IBM and NASA have unveiled what they describe as the most advanced open-source artificial intelligence model for analysing the Sun, aiming to improve forecasting of solar storms that threaten satellites, power grids and telecommunications.
The model, named Surya after the Sanskrit word for the Sun, was trained on nine years of high-resolution data from NASA’s Solar Dynamics Observatory.
Researchers say it can predict solar flares up to two hours in advance, with testing showing a 16% improvement in flare classification accuracy compared with previous approaches.
The launch comes as scientists and policymakers warn of rising risks from space weather. Solar flares and coronal mass ejections can disrupt GPS navigation, trigger blackouts, and endanger astronauts.
Lloyds of London has estimated that a major solar storm could cause up to $2.4 trillion in global economic losses over five years.
“Think of this as a weather forecast for space,” said Juan Bernabe-Moreno, director of IBM Research Europe, UK and Ireland.
“Just as we work to prepare for hazardous weather events, we need to do the same for solar storms. Surya gives us unprecedented capability to anticipate what’s coming.”
Unlike traditional prediction methods, which rely on partial satellite views, Surya was trained on what IBM and NASA call the largest curated dataset of high-resolution heliophysics observations.
The dataset, which has also been released for open use on Hugging Face, is intended to help researchers study tasks such as predicting solar winds, mapping solar flares, and analysing the Sun’s extreme ultraviolet output.
IBM said the move allows scientists and developers worldwide to build on the model for applications ranging from space weather monitoring to protecting agriculture and transport networks.
The project faced major technical challenges because of the size of the images involved, which are 10 times larger than typical AI training data.
IBM engineers developed a multi-architecture solution to handle the scale efficiently, allowing the model to capture solar features at levels of detail previously unseen in large-scale AI.
“We are advancing data-driven science by embedding NASA’s deep scientific expertise into cutting-edge AI models,” said Kevin Murphy, chief science data officer at NASA.
“This empowers broader understanding of how solar activity impacts critical systems and technologies that we all rely on here on Earth.”
