Investors price Kevin Mandia's seven-month-old Armadin above $2.5 billion

Armadin had raised $189.5 million before this round. That is the gap between its $445 million total and the new $255.5 million ^[1]. The company describes the valuation only as "over $2.5 billion" ^[1]. If that is a post-money figure, the Series B bought at most about 10.2% of the company, and the pre-money valuation was roughly $2.24 billion ^[2][3]. Armadin is now valued at about 5.6 times all the cash it has ever taken in ^[4]. The company did not disclose revenue, customer count or contract pricing.

I'd look at Accel's seat first. "That was our conviction when we led Armadin's Series A, and it's even more pressing today," said Ping Li, a partner at Accel ^[12]. Accel led the A and co-led this round with Andreessen Horowitz ^[5]. One of the two firms setting the new price therefore already held a stake that the price marks up. The new names were Bain Capital Ventures and Redpoint, and six existing investors came back, Google Ventures and In-Q-Tel among them ^[2].

The product claim is about how often a company tests itself. According to Armadin's release, frontier AI models are shortening the gap between a vulnerability's disclosure and a working exploit, so "periodic penetration tests cannot keep pace" ^[4]. Scanners fail in a different way, the release says, because they score each finding in isolation ^[7]. Armadin's agents chain individually low-severity weaknesses into validated paths that run from remote code execution at the perimeter, through lateral movement, to full cloud compromise ^[5]. Mandia said "AI lets an attacker find and chain weaknesses faster than any human team can respond" ^[9].

This could go about three ways. Buyers could move the scheduled pen-test budget to continuous agents. A $2.5 billion price needs that outcome. They could also keep the scheduled test and buy Armadin as an extra line beside it, so the security team pays twice and Armadin's market is smaller than its pitch describes. Or the price could mostly reflect Mandia himself. He founded Mandiant, the company Google acquired in 2022 ^[10].

I think investors are paying for the first outcome, while the evidence available supports a narrower claim: that Armadin runs agentic attack campaigns in production for Fortune 500 enterprises and government customers ^[6]. The counter-case is that a founder with Mandia's record can get into those accounts on reputation. If so, production deployments seven months after stealth may say more about his access than about any budget moving. I would be wrong if those customers start saying they cancelled scheduled penetration tests and used the agents in their place.

The money will go to scaling the platform and to research, training and go-to-market ^[8]. "We invested because we believe Armadin will become the defining security company of the AI era," said David George, a general partner at Andreessen Horowitz ^[11].