[Action]: IBM and NASA launched an open-source AI model on Thursday, September 10, to help scientists analyze lunar observation data. [Context]: The NASA-IBM Lunar Foundation Model is part of the Prithvi family of open foundation models covering geospatial and weather applications. [Stat]: The NASA-IBM Lunar Foundation Model was trained on more than 30 data layers from 9 instruments in 4 NASA missions, including the Lunar Reconnaissance Orbiter. [Action]: The AI model identifies potential ice deposits in permanently shadowed areas, maps craters for safe landing sites, and studies volcanic geological features. [Stat]: Benchmark tests show the NASA-IBM Lunar Foundation Model is 23 percent more accurate at identifying lunar surface features than widely used methods. [Reason]: Lunar ice is a priority for space agencies as it indicates the presence of water and oxygen, necessary for future Moon bases and Mars mission rocket fuel. [Timeline]: NASA's Artemis program plans to send astronauts back to the Moon in 2028 to test technologies for sustainable lunar habitation and prepare for Mars exploration.