Foundation models accelerate space and climate science

On AI in Action, IBM researcher Campbell Watson explains how foundation models are accelerating discovery across Earth and space science. Moving beyond traditional numerical methods, his team applies concepts from large language models to multimodal satellite data to build powerful, open-source AI systems. In collaboration with NASA and the European Space Agency, they have developed foundation models for Earth observation, weather and heliophysics. They are using AI for sustainability use cases, such as flood detection, biodiversity monitoring and solar flare forecasting. Designed for hybrid cloud environments and even deployed in orbit, these models point toward a future where AI and quantum computing unlock deeper planetary insights.