Questions
SubTerra Intelligence is a joint venture between Anumá Aerospace, Genetix Computing, and Wieder, Inc., building a mine and unexploded-ordnance (UXO) detection and mapping platform. The three companies each hold an equal one-third stake in the venture.
It combines three layers into one system: the Anumá Airship carries sensor payloads over contaminated ground, up to nine sensor technologies collect the raw data, and Genetix Software analyzes that data to identify and map buried mines and unexploded ordnance.
Genetix Software only classifies a finding when the evidence matches a validated precedent. When a case is uncertain, it escalates to a human expert rather than guessing, and every verified answer is permanently added to the system's ground-truth library. That is a meaningful difference from typical AI systems, which output a probability rather than a checked answer — and here, a wrong guess is not an acceptable failure mode.
The Anumá Airship is at TRL 4–5 on the standard 1–9 Technology Readiness Level scale — already built and flight-tested, not a concept. Genetix Software's core detection engine is roughly 90% built. The next phase of work is field validation in Ukraine.
We start with agricultural land in Ukraine, where the need is most immediate. From there, the same platform extends to roads, forests, and every other terrain that can be mapped from the air — working toward all 139,000 square kilometers of contaminated territory over time, at an estimated remediation cost of €35 billion.
Not today. The platform operates on land that's already reachable. As more land becomes reachable over time — whether the war ends or the front line shifts — the same platform extends there too, working toward all 139,000 km².
Hardware manufacturers have little incentive to build open software that connects their own sensors to a competitor's equipment. The humanitarian mine-action sector is fragmented and has historically been underfunded. And subsurface signal processing is a genuinely harder problem than aerial image recognition. All three of those conditions are changing at once, which is part of why the timing works now.
Reach out through our contact page — we are actively in conversation with funders, mine-action organizations, and technology partners.