Every second, the sun breathes. It releases a constant stream of charged particles—an invisible river of plasma flowing through the solar system. This phenomenon, known as the solar wind, shapes the very space environment that Earth lives in. Most days, the wind is gentle, brushing past our planet without consequence. But sometimes, the current surges. Bursts of high-speed particles can batter Earth’s magnetic shield, disturb the upper atmosphere, and ripple down into the technologies we depend on.
In recent years, humanity has learned just how vulnerable we are. In 2022, for example, a powerful surge of solar wind knocked 40 freshly launched Starlink satellites out of orbit, costing SpaceX millions of dollars and reminding the world of our fragile relationship with the star we orbit. Space weather is no longer a distant curiosity—it is a challenge to modern civilization.
Now, a team of scientists at New York University Abu Dhabi (NYUAD) has unveiled a groundbreaking solution: an artificial intelligence system that can forecast solar wind speeds with unprecedented accuracy, offering up to four days of advance warning. Published in The Astrophysical Journal Supplement Series, this research could mark the beginning of a new era in space weather prediction.
A Challenge Written in Light
Predicting the solar wind has always been one of space science’s toughest problems. Unlike weather on Earth, which can be tracked with satellites, radar, and ground observations, the solar wind emerges from deep, turbulent regions of the sun’s atmosphere. The patterns are complex, subtle, and often hidden from human eyes. Traditional physics-based models have made progress, but their forecasts often fall short, leaving satellite operators, astronauts, and power grid managers with little time to prepare.
The NYUAD researchers approached the challenge from a new direction: let artificial intelligence learn directly from the sun itself. Instead of relying only on equations or text-based datasets, the AI analyzed high-resolution ultraviolet images from NASA’s Solar Dynamics Observatory, a spacecraft that has been observing the sun in exquisite detail for more than a decade. By combining these images with historical solar wind records, the AI was trained to detect faint, often-overlooked patterns that precede changes in solar wind speed.
The result? A forecasting model that improves accuracy by 45% compared to current operational methods, and 20% better than earlier AI attempts. In the world of space weather prediction, where every percentage point can mean the difference between safety and disaster, this leap is monumental.
Seeing the Sun Through AI’s Eyes
What makes this advance so remarkable is the way AI redefines how we “see” the sun. Humans perceive only the visible surface, a bright, shifting face of boiling plasma. But in ultraviolet wavelengths, the sun reveals intricate structures—loops of magnetic energy, coronal holes, and flashes of activity. These features often hold the clues to when and how the solar wind will surge.
The AI system acts like a tireless observer, scanning thousands of these images, learning correlations no human could easily spot. It recognizes when a dark patch in the corona hints at an upcoming gust of solar wind, or when subtle changes in magnetic loops signal turbulence on the way. By transforming this information into predictions, the AI offers not just raw data but actionable knowledge: a warning that can be issued days before a storm arrives.
“This is a major step forward in protecting the satellites, navigation systems, and power infrastructure that modern life depends on,” said Dattaraj Dhuri, the study’s lead author. “By combining advanced AI with solar observations, we can give early warnings that help safeguard critical technology on Earth and in space.”
Why Space Weather Matters on Earth
It’s easy to think of space weather as something far removed from everyday life, but our modern world is more exposed to it than ever. The same solar particles that create breathtaking auroras at high latitudes can also wreak havoc on human systems.
When a strong solar storm strikes, the following risks arise:
- Satellites: High-speed particles can strip electrons, fry circuits, and drag satellites out of orbit by heating the atmosphere.
- Power Grids: Fluctuations in Earth’s magnetic field can induce massive currents in power lines, risking blackouts across cities or even nations.
- Navigation and Communication: GPS systems, radio communications, and even aviation can be disrupted.
- Astronaut Safety: Beyond Earth’s shield, astronauts are vulnerable to radiation exposure that could endanger missions to the Moon or Mars.
As humanity pushes deeper into space with thousands of satellites, lunar bases, and plans for crewed missions to Mars, the importance of accurate solar wind forecasts will only grow. The NYUAD AI model could become a critical shield, not of steel or concrete, but of foresight.
The Future of Forecasting
The breakthrough is more than just a technical achievement—it represents a new philosophy of science. For centuries, humans have studied nature by observing, experimenting, and then building mathematical models. AI offers a new companion to this process, one that can sift through vast amounts of data, uncovering patterns that escape traditional analysis.
The NYUAD team’s work also shows how collaboration across disciplines can fuel discovery. Here, space science met artificial intelligence, and the result was not only a better forecast but a glimpse into how AI may reshape the future of astrophysics. As Shravan Hanasoge, co-principal investigator at the Center for Space Science (CASS), explained, the model bridges the gap between raw solar observations and actionable forecasts, something that could redefine how we defend against space weather.
In the coming years, the model may be refined further with more data, higher-resolution images, and perhaps real-time integration with operational forecasting systems. With each step, humanity gains a little more resilience against the sun’s unpredictable moods.
A New Dawn in Space Weather Protection
The story of solar wind forecasting is more than a scientific triumph—it is a reminder of our delicate place in the cosmos. Earth, for all its beauty and stability, is tethered to a star that can flare and surge with little warning. Yet, instead of fear, we have chosen curiosity. Instead of surrender, we build tools to understand and adapt.
The AI model created at NYU Abu Dhabi is a powerful expression of this spirit. By teaching machines to read the hidden language of the sun, we have taken a bold step toward safeguarding the technologies that define our modern lives. From satellites in orbit to the electricity powering our homes, this work brings us closer to living safely beneath the restless gaze of our star.
More information: Dattaraj B. Dhuri et al, A Multimodal Encoder–Decoder Neural Network for Forecasting Solar Wind Speed at L1, The Astrophysical Journal Supplement Series (2025). DOI: 10.3847/1538-4365/adf436
