Kevin Mandia's Armadin raises $255.5 million at a $2.5 billion-plus valuation

Armadin, the AI security startup led by Mandiant founder Kevin Mandia, raised a $255.5 million Series B at a valuation above $2.5 billion, according to reports published on October 1, 2026. Andreessen Horowitz (a16z) and Accel co-led the round, which brings the company's total funding to about $445 million.

What was announced

• Investors: a16z and Accel (an existing backer) co-led. Bain Capital Ventures and Redpoint joined as new investors. Existing investors 8VC, Ballistic Ventures, GV, In-Q-Tel, Kleiner Perkins and Menlo Ventures also participated, per SecurityWeek.
• Use of funds: scaling the platform and expanding research, model training and go-to-market work.
• Prior funding: SecurityWeek says Armadin appeared quietly in late 2025 with a $24 million seed round and formally launched in March 2026 with $189.9 million. SecurityWeek's reported total of about $445 million matches the Series B plus the March figure, so it may exclude the seed.

The exact valuation is "more than $2.5 billion" and was not attributed to a named source.

What the product does

Armadin sells offensive security testing run by AI. It deploys swarms of specialized agents that probe a company's systems the way an attacker would, then link small weaknesses into validated attack paths, for example from an internet-facing flaw through lateral movement to cloud resources. The report shows defenders the path, its likely impact and how to fix it, instead of a long list of separate findings. The company calls its simulations "AI Hyperattacks" and says agents run through a control layer with a safety model.

In a statement, Mandia said: "AI lets an attacker find and chain weaknesses faster than any human team can respond."

Claims we could not verify

• Armadin says it already runs agentic attack campaigns in production for Fortune 500 companies and government customers. No customers are named.
• In August 2026, Armadin and TENEX.ai ran a three-day live exercise. The company reported 1,300 attacks, 26,000 agents, 238 findings and 38 validated attack paths. These are company figures.
• No revenue, customer count or pricing was disclosed.

Why it matters

Security teams buying software now face attackers using the same AI agents that vendors are selling. Armadin's pitch is continuous automated testing in place of a periodic penetration test, which changes how security tools are priced and bought: by coverage and frequency instead of consultant days. If you run a SaaS product, customers and auditors may start asking for results from automated attack simulation, not only a yearly report. If you buy security tools, ask any vendor in this category for named customers, how the agents are kept inside a safe sandbox, and what happens to data they touch. For the wider shift, see AI agents are coming for the SaaS stack.