Last Updated on September 11, 2026 by YeJahan
YEJAHAN DESK
IBM and NASA have released an open-source AI model designed specifically for analyzing decades of data from lunar observation missions. The Nasa-IBM Lunar Foundation Model is now available to scientists worldwide.
The tool was trained on more than 30 layers of data collected by nine instruments during four separate NASA missions, including the successful Lunar Reconnaissance Orbiter mission. This dataset forms part of IBM and NASA’s growing collection of open-source AI models known as Prithvi, which cover a range of applications from geospatial analysis to weather forecasting.
This new model is a game-changer for lunar scientists, making it much easier to pinpoint hidden ice deposits inside the Moon’s permanently shadowed craters. It also automates the tedious process of mapping out craters and ancient volcanic structures with incredible detail—tasks that used to require hours of manual review or flawed, low-res automated tools.
During recent tests, the model easily outperformed current methods, spotting key lunar features up to 23% more accurately than standard alternatives.
Finding water ice on the Moon is a top priority for space agencies like NASA. Ice means access to water and oxygen, which are the exact raw materials needed to manufacture rocket fuel on-site for future missions to Mars. By launching today, this tool is arriving just in time to help plan the logistics for long-term space exploration.
It fits perfectly into NASA’s broader timeline. The Artemis program aims to take astronauts back to the Moon by 2028, using the mission to test technologies that will allow humans to survive on the lunar surface for long stretches and eventually build permanent bases.
