NVIDIA Earth-2 Weather Analytics: A Step Forward for Climate Tech

SAN JOSE, Calif., March 18, 2025 (GLOBE NEWSWIRE) -- At the GTC event, NVIDIA unveiled the NVIDIA Omniverse Blueprint for Earth-2 weather analytics, aimed at enhancing the accuracy of weather forecasting.

Over the past decade, climate-related weather events have caused a staggering $2 trillion impact on the global economy. The new Omniverse Blueprint provides cutting-edge technologies designed for organizations to enhance their risk management and disaster preparedness capabilities.

The NVIDIA Omniverse Blueprint for Earth-2 weather analytics includes reference workflows, which feature NVIDIA GPU acceleration libraries, a physics-AI framework, development tools, and microservices. This provides a streamlined path for businesses to move from initial prototyping to full-scale production of weather forecasting models.

Included in the blueprint are easy-to-use NVIDIA NIM™ microservices designed for NVIDIA Earth-2. Notable services like CorrDiff, used for downscaling data, and FourCastNet, aimed at predicting global atmospheric dynamics, are already benefiting various weather tech companies, researchers, and government agencies in crafting solutions to counteract extreme weather risks.

"As extreme weather events and natural disasters become more frequent, they pose significant threats to lives and properties worldwide," stated Jensen Huang, NVIDIA's founder and CEO. "The NVIDIA Omniverse Blueprint for Earth-2 is set to empower industries globally in their efforts to prepare for and respond to climate change and disaster challenges."

Collaborative Ecosystem

A range of leading climate tech companies, including G42, JBA Risk Management, and Spire, are incorporating the Omniverse Blueprint to develop innovative AI-enhanced weather solutions.

In collaboration with proprietary data across the $20 billion climate tech sector, the NVIDIA Earth-2 platform allows developers to produce timely alerts and forecasts within seconds. This is a significant improvement compared to traditional CPU-based modeling, which could take minutes or even hours.

G42 has begun integrating components from the Omniverse Blueprint with its AI-driven forecasting models, providing advanced solutions to the UAE's National Center of Meteorology for better weather forecasting and disaster management.

Andrew Jackson, CEO of Inception (a G42 company), remarked, "We are leveraging AI to revolutionize forecasting, empowering governments and enterprises to better cope with the impacts of extreme weather in an ever-evolving environment. Our collaboration with NVIDIA’s CorrDiff model allows us to transform large-scale weather data into highly localized forecasts, ultimately enabling rapid decision-making for community safety and infrastructure resilience."

Another player, Spire Global, is using AI elements from the blueprint to develop new products that combine its satellite data with weather models, producing medium-range and sub-seasonal forecasts extending up to 45 days. Spire's cutting-edge models, powered by NVIDIA GPUs, are operating at speeds over 1,000 times faster than traditional physics-based models, allowing for extensive ensemble forecasts that outline a range of potential weather scenarios.

Organizations such as Taiwan's Central Weather Administration and The Weather Company, in addition to 3D mapping site Ecopia, spatial analysis company ESRI, and green energy firm GCL Power, are also exploring the capabilities of Earth-2.

Innovative AI in Climate Tech

The Earth-2 platform provides a suite of tools, microservices, and advanced AI weather models for in-depth visualization and simulation of global weather patterns.

The CorrDiff service, featured within the Omniverse Blueprint, functions as an NVIDIA NIM microservice and is exceptionally efficient — 500 times faster and 10,000 times more energy-efficient than traditional CPU methods, producing high-resolution numerical weather predictions.

Additionally, the Omniverse Blueprint facilitates independent software vendors in creating and deploying AI-enhanced solutions, allowing observational data to refine the accuracy and efficiency of their offerings.

Esri, renowned for its geospatial technology, is working with NVIDIA to connect its ArcGIS platform with Earth-2, while OroraTech is examining ways to integrate its data platform into the Omniverse Blueprint for Earth-2.

Tomorrow.io is collaborating with NVIDIA to utilize its real-time satellite data to help build a digital twin of Earth, invaluable for training and enhancing the next generation of AI models.

A standout feature of the new blueprint is the NVIDIA Omniverse™, a platform used for developing three-dimensional workflows and applications based on OpenUSD. The Earth-2 blueprint showcases how developers can effectively leverage Omniverse SDKs and microservices to architect RTX™-powered visual pipelines for rendering complex geospatial and weather data.

High-Performance Weather Forecasting

The Omniverse Blueprint taps into the NVIDIA DGX™ Cloud platform to showcase comprehensive acceleration for AI-powered weather forecasting. Utilizing NVIDIA supercomputers like DGX GB200, HGX™ B200, and OVX™, the blueprint paves the way for rapid and extensive global climate simulations.

For further details, watch the NVIDIA GTC keynote and register for additional sessions with NVIDIA and industry experts during the event, which runs through March 21.

About NVIDIA

NVIDIA (NVDA) is recognized as a leader in accelerated computing technologies.

climate, technology, AI, forecasting, NVIDIA