Swiss researchers train AI on NASA climate data
Researchers at Switzerland's ETH and the Swiss National Supercomputing Centre copied about 100 petabytes — roughly six billion NASA files — onto servers adjacent to the Alps supercomputer over about a year. They are training AI statistical models on that dataset to speed up and expand weather and hazard forecasting, with models able to run a global forecast in about a minute after training. The approach is expensive to train but cheap to run, and could allow more frequent checks and targeted monitoring that help flag hazards such as glacier collapse or flood hotspots earlier.
About 100 petabytes of NASA data were copied to a supercomputer.
Context
NASA made vast climate and Earth observation data publicly available. Swiss teams moved those files near a top supercomputer to train AI models. Next, they will test and use the models for faster forecasts and targeted monitoring.
The full analysis
19 dimensions on this story — world impact, market read, and what happens next.
- Full ContextLocked
- Affected SectorsLocked
- Stock ImpactLocked
- Economic IndicatorLocked
- Investor RelevanceLocked
- Professional RelevanceLocked
- Watch PointsLocked
- Probability of ChangeLocked
- Debate PointsLocked
- Prerequisite KnowledgeLocked
- Follow-up QuestionsLocked
- Pros & ConsLocked