TLDR
- IBM and NASA have released the NASA-IBM Lunar Foundation Model as an open-source tool for lunar science
- The model outperforms existing methods by up to 23% in identifying key surface features on the Moon
- It can detect potential ice deposits, volcanic formations, and craters using multi-mission data
- The model was trained on data from nine instruments across four space missions
- It joins IBM and NASA’s existing Prithvi family of open scientific foundation models
IBM and NASA have launched a new open-source AI model designed to help scientists study the Moon’s surface. The model, called the NASA-IBM Lunar Foundation Model, was announced on September 10, 2026.
What if we could explore the Moon with AI?
Meet the @NASA–@IBM Lunar Foundation Model, a new open‑source AI trained on 17 years of lunar data to map craters, identify geologic features, and even help predict where ice may hide in the Moon’s darkest regions. 🌙 pic.twitter.com/bMWlw9qdU8
— NASA Science (@NASAScience_) September 10, 2026
The tool was built to help researchers analyze decades of lunar data more efficiently. Scientists currently have to sort through maps and images by hand or rely on lower-quality models that lack scientific accuracy.
What the Model Can Do
The AI model pulls together data from nine instruments across four space missions. These include NASA’s Lunar Reconnaissance Orbiter, NASA’s GRAIL mission, and Japan’s SELENE/Kaguya mission.
One key use is detecting potential water ice deposits in permanently shadowed regions of the Moon. Finding ice is important because it could provide water and oxygen for a future Moon base, and fuel for missions to Mars.
The model reduced error in identifying areas likely to contain lunar ice by up to 22% compared to a widely used model called SwinV2-B. That is a meaningful improvement for scientists trying to pinpoint resources on the lunar surface.
The model also helps researchers study volcanic features called Irregular Mare Patches. It showed a 3% improvement over existing methods in capturing the extent of these formations.
For crater detection, it outperformed SwinV2-B by nearly 19% at context-scale resolution, using only half the training data. Crater mapping is used by NASA to pick safe landing sites and plan long-term lunar infrastructure.
Open Data for the Scientific Community
Alongside the model, IBM and NASA released the first open-source unified lunar dataset of its kind. It brings together more than 30 spatially aligned layers from the four missions into a single machine-learning-ready format.
The dataset includes tens of thousands of images and maps showing the geophysical properties of the Moon’s surface and subsurface. Previously, no publicly available unified dataset like this existed.
Kevin Murphy, chief science data officer at NASA, said the goal is to make it easier for scientists to explore and use the data NASA has collected over decades.
IBM’s Juan Bernabe-Moreno said the model gives researchers a shared starting point to explore the Moon at scale, rather than building new systems for every scientific question.
The NASA-IBM Lunar Foundation Model is now part of IBM’s Prithvi family of open foundation models. That family already covers geospatial, weather, and heliophysics applications.
IBM stock was trading down about 1.38% at the time of the announcement.
Stop guessing and start investing with confidence. KnockoutStocks gives you the AI insights, market intelligence, and stock research you need to spot opportunities, cut through the noise, and make smarter investment decisions — all in one powerful platform.
Sign up today and get 50% OFF full access to our premium stock picks.
Simply use coupon code SPECIAL50 at checkout to claim your exclusive discount.
