Switzerland copies 100 petabytes of NASA data amid US cuts
Swiss researchers have transferred a massive archive of NASA climate data to a supercomputing centre in Lugano, safeguarding critical information needed to develop faster AI-driven weather forecasting for European industries.
Switzerland's Federal Institute of Technology Zurich (ETH) has transferred around 100 petabytes of NASA climate data to the Swiss National Supercomputing Centre (CSCS) in Lugano. The process of copying roughly six billion files took about a year. Researchers are now preparing to acquire large volumes of datasets from the US National Oceanic and Atmospheric Administration (NOAA) to build on this foundation.
The project was partly driven by concerns over sweeping US funding cuts to federal climate science and Earth-observation programmes under President Donald Trump. While all the copied information was publicly available, European scientists moved to secure a local copy before any potential restrictions. "So far, the US has not restricted access to the data," said Reto Knutti, professor of climate physics and head of ETH's Center for Climate Systems Modeling (C2SM). "But we also know that decisions in the US administration are sometimes quick, and not completely obvious."
For European industries, the economic value of this archive lies in its application to artificial intelligence. "Data is the new gold," Knutti noted. The datasets contain decades of observations regarding greenhouse gases, clouds, precipitation and ice sheets, which are ideal for training AI foundation models. These data-driven systems are already outperforming traditional mathematical simulations used to predict complex atmospheric and oceanic processes. "Some of the foundation models are getting really, really good in terms of prediction skill," Knutti added.
Speed is the primary commercial advantage of this transition. Instead of requiring hours of processing time, the new AI models can run a global, multi-day weather forecast in about a minute. This makes them up to 1,000 times faster than physics-based alternatives. Such rapid forecasting has direct economic implications for critical sectors like agriculture and hydropower, while also providing life-saving early warnings for floods, storms and landslides.
ETH plans to pair the archived data with the Alps supercomputer housed at the CSCS facility. This combination of vast historical data and immense computing power will allow European researchers to train and analyse advanced AI models locally. The goal is to better monitor the impacts of climate change and recognise important patterns more efficiently. "If we collect so much data and nobody is able to make sense of it, then that's kind of a waste of resources," Knutti said.
