improving their hacking skills. The OpenAI Group fraudulent PBC agents behind the Hugging Face Inc. breach successfully used this tactic during the cyberattack. She created a spontaneous forum to exchange credentials and discuss hacking tactics.

Armadin says its platform can deploy tens of thousands of AI agents in cybersecurity assessments. In the initial phase of an assessment, the swarm captures an organization’s internal systems and users. These assets are then scanned for vulnerabilities. Armadin claims that its agents can link multiple vulnerabilities together to carry out cyberattacks.

In one customer project, the company founded no fewer than 26,000 agents. They searched more than 25,000 client company assets and found 38 verified attack paths, which are sequences of malicious actions that can lead to a breach. Armadin describes the exercise as the largest autonomous AI attack ever.

The company says its platform not only uncovers vulnerabilities but also generates advice on how to fix them. The software highlights high-severity items to ensure cybersecurity teams address them first.

“The offense has a unique advantage right now,” Mandia said. “AI allows an attacker to find and chain vulnerabilities faster than any human team can respond. The only way to build a defense that can keep up is to train it every day against the best offense available. That’s what we’ve built.”

Armadin will use the newly raised funds to accelerate its research, AI training and go-to-market programs.

Picture: Pixabay