Mistral and Nvidia want to speed up industrial design with physics AI
decoder ·
Mistral and Nvidia have announced a strategic partnership aimed at accelerating industrial design and engineering simulations through the integration of physics-based AI. The collaboration combines Mistral Large 4, a massive language model with approximately one trillion parameters, with specialized physics AI tools. This technology will be integrated into Nvidia’s CUDA-X and Warp software libraries alongside Omniverse to facilitate 3D simulations. The primary objective is to create a seamless workflow that connects product design, simulation processes, and AI training for sectors such as aerospace, automotive, and semiconductors. Additionally, the companies plan to release public benchmarks and open datasets to help test and compare physics models and engineering agents. This initiative builds on their previous joint development of Mistral NeMo and their involvement in the Nemotron Coalition. Mistral Large 4 is scheduled to launch as an open-weight model in late October after being trained on 3,800 Grace Blackwell GPUs in Europe.
decoder