Seven months after a public launch, Kevin Mandia’s Armadin just printed one of the densest late-stage cyber checks of 2026. On October 1, 2026, the Palo Alto AI-native offensive security company announced a $255.5 million Series B co-led by Andreessen Horowitz and Accel. Company press (PR Newswire) and Reuters both state the round brings Armadin’s valuation to over $2.5 billion. Total funding: $445 million. That is not a seed narrative. That is consolidation capital.
Round, as sourced. Co-leads: a16z and Accel. New participants: Bain Capital Ventures and Redpoint. Returning: 8VC, Ballistic Ventures, Google Ventures, In-Q-Tel, Kleiner Perkins, and Menlo Ventures (PR Newswire). SecurityWeek adds the path: about $24 million seed in late 2025, formal launch in March 2026 with $189.9 million already raised, then this Series B. Post-money structure beyond the “over $2.5B” line, ARR, and exact customer logos: Undisclosed.
What they claim to sell. Armadin’s wire copy frames an “agentic” offensive platform: an autonomous swarm of specialized agents that chain low-severity findings into validated kill chains — from unauthenticated remote code execution at the perimeter through lateral movement to cloud compromise — so enterprise and government teams can sever paths before adversaries use them. Mandia’s line on the PR: offense is uniquely advantaged because AI lets attackers find and chain weaknesses faster than human response. Use-of-proceeds language: scale the platform, research, training, and go-to-market. Production use at Fortune 500 and government customers: company-stated, not a CyberMerge census.
Why the investor mix matters. a16z and Accel co-leading a mid-nine-figure B is a category bet, not a seed fashion print. In-Q-Tel’s return signals government adjacency without proving classified revenue. Bain Capital Ventures and Redpoint joining as new money widens the growth-equity bench. None of that invents ARR. It does tell you where dry powder wants density: continuous, AI-driven offense that produces exploitable paths, not another scanner that ranks CVEs in isolation.
POV — mega rounds densify offense, squeeze the middle. Boards should not read this as “AI security is hot.” Read it as capital choosing a chokepoint. Agentic offense that continuously proves exploitable paths competes with — and eventually absorbs — periodic pen-test shops and scanner vendors that still score findings one at a time. Platform buyers will ask whether their existing ASM, vulnerability management, and red-team spend already covers that kill-chain proof, or whether they need a Mandia-class autonomous offense layer. Mid-tier tools without a clear attach to that layer get priced as features inside someone else’s platform renewal.
Underwriting discipline. Assert what is public: $255.5M Series B, co-leads a16z + Accel, valuation over $2.5B (company + Reuters), total raised $445M, seven-month post-launch clock (SecurityWeek). TAM, win rates, and ARR: Undisclosed. Watch next: which offensive-AI or continuous-assessment independents get acquired, which scanner vendors pivot into “agentic” language without product proof, and whether a valuation over $2.5B marks a ceiling that forces category M&A rather than more Series Bs at this altitude.
Desk sheet — sourced only: PR Newswire Oct 1, 2026 (Armadin) — $255.5M Series B; valuation over $2.5B; $445M total; a16z + Accel co-lead; BCV + Redpoint new; returning 8VC, Ballistic, GV, In-Q-Tel, Kleiner, Menlo; use of proceeds platform/research/GTM. Reuters Oct 1, 2026 — same round size and “more than $2.5 billion” valuation. SecurityWeek Oct 1, 2026 (Mike Lennon) — Mandia founder framing; $445M total about seven months after launch; prior $189.9M at March 2026 launch; $24M seed late 2025. ARR / dilution / board seats: Undisclosed.