IBM and NASA unveil Surya, an AI model to forecast solar storms

Solar storms

Solar weather has always been hard to predict. Unlike everyday forecasts, scientists only get fragmented satellite views of the Sun, making it nearly impossible to anticipate solar flares or coronal mass ejections with precision. That’s where Surya comes in.

Named after the Sanskrit word for the Sun, Surya is the most advanced open-source AI model built to decode high-resolution solar images and anticipate solar activity. Developed jointly by IBM and NASA, it aims to forecast events that could disrupt satellites, power grids, aviation systems, and even farming — essentially, the backbone of modern civilisation.

Solar storms may sound abstract, but their economic and human costs are very real. Lloyd’s has estimated that a severe solar storm could cost the global economy up to $2.4 trillion over five years, with individual events triggering billions in damages. Recent bursts of solar activity have already forced airlines to divert flights, jammed GPS signals, and degraded satellites. For astronauts and spacecraft outside Earth’s protective atmosphere, the risks are even higher.

Surya was trained on nine years of high-resolution data from NASA’s Solar Dynamics Observatory — images that are 10 times larger than typical AI training data. To handle this scale, IBM had to design a custom multi-architecture system. The result: a model that can spot and classify solar flares with 16% greater accuracy than previous methods.

Crucially, Surya can also visualise where a flare is likely to occur up to two hours in advance — the first AI system to do so. Think of it as a weather radar for space, offering scientists a chance to prepare for incoming solar storms rather than react after the fact.

Surya isn’t staying locked inside IBM’s labs. By releasing it on Hugging Face, IBM and NASA are democratising access, letting researchers worldwide adapt it for their own industries and regions. It also joins IBM’s “Prithvi” family of models, which already includes tools for climate and geospatial research.

As Kevin Murphy of NASA put it: "Surya isn’t just about prediction — it’s about accelerating discovery. In a world leaning more heavily on satellites and space infrastructure, that could make all the difference."