This summer, students at the Weatherhead School of Management partnered with a large insurance company to explore how AI supercomputing can transform weather prediction and risk management. Using Weatherhead’s newly deployed AI supercomputer powered by NVIDIA GPUs, the team experimented with cutting-edge forecasting models: transformer-based Pangu-Weather, graph neural network approaches like GraphCast, and diffusion-based systems such as CorrDiff, StormCast, and cBottle. These advanced models were benchmarked against real-time data from national weather providers, allowing the team to evaluate performance across different forecasting horizons and event types.
To translate scientific advances into business value, the team created the Insurance Weather Center—a unified digital hub where forecasts from multiple AI models can be compared, analyzed, and visualized. The platform integrates directly with insured asset and claims data APIs, enabling insurers to overlay severe weather predictions with customer exposure. Crucially, the system can identify policyholders in regions projected to be affected by storms and potentially send alerts ahead of time, helping protect lives, reduce losses, and strengthen customer trust.
This initiative also demonstrates Weatherhead’s deepening commitment to AI-powered business innovation. By investing in world-class computational infrastructure and hands-on partnerships with industry leaders, Weatherhead is preparing its students to harness artificial intelligence not only as a technological tool, but as a driver of strategic advantage in industries from insurance to finance and beyond.