• The NASA‑IBM Lunar Foundation Model turns decades of lunar observations into a foundation for discovery, helping scientists surface patterns across data at a scale no single instrument has provided

• The model exceeds widely used methods by up to 23% in identifying key geographic features on the Moon’s surface, including potential ice deposits, craters and volcanic formations, to support a sustained return to the Moon

Sep 10, 2026

YORKTOWN HEIGHTS, N.Y., SEPT 10, 2026 – IBM (NYSE: IBM) and NASA today announced the open-source release of the NASA‑IBM Lunar Foundation Model, one of the first publicly available foundation models for scientific exploration of the Moon, now available. Trained on an extensive lunar observation dataset curated by IBM and NASA researchers, the model can help scientists turn decades of complex, multi-instrument data into insights to support the establishment of a sustained human presence on the Moon.

The Moon’s topography is ever-changing. Over time, the Moon’s surface has formed craters, distributed ice and even experienced volcanic activity. For decades, sensors and instruments have continuously observed the Moon, generating petabytes of data, but to study the Moon’s surface, scientists need to either sift through maps and images by hand or use low resolution, task specific machine learning models. These methods can be computationally intensive and can lack the degree of scientific accuracy needed to identify and analyze geographic features. The newly released NASA-IBM Lunar Foundation Model will help researchers accelerate scientific progress by identifying hidden relationships between many different types and resolutions of lunar data.

Researchers could use the NASA-IBM Lunar Foundation Model to investigate multiple lunar phenomena, including:

• Potential Lunar Ice Deposits: Permanently shadowed regions are among the Moon’s most difficult environments to observe, yet they may contain lunar ice below the surface. Lunar ice indicates the presence of water and oxygen — resources considered essential for a future Moon base and producing rocket fuel for future missions to Mars. The NASA-IBM model combines multimodal and multi-resolution observations to predict where ice may be present on the lunar surface. A NASA-IBM authored technical paper shows that the NASA-IBM model reduced error (RMSE) in identifying areas with high potential for lunar ice up to 22% compared to the SwinV2-B (ImageNet) model.[1]
• Volcanic History: Scientists study lunar volcanic features, called Irregular Mare Patches, to better understand the Moon's volcanic history and thermal evolution. In addition, identifying these changing regions is strategic for future surface operations. Using imperfect labels, the model better captures the extent of the volcanic features than the SwinV2-B (ImageNet model) by 3%, bringing comparable accuracy with greater efficiency and lower fine-tuning costs.[2]
• Crater Detection: Craters are one of the most important and distinguishing features of the Moon and can reveal important clues about its history, such as the age of different terrains, their geology, and the chemical composition of the early lunar interior. Crater mapping also helps NASA select safe landing sites, avoid hazards such as steep slopes and boulders, and plan locations for long-term lunar infrastructure. With the model, researchers can now identify, contextualize, and classify craters at meter-scale resolution with the cited paper showing comparable accuracy as state-of-the-art models like SwinV2-B while offering greater efficiency and lower fine-tuning costs. At context-scale resolution (~100 meters), it outperforms SwinV2-B by nearly 19% using just half the training data.[3]

“NASA has spent decades building an extraordinary scientific record of the Moon, but collecting data is only part of the job,” said Kevin Murphy, chief science data officer and acting chief data and AI officer at NASA Headquarters in Washington. “We also have to make data easier for scientists to explore and use. The NASA-IBM Lunar Foundation Model shows what’s possible when we bring AI to NASA’s petabytes of scientific data. That’s a real opportunity we see with AI: turning large-scale data into new discoveries.”

“Uncovering the mysteries of the Moon requires an ability to learn from an extraordinary volume of scientific data,” said Juan Bernabe-Moreno, Director of IBM Research Europe, UK and Ireland. “The NASA-IBM Lunar Foundation Model gives scientists a foundation to explore the Moon at scale, connecting observations across instruments, revealing patterns that are difficult to see in isolation, and providing an open platform the global research community can build on.”

Despite the wealth of lunar data, no publicly available, unified dataset exists that brings the multi-modal, multi-resolution data into a common framework suitable for modern machine learning. Alongside the model, IBM and NASA scientists built the first open-source lunar dataset of its kind, a unified, machine learning ready lunar dataset aggregating over 30 spatially-aligned layers from nine instruments across four missions. The dataset combines tens of thousands of images and maps showing unique geophysical properties of the lunar surface from NASA’s Lunar Reconnaissance Orbiter (LRO) and NASA’s GRAIL mission and incorporated complementary lunar data from the Japanese Aerospace Exploration Agency’s SELENE/Kaguya for a rich, multi-modal view of the lunar surface and subsurface available for the lunar science community to build on.

The model extends an established IBM and NASA collaboration that transforms valuable scientific data into an openly available foundation for discovery for the entire scientific community. By open sourcing the model, scientists and researchers have access to cutting edge AI systems to accelerate progress in lunar exploration. It joins the Prithvi family of open foundation models, spanning geospatial, weather, heliophysics and now the Moon. Together, these models advance a broader vision: instead of building a new algorithmic system for every scientific question, researchers can start from a shared model and adapt it to new tasks to accelerate discovery across domains.

Maury Chasteau-Simien
IBM Research Communications 
Maury@ibm.com

more articles