NextFin News - On January 26, 2026, Nvidia announced the launch of its Earth-2 family of open-source weather and climate AI models at the American Meteorological Society’s annual meeting in Houston. The suite, which includes Earth-2 Medium Range, Nowcasting, and Global Data Assimilation, is designed to provide faster and more cost-effective forecasting compared to traditional numerical weather prediction (NWP) methods. According to Nvidia, these models allow scientists, startups, and government agencies to run complex simulations on standard GPUs rather than relying exclusively on multi-million dollar supercomputer clusters.
The Earth-2 Medium Range model, built on the new Atlas transformer architecture, delivers 15-day global forecasts across more than 70 variables, including temperature and wind speed. Simultaneously, the Earth-2 Nowcasting tool utilizes generative AI to provide kilometer-resolution local storm predictions up to six hours in advance. Early adopters, including the Israel Meteorological Service and the U.S. National Weather Service, are already testing the technology. Director Amir Givati of the Israel Meteorological Service reported a 90% reduction in compute time for high-resolution forecasts, highlighting the immediate efficiency gains offered by the platform.
The release of Earth-2 represents a direct challenge to the current AI meteorological benchmark set by Google DeepMind’s GenCast. Nvidia claims its new models outperform GenCast across dozens of weather variables, though scientific observers note that these claims currently lack peer-reviewed validation. The competition between these tech giants underscores a broader industry shift: the transition from physics-based models that solve complex fluid dynamics equations to data-driven AI models that identify patterns in historical atmospheric data. Mike Pritchard, Nvidia’s director of climate simulation research, noted that once trained, AI models can be up to 1,000 times faster than traditional systems, enabling the execution of massive "ensemble" forecasts to detect rare, high-impact weather events.
From a financial and operational perspective, the democratization of these tools is poised to reshape the energy and insurance sectors. Grid operators like Southwest Power Pool are evaluating Earth-2 for intraday wind forecasting to better balance power loads, while insurance firms such as AXA are using similar AI architectures to simulate thousands of hypothetical hurricane scenarios for risk modeling. By open-sourcing these models on platforms like GitHub and Hugging Face, Nvidia is lowering the barrier to entry for smaller nations and private enterprises that previously could not afford the infrastructure required for sovereign weather intelligence.
However, the rapid adoption of AI in meteorology is met with measured caution from the scientific community. A 2025 study from the University of Geneva indicated that while AI models excel at standard patterns, they often struggle with record-breaking extreme events where historical data is sparse. As U.S. President Trump’s administration continues to emphasize infrastructure and national security, the reliability of these tools becomes a matter of public policy. Weather forecasting is increasingly viewed as a national security asset; the ability to predict storm trajectories with kilometer-level precision can determine the success of emergency evacuations and the stability of the national power grid.
Looking forward, the success of Earth-2 will likely hinge on its integration into the broader AI ecosystem. Nvidia’s simultaneous $2 billion investment in CoreWeave, announced today, suggests a strategic push to ensure that the "AI factories" of the future have the specialized cloud capacity to run these heavy climate workloads. As the industry moves toward the February 25 earnings report, investors will be watching closely to see if Nvidia’s software-led strategy in climate tech can maintain its competitive moat against in-house chip developments from rivals like Microsoft and Google. The convergence of generative AI and climate science suggests that the next decade of meteorological advancement will be defined not by the size of the supercomputer, but by the sophistication of the neural network.
Explore more exclusive insights at nextfin.ai.
