IBM and NASA have released the Lunar Foundation Model, an open AI model for mapping the Moon’s surface. It was trained on more than 30 data layers drawn from nine instruments across four NASA missions, including Lunar Reconnaissance Orbiter data.
Its uses are practical: identifying ice, mapping landing zones and studying volcanic features. Accuracy is reported to be up to 23% better than existing methods. The model joins Prithvi, the IBM–NASA family of open geospatial and weather models.
What stands out is not only the result but the pattern: large institutions now release foundation models openly so others can refine them. The same pattern is emerging in weather, agriculture and disaster mitigation — areas highly relevant to Indonesia.