IBM is launching a new open source AI model to get NASA back to the Moon — and making petabytes of lunar data available to study

IBM and NASA have collaborated to launch an open source AI model designed to aid lunar research. The new NASA-IBM Lunar Foundation model, available on Hugging Face, enables researchers to analyze decades of lunar observation data, identifying critical geological features essential for NASA's goal of establishing a sustained human presence on the Moon.

Trained on a vast multimodal NASA dataset, the model will also make petabytes of lunar data readily accessible for public usage and analysis. This release aims to streamline the process of studying moon data, which has traditionally relied on manual analysis or low-resolution AI models. By providing a single foundation model that can adapt to various lunar geologic features, researchers will be able to investigate a broader range of lunar aspects with greater accuracy, such as locating potentially hazardous locations, ice deposits, volcanic features, and craters.

This will not only enhance our understanding of the Moon's geological history but also facilitate the safe navigation of future astronauts and the identification of essential resources for establishing a human base on the Moon.