NVIDIA and its CEO Jensen Huang had their presence at Berlin Summit through a keynote that focuses on climate research augmented with AI power and digital twin technologies.
At the current state, the very best hardware, the NVIDIA GH200 Grace Hopper Superchip, is capable of delivering up to 10x higher performance in applications running TB-levels of data which is also the scale of weather prediction software runs.
Speaking of software, Sir Leather Jacket also mentioned NVIDIA Modulus – an open-source framework dedicated to physics-based machine learning models for building, training, and fine-tuning, alongside the global, data-driven weather forecasting model FourCastNet.
Its ability to learn and predict the path of Hurricane Harvey by modeling the Coriolis force, the effect of the Earth’s rotation, on the storm, shows how powerful is accelerated computing in achieving all of these.
Cloud computing also empowers interactive visualization of global-scale climate data that is high-resolution, where in this case Berlin, for monitoring and predicting climate shifts and weather changes that can be any city such as Tokyo, Buenos Aires, and more.
Jensen also outlined NVIDIA’s continuous effort in bringing more powerful computing solutions, both hardware and software, that were deemed impossible just a few years prior.