G42 and NVIDIA have launched a hybrid AI-based weather forecasting model capable of generating 200-metre resolution simulations tailored for urban environments, starting with Abu Dhabi. The development is part of their ongoing collaboration under the Earth-2 Climate Tech Lab initiative, based in the UAE.
Developed by Inception, a subsidiary of G42, the custom generative model—CorrDiff—uses NVIDIA’s accelerated computing technology to produce high-resolution weather forecasts and predict extreme climate events. CorrDiff adapts to local weather conditions, including fog, which is a key challenge for the UAE.
G42 said it has already conducted fog simulations across the UAE, improving forecasting accuracy and operational readiness in sectors such as aviation, logistics, and emergency response. Core42, G42’s infrastructure arm, will host the system, providing computing and inference services powered by NVIDIA.
The hybrid model integrates traditional physics-based forecasting with AI-powered data modelling, improving both the speed and cost-efficiency of weather prediction systems. According to G42, this results in better early warnings for weather events and more accurate decision-making tools for governments and businesses.
Dr Abdulla Al Mandous, Director-General of the UAE’s National Center of Meteorology, said the model enables “real-time, localised weather predictions” that improve adaptation strategies and forecasting accuracy.
Dr Andrew Jackson, CEO of Inception, said CorrDiff’s utility extends beyond the UAE and can be deployed in other regions facing climate-related challenges. He cited use cases including aviation, renewable energy planning, and urban mobility systems.
NVIDIA’s Dion Harris added that the Earth-2 platform supports high-resolution AI weather models that require “immense computational power,” and that such systems will be central to improving climate resilience globally.
G42 is evaluating opportunities to expand the AI model across climate-vulnerable regions in Africa, South Asia, and Southeast Asia. The move aligns with broader trends in AI deployment for climate adaptation, including the use of generative models for real-time, granular simulations of weather impacts.
The partnership follows a global push by governments and the private sector to use AI for improved disaster preparedness, as climate volatility increasingly disrupts infrastructure, energy systems and economic activity.
