IBM and NASA have released the NASA-IBM Lunar Foundation Model, an open-source AI model trained on decades of lunar observation data from missions and instruments including the Lunar Reconnaissance Orbiter. In benchmark testing, IBM and NASA said the model identified important lunar surface features up to 23% more accurately than widely used methods, supporting tasks such as mapping craters, spotting potential ice deposits and studying volcanic features. The lunar model joins IBM and NASA's broader Prithvi family of open-source foundation models, which already covers geospatial analysis and weather, and NASA has demonstrated Prithvi in orbit, suggesting the technology could run directly on satellites. The project is strategically positive for IBM's AI credentials but carries limited near-term revenue, since the model is open-source and no major commercial product or large new NASA contract was announced. The commercial opportunity depends heavily on the pace of lunar exploration, including NASA's Artemis program and China's plans to search for ice near the Moon's south pole.
International Business MachinesIBM and NASA released the open-source NASA-IBM Lunar Foundation Model, an AI model that outperformed existing methods by up to 23% on lunar surface tasks, boosting IBM's AI credentials.
