Nvidia has introduced a new family of open-source AI models designed to revolutionize weather forecasting by making it faster, more accurate, and more cost-effective. The initiative, branded under the “Earth-2” model family, represents the first fully open AI weather stack, allowing developers, governments, enterprises, and organizations to build and deploy tailored weather forecasting applications. Among the models is Earth-2 Medium Range, which employs a novel architecture called “Atlas.” These AI-driven models move beyond traditional physics-based methods, enabling quicker processing of climate data and improved predictive capabilities.
This development aligns with Nvidia’s broader strategic focus on open-source AI software, enhancing accessibility and innovation in the meteorological domain. By open-sourcing these models, Nvidia lowers barriers for stakeholders worldwide to integrate advanced weather insights into decision-making processes, disaster response, agriculture, and other sectors heavily dependent on weather accuracy. The open approach also fosters collaboration and accelerates research within the climate tech community.
Given the growing urgency to address climate change and extreme weather events, Nvidia’s Earth-2 models come at a critical time, providing scalable tools that can adapt to diverse geographic regions and forecasting needs. This move positions Nvidia at the forefront of combining AI with climate science, reinforcing their dominance in GPU-accelerated computing while contributing valuable capabilities to global weather prediction efforts. The company supports these models with resources and tools, aiming to democratize access to high-quality weather AI technologies.
