IBM and NASA Develop Digital Twin AI Model of the Sun to Predict Solar Weather
IBM and NASA Develop Digital Twin AI Model of the Sun to Predict Solar Weather
IBM and NASA have introduced Surya, an advanced open-source AI foundation model designed as a digital twin of the Sun. Built on nine years of high-resolution data from NASA’s Solar Dynamics Observatory, the model interprets solar activity to forecast events such as solar flares, solar winds, and storms that threaten satellites, GPS, telecommunications, aviation, and power grids. Surya is freely available on Hugging Face to accelerate scientific discovery and democratize access to AI worldwide.
The Sun, though 93 million miles away, directly impacts modern life. Solar flares and coronal mass ejections can disrupt navigation, damage satellites, and pose radiation risks. Lloyd’s has estimated that a severe solar storm could cause $2.4 trillion in global economic losses over five years, with $17 billion lost in a single scenario. Recent solar activity has already disrupted GPS services, forced airline diversions, and harmed spacecraft hardware.
Surya addresses these risks by unifying diverse datasets into a single AI-powered virtual replica of the Sun. Using a long-range vision transformer and spectral gating, it processes extremely large-scale images with reduced memory usage and greater precision. In testing, Surya achieved a 16% improvement in solar flare classification accuracy and became the first model to visually predict solar flares up to two hours in advance, doubling the lead time compared to traditional methods.
Juan Bernabe-Moreno of IBM Research described Surya as “a weather forecast for space,” offering an unprecedented capability to anticipate solar storms. NASA’s Kevin Murphy emphasized that embedding heliophysics data into AI allows faster and more precise analysis of the Sun’s dynamics, empowering broader understanding of its impact on Earth’s critical systems.
Surya also demonstrated adaptability by integrating data from missions such as the Parker Solar Probe and SOHO, proving effective in predicting flare activity, solar winds, and other heliophysical phenomena. According to lead scientist Andrés Muñoz-Jaramillo of the Southwest Research Institute, the goal is to maximize Earth’s lead time in preparing for extreme solar events.
This initiative is part of IBM and NASA’s broader collaboration in AI, including the Prithvi family of foundation models for weather and geospatial applications. By framing Surya as a digital twin of our star, both organizations provide a powerful tool to safeguard technology infrastructure, advance heliophysics, and expand applications of AI to planetary science and Earth observation.
Source info here – Have a Story? Address it to the Editor and submit it here
NASA’s Digital Twin Research
NASA has been a pioneer in digital twin research since the Apollo program, when engineers created “living models” of spacecraft to simulate and troubleshoot failures in real time, most famously during Apollo 13. Today, its labs focus on high-fidelity digital twin platforms for Earth system modeling, such as wildfire and environmental monitoring, to better predict dynamic natural phenomena. At the Michoud facility, NASA operates a real-time digital twin of its two-million-square-foot rocket assembly plant, combining LIDAR, photogrammetry, and AI for operational simulation. The JSTAR lab also leverages digital-twin simulations to test spacecraft software and enhance mission safety.
IBM’s Digital Twin Labs
IBM’s research labs are advancing digital twin frameworks across industries, using foundation models and generative AI to improve simulation and prediction. Projects include digital twins of battery systems, designed to accelerate innovation in energy storage and grid applications. IBM researchers are also exploring how generative AI can create multiple plausible states of physical systems, supporting real-time monitoring, fault detection, and predictive maintenance in utilities and manufacturing. The labs emphasize the role of digital twins as continuously updated, data-driven replicas that integrate live sensor feeds with advanced modeling—positioning them as powerful tools for engineering, energy, and infrastructure resilience.
Featured image Source: Live Science
Disclaimer
The information provided in this article is for general informational purposes only and from publicly available sources. While we strive for accuracy, we do not make any representations or warranties, express or implied, regarding the completeness, reliability, or validity of the content. This article does not make any direct claims about specific companies, individuals, or organizations. Any references to reports or external sources are for context and do not imply endorsement or verification of any specific allegations. Readers are encouraged to conduct their own research and seek professional advice before making business decisions. We disclaim any liability for any losses or damages incurred as a result of reliance on the information provided.