Giant 3D map shows almost every building in the world

Researchers published GlobalBuildingAtlas, a near‑global 3D map covering about 97% of buildings worldwide—2.75 billion structures mapped at 3×3 m resolution—with footprints, heights and volumes in Earth System Science Data; the product was generated by applying deep‑learning height prediction trained on LiDAR from 168 cities to roughly 800,000 satellite scenes from 2019. The study finds stark spatial contrasts—Asia contains about 1.22 billion buildings and 1.27 trillion m³ of built volume, while Africa has 540 million buildings but only 117 billion m³—and city‑level per‑capita volume varies widely (e.g., Finland vs Greece, Niger far below the world average). The dataset is intended to improve disaster‑risk assessment, climate modelling, urban planning and UN SDG monitoring and demonstrates that traditional 2D built‑area metrics can obscure important differences in infrastructure and living conditions.

Analysis

Researchers published the GlobalBuildingAtlas on 1 December in Earth System Science Data, delivering a near‑global 3D inventory that covers roughly 97% of buildings (about 2.75 billion structures) with 3 m × 3 m spatial resolution including footprints, heights and volumes. The product was generated by applying deep‑learning height prediction trained on LiDAR reference data from 168 cities (mainly Europe, North America and Oceania) to roughly 800,000 satellite scenes from 2019. The study quantifies stark spatial contrasts: Asia contains ~1.22 billion buildings and ~1.27 trillion m3 of built volume while Africa has ~540 million buildings but only ~117 billion m3, and city‑level per‑capita volume varies widely (Finland six times Greece; Niger 27× below world average). The authors and quoted researchers position the dataset as immediately relevant to disaster‑risk assessment, climate modelling, urban planning and UN SDG monitoring because volumetric measures reveal infrastructure and living‑condition differences that 2D metrics obscure. Market signals attach a mildly positive tone with limited immediate market impact (sentiment_score 0.25; market_impact_score 0.12). Key limitations for commercial use include geographic bias in LiDAR training data and reliance on 2019 imagery, so accuracy and applicability will vary by region and over time.

Key Decisions for Investors

• Consider initiating focused due diligence on public and private geospatial‑data, remote‑sensing AI and catastrophe‑modelling vendors that could commercialize or integrate the GlobalBuildingAtlas, but defer material allocation until concrete licensing or partnership announcements are evident
• Insurers and reinsurers should commission validation pilots to test how incorporating building volumes from this data changes exposure and pricing models, particularly in regions where LiDAR training coverage is thin
• Infrastructure and real‑estate investors should incorporate per‑capita building volume metrics into macro and city‑selection frameworks to refine demand and capex assumptions for urbanising markets
• Monitor for follow‑on developments — temporal updates, third‑party validation studies, government/UN adoption or commercial distribution deals — as triggers to increase conviction or deploy capital