IBM and NASA's New Moon AI Is Really a Data Infrastructure Project
The model and dataset combine more than 30 aligned layers from nine instruments to support ice detection, crater mapping and volcanic studies.
- NASA and IBM released an open AI model to help scientists identify lunar ice, map craters, and study volcanic features across the Moon.
- Previously, disparate instruments on missions like the Lunar Reconnaissance Orbiter, GRAIL, and SELENE Kaguya collected data at varying resolutions, complicating analysis and creating data silos.
- Performance tests reveal the model reduced error in potential ice detection by up to 22% compared with a SwinV2 baseline, while improving crater detection by nearly 19% using lightweight LoRA adapters.
- NASA's Commercial Lunar Payload Services initiative has awarded 17 lunar deliveries worth up to $2.6 billion through November 2028 to support a Moon Base near the lunar South Pole.
- Complementing these efforts, the project extends the strategy behind the Prithvi Earth foundation model, which researchers deployed aboard orbital platforms in 2026 to create reusable tools for future scientific questions.
21 Articles
21 Articles
For decades, NASA has collected one of the largest data files on the Moon. Now, a new AI tool created with IBM promises to help transform millions of images and maps into clues on where to land, look for ice and study the evolution of the lunar surface
NASA and IBM Release Open-Source Tools to Map the Moon with Python and AI
NASA and IBM have released open-source Python tools and Jupyter notebooks that simplify processing Lunar Reconnaissance Orbiter data into maps, including image alignment, elevation modeling, and AI-based feature detection. The effort lowers technical barriers, promotes reproducibility, and broadens participation in lunar science.
NASA and IBM have developed a new artificial intelligence model specifically designed for analyzing lunar data.
IBM and NASA's New Moon AI Is Really a Data Infrastructure Project
NASA IBM Lunar Foundation Model turns decades of fragmented Moon observations into a reusable scientific system. With more than 30 aligned data layers from nine instruments across four missions, you can see why the bigger breakthrough is the infrastructure underneath the AI.
This update could change the rules of lunar exploration
Coverage Details
Bias Distribution
- 72% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
