Will the Sentinel‑2 archive be compressed at a ratio of 1000× before 2032?

Prediction market on metaculus. [Sentinel‑2](https://sentinels.copernicus.eu/web/sentinel/copernicus/sentinel-2) is a satellite constellation operated by the European Space Agency that captures multispectral images of Earth’s land and coastal waters. Operating since 2015, data from the project is used across several fields, including agricultural monitoring, disaster management, land cover classification, and water quality. The project produces [several datasets](https://sentiwiki.copernicus.eu/web/s2-products), with Level-1C and Level-2A being the publicly released measurement data used by scientists. As of 2025, the complete L1C+L2A dataset was [over 50PB](https://lps25.esa.int/lps25-presentations/presentations/2505/_2505.pdf), with roughly 3-4PB being added each year.  Due to the massive size, significant research has gone towards compressing Earth observation data. In recent years, [specialised AI models](https://embed2scale.eu/2026/03/16/embed2scale-introduces-terracodec-neural-compression-for-optical-earth-observation-data/) have achieved [substantial improvements](https://remotesensing.vito.be//news/corsa-eo-data-compression-and-processing-ai) over traditional methods. Nonetheless, the full dataset remains unwieldy for local storage.

Resolves: 12/31/2031.