엔비디아 Earth-2, 기후 기술 날씨 예측 정확도 향상
무슨 발표인가
- —지난 10년 기후 영향 2조 달러
- —NVIDIA Omniverse Blueprint for Earth-2 발표
- —CorrDiff, FourCastNet 마이크로서비스 포함
기후 기술사, 연구기관, 정부 기관 사용 중
원문 (영어)
GTC— NVIDIA today announced the NVIDIA Omniverse Blueprint for Earth-2 weather analytics to accelerate the development of more accurate weather forecasting solutions. Climate-related weather events have had a $2 trillion impact on the global economy over the last decade.
The new Omniverse Blueprint equips users with the latest technologies to help global organizations improve risk management and disaster preparedness. The NVIDIA Omniverse Blueprint for Earth-2 weather analytics offers reference workflows — including NVIDIA GPU acceleration libraries, a physics-AI framework, development tools and microservices — to help enterprises go from prototyping to production with weather forecast models.
Easy-to-deploy NVIDIA NIM microservices for NVIDIA Earth-2 are also part of the blueprint, including CorrDiff for downscaling and FourCastNet for predicting global atmospheric dynamics of various weather and climate variables. These are already being used by weather technology companies, researchers and government agencies to derive insights and mitigate risk from extreme weather events.
“We’re seeing more extreme weather events and natural disasters than ever, threatening lives and property,” said Jensen Huang, founder and CEO of NVIDIA. “The NVIDIA Omniverse Blueprint for Earth-2 will help industries around the world prepare for — and mitigate — climate change and weather-related disasters.”
Ecosystem Support Industry-leading climate tech companies including AI company G42, JBA Risk Management, Spire and others are using the blueprint to develop unique AI-augmented solutions.
원문: NVIDIA News — "Climate Tech Companies Adopt NVIDIA Earth-2 for High-Resolution, Energy-Efficient, More Accurate Weather Predictions and Disaster Preparedness" (2025-07-03) 공식 원문: https://nvidianews.nvidia.com/news/nvidia-earth-2-climate-tech-weather-prediction-disaster-preparedness