NASA

Overview

NASA is the US space agency, surfaced in the AI Digest corpus as a government partner for domain-specific foundation models trained on its own satellite instrument data corpora. Opening appearance is a co-release with IBM of an open-source lunar foundation model trained on 17 years of Lunar Reconnaissance Orbiter data.

Timeline

• 2026-10-05-AI-Digest — NASA and IBM shipped an open-source lunar foundation model on 2026-10-04, trained on ~17 years of Lunar Reconnaissance Orbiter data (~2M tile bundles). Headline capability: improved polar-ice-deposit prediction; open-source licensing lets third-party labs fine-tune for other lunar-science tasks (regolith composition, crater dating, resource prospecting). Pattern to watch (government agency + hyperscaler co-release, open-weight, narrow-domain) is the one worth tracking for other NASA / NOAA / ESA data corpora over the next 6-12 months. Log against MOC - AI Infrastructure.

Key Developments

1. Lunar Foundation Model Co-Release with IBM (October 4, 2026): The ~17-year LRO data corpus + ~2M tile bundles represents foundation-model recipes jumping cleanly out of language and vision into single-instrument-stream satellite imagery; co-release with IBM under open-source licensing is the government-agency + hyperscaler shape to watch for other agency data corpora.

Related

See also: IBM, MOC - AI Infrastructure.