Earth Observation & Weather Data Federation with AI Embeddings
Embed2Scale wants to unlock the true potential of the Copernicus Programme, leveraging AI-based data compression to streamline the exchange of vast geospatial information. This initiative aims to pioneer compressed embeddings, enabling quicker access, decentralized applications and accelerated analytics across four different domains: maritime awareness, aboveground biomass estimation, climate and air pollution prediction, and crop stress & early yield detection.
Revolutionising Geospatial Data with AI Compression
In this introduction to Embed2Scale, our team shares the ambitious vision behind the project: transforming how Earth Observation (EO) data is accessed, processed and utilised. Leveraging AI-powered embeddings, Embed2Scale aims to solve the data gravity problem and democratise EO data, making it accessible for real-world applications like climate change monitoring, vegetation analysis and more. Learn how our innovative AI compressors reduce data size by up to 1000x, enabling faster, more efficient global assessments at scale. Hear from our expert partners on the groundbreaking potential of AI in environmental and satellite data analysis.
Check out latest updates!
Embed2Scale to launch the Earth2Vec Xchange session at ESA-NASA Workshop
The ESA-NASA International Workshop on AI Foundation Model for Earth Observation, taking place from 5–7 May 2025 at ESA’s ESRIN (Frascati, Italy), will feature a dedicated session initiated by Embed2Scale: This interactive ideation session—moderated by Isabelle Wittmann (IBM Research Europe), Conrad Albrecht (DLR) and Thomas Brunschwiler (IBM Research Europe)—aims to officially launch the Earth2Vec Xchange,…
Enhancing Biomass Mapping with Super-Resolution powered by Sentinel-1/2
We are pleased to highlight a new scientific contribution partially supported by Embed2Scale: “GSR4B: Biomass Map Super-Resolution with Sentinel-1/2 Guidance”, which was presented at ISPRS Geospatial Week 2025 in Dubai, on April 8, 2025. This research, led by the University of Zurich team, demonstrates how existing low-resolution aboveground biomass maps can be super-resolved to 10m…