.svg){kind=logo}
Variable-resolution grids, JEDI-powered data assimilation, ensemble forecasting at scale.
Sharper grids. Smarter AI. No excuses.
GRAF's grid tightens where the atmosphere demands it, resolving weather at kilometer scale. Coarse over open ocean, fine over the storm, so compute goes where the forecast is hardest, not everywhere at once.
First implementation of the JEDI system, advancing how machine-driven forecasting improves at the model level. Every observation, satellite, radar, aircraft, and ground station, folded into the model as it runs.
GRAF runs around the clock, refreshing the global forecast every hour and producing 2.2 billion forecasts a day for 2.2 billion locations. Ensembles quantify the uncertainty, so operators see not just the forecast but how confident it is.
2.2 billion forecasts produced daily · trillions of API calls per month
Machine-learning methods reconciling multiple model outputs, running in production long before the hype cycle gave it a name.
A proprietary global atmospheric model built at the intersection of supercomputing and machine learning, not a wrapper on someone else's.
A proprietary network of ground-truth weather stations feeding the models real-world data most AI-weather startups don't have access to.
To put GRAF behind the decisions that can't wait for certainty, talk to our team.
Contact us