NASA and IBM Deploy AI Model Trained on 17 Years of Lunar Data to Map Moon’s Ice, Craters and Volcanoes

NASA announced the rollout of a new artificial‑intelligence system designed to accelerate the analysis of the Moon’s surface. The NASA‑IBM Lunar Foundation Model, built on more than a decade and a half of orbital imagery and topographic maps, can automatically identify features such as water‑ice deposits, impact craters and ancient volcanic structures.

The model emerged from a joint effort that pairs NASA’s planetary science teams with IBM Research and a consortium of universities. By feeding the AI a continuous stream of lunar data collected since the early 2000s, the collaborators created a foundation model capable of recognizing patterns that would take human analysts months to discern.

Researchers say the system’s ability to flag ice‑rich regions could prove pivotal for future exploration plans that envision using lunar water as a resource for life‑support and fuel production. Likewise, rapid mapping of crater distributions helps refine impact‑chronology models, while automated detection of volcanic domes offers fresh clues about the Moon’s thermal evolution.

Historically, lunar cartography has relied on manual interpretation of images from missions such as the Lunar Reconnaissance Orbiter and earlier Apollo surveys. Those methods, while thorough, are labor‑intensive and limited by human subjectivity. The new AI approach promises consistent, scalable analysis across the entire lunar surface, opening the door to more comprehensive scientific studies.

The initiative also reflects a broader shift in planetary science toward “foundation models” – large‑scale AI frameworks that can be fine‑tuned for specific research tasks. By establishing a baseline model trained on a wide swath of lunar data, scientists can later adapt it to investigate niche questions, such as the mineralogy of permanently shadowed regions or the morphology of newly discovered pits.

Looking ahead, NASA plans to integrate the model’s outputs into mission planning tools for upcoming lunar landers and habitats. The agency also intends to share the model with the international research community, fostering collaborative studies that could accelerate the timeline for sustained human presence on the Moon. As the AI continues to learn from fresh data streams, its predictive power is expected to grow, offering an ever‑more detailed portrait of Earth’s nearest celestial neighbor.