Soon after large parts of the US were hit by the biggest snowstorm of the season, Nvidia announced a new suite of open-source artificial intelligence models designed to improve access to detailed weather and climate data.
The company said its Earth-2 models will enable users to track changes in weather and climate with greater precision, offering 15-day global forecasts, localised storm predictions and high-resolution snapshots of atmospheric conditions. The tools provide medium-range forecasts alongside highly detailed storm modelling.
Speaking at the American Meteorological Society annual meeting, Nvidia said the release represents “the first open set of models and tools” allowing developers to combine different AI-driven weather and climate capabilities in a single framework. The company said the move will make AI-powered forecasting more accessible to scientists, businesses and government agencies.
Mike Pritchard, Nvidia’s director of climate simulation, said the open-source approach is intended to democratise access to weather and climate intelligence, particularly for less developed countries that are often most exposed to extreme events such as floods and droughts.
“Ensuring these models are open and free is really important,” he said.
The initiative comes amid growing corporate investment in weather and climate monitoring, as concerns rise in the US over access to public weather data following staff cuts at the National Oceanic and Atmospheric Administration under the Trump administration.
A growing number of startups and major technology companies are also entering the AI weather space. Microsoft has developed its Aurora AI forecasting model, while Google has partnered its NeuralGCM model with systems from the European Centre for Medium-Range Weather Forecasts to deliver granular forecasts for smallholder farmers in India.
Beyond the technology sector, companies across industries are investing in new tools to assess the impact of weather and climate change on infrastructure, workforces and supply chains.
Nvidia said AI-driven forecasting can significantly reduce time and costs by replacing traditional supercomputer-based, physics-driven models with data-driven AI systems.
Developers from The Weather Company, Taiwan’s Central Weather Administration and the Israel Meteorological Service are already using Nvidia’s AI capabilities, alongside energy companies including TotalEnergies and Eni. Pritchard said collaboration with The Weather Company led Nvidia to enhance one model so it can distinguish complex precipitation types such as freezing rain, hail and snow.
The breadth of early adoption has surprised the company, he added, pointing to insurance firms using the models to map climate-related risk.
“We don’t know what the ecosystem will build on top of this, but that’s kind of the point,” Pritchard said.
