Google LLC
US technology company, wholly-owned subsidiary of Alphabet Inc. (NASDAQ GOOGL/GOOG) since the 2015 restructuring. EO-Atlas relevance is Google Earth Engine (GEE), a cloud geospatial analysis platform launched in 2010 that hosts over 80 petabytes of third-party satellite imagery (Landsat, Sentinel, MODIS, ALOS/PALSAR-2 and others) and provides distributed compute for global-scale analysis. Google is not an EO operator, sensor manufacturer, or alert producer. GEE is the substrate and distribution channel for GLAD-L, GLAD-S2, RADD, DIST-ALERT, and the Hansen Global Forest Change dataset; producers are UMD, WUR, and NASA respectively.
Google LLC (Alphabet Inc. subsidiary, NASDAQ GOOGL/GOOG) occupies the platform and distribution layer of the Earth observation ecosystem. Founded in 1998 and restructured under Alphabet in 2015, Google does not operate satellites, manufacture sensors, or produce raw imagery data. Its EO relevance is entirely upstream infrastructure and downstream ML-derived data products.
Google Earth Engine (GEE), launched in 2010, is a cloud geospatial analysis platform hosting more than 80 petabytes of satellite imagery, including Landsat 4 through 9, Sentinel-1 and Sentinel-2, MODIS, ASTER, ALOS PALSAR-2, and the ISS-mounted EMIT hyperspectral instrument, among others.[^gee-concepts] GEE provides distributed compute via Python and JavaScript APIs, supporting both interactive and planetary-scale batch processing modes. Since its transition to a Google Cloud product, access is split between a free noncommercial tier (research and nonprofit, subject to verification) and a paid commercial tier billed in Earth Engine Compute Units (EECU).[^gee-access] The platform functions as the compute substrate for a significant fraction of global forest monitoring infrastructure, including Global Forest Watch (WRI-led), MapBiomas, and the GLAD-L alert system (UMD).
Google Cloud Public Datasets extends this hosting role: Sentinel-2 (10-60m, 2015-present) and Landsat Collection 1 archives are mirrored in Google Cloud Storage at no cost to users, providing raw file access independent of the GEE compute environment.[^gcp-public-datasets]
On the derived data product side, Google produces two datasets of direct EO-Atlas relevance. Open Buildings (v3, May 2023) delivers 1.8 billion ML-derived building detections across 195 countries and 58 million km^2, using deep learning on high-resolution satellite imagery; the dataset is dual-licensed under CC BY-4.0 and ODbL.[^open-buildings-v3] Dynamic World V1, developed jointly with the World Resources Institute and National Geographic Society, is a near-realtime global land cover classification at 10-metre resolution, updated every 2-5 days from Sentinel-2 L1C input across nine land cover classes; it is peer-reviewed and CC-BY licensed.[^dynamic-world-paper]
Google is not a data producer for alert systems hosted on its platform. GLAD-L alerts are produced by the UMD GLAD lab; RADD alerts by Wageningen University and Research; Hansen Global Forest Change maps by the University of Maryland. Google supplies the compute and hosting; the scientific content and methodology attribution belongs to the respective producing institutions.