We launched Atlas in January 2025 with 0 hectares under monitoring. By December, Atlas was actively tracking 499,642 hectares — not as a static snapshot, but as change through time, with methods and provenance you can trace end-to-end.
That progress happened in the real world, with real teams. 187 users across 25 organisations in 7 countries used Atlas to replace assumptions with evidence — and to make nature legible inside decisions that normally move faster than ecosystems can recover. The goal wasn’t “more data”. It was clearer decisions. Over the year, Atlas produced 92 insights: defensible conclusions that teams could carry into planning, governance, risk management, and reporting — without having to dilute the science to make it usable.
Some of the organizations active on Atlas at the end of December 2025:
Atlas also reflected the reality of ecosystems: complexity is not noise; it’s the signal. We grew from 13 to 19 Key Nature Indicators, spanning 6 dimensions (water, biodiversity, carbon, climate, vegetation, and risks), because nature doesn’t live in silos. Ecosystems are interconnected systems: trade-offs are common and second-order effects are the rule. The path to understanding isn’t simplification; it’s embracing the multidimensional and making it operational — fusing sources, maintaining scientific traceability, and turning complexity into usable intelligence. That’s what frontier tech is for.
That translated into concrete magnitudes. Across the monitored areas, Atlas consolidated 35.9 megatonnes of total biomass carbon, 143,701 biodiversity observations, 6,958 observed species, and 3,541 IUCN Red List species — with an average LemuRank of 64% (the higher the percentage, the higher the conservation priority). More than statistics, these are context for deciding what to protect, where to act first, and why.
In parallel, Lemu Nge kept delivering the kind of consistent observation that makes Nature Intelligence scalable. Across 5,446 full orbits, we executed 411 sessions that produced 276 captures, generating 8,000+ images — one for each of the 32 hyperspectral bands — and 179.22 GB downlinked from space, representing 291 gigapixels. It’s operational proof: hyperspectral data with the quality needed to infer biodiversity can be collected reliably from orbit, then processed into something actionable on the ground.
If 2024 was the year Lemu Nge reached orbit, 2025 was the year Atlas became operational. Now 2026 is set up to be the year impact compounds — scaling to more customers in more countries, earning trust through successful deployments, and showing, repeatedly, that nature doesn’t compete with business: it strengthens it, with measurable financial returns.
