From March 17th to 21st, the 2025 NVIDIA GTC (NVIDIA GPU Technology Conference) was held in San Jose, California, USA. At this GTC conference, GCL Group, in collaboration with Peking University and NVIDIA, achieved in-depth industry-academia-research collaboration and jointly launched a large-scale PV power prediction model based on the NVIDIA Earth-2 platform. This marks a breakthrough in the intelligent transformation of the energy industry.
The model innovatively constructs a triangular support system of “industry data + research capabilities + supercomputing power”: relying on GCL Group’s years of practical experience and massive data accumulation in the clean energy field, combined with Peking University’s cutting-edge research achievements in artificial intelligence, big data analysis, and other fields, and utilizing weather and climate prediction data from the NVIDIA Earth-2 platform, it can accurately predict new energy generation power.
NVIDIA Earth-2, as a climate digital twin cloud platform, combines the powerful functions of AI, GPU acceleration, physical simulation, and computer graphics, enabling weather and climate predictions and data generation worldwide with unprecedented accuracy and speed. By leveraging the predicted data, the accuracy of PV power generation forecasting will be improved by 3 to 5 percentage points compared to relying solely on time series models.
