Mistral and NVIDIA Partner on Engineering AI

Mistral and NVIDIA are collaborating on AI-native engineering workflows that combine Mistral Large 4, data generation and physics modeling pipelines with NVIDIA accelerated computing. Using NVIDIA CUDA-X libraries and NVIDIA Warp, the work aims to speed up engineering simulations and connect simulation, AI model development and design in closed-loop workflows. The companies also plan to integrate Mistral models with NVIDIA Omniverse libraries for GPU-accelerated 3D scene generation and SimReady asset creation.

The partnership will develop public benchmarks and open datasets to evaluate physics models, giving researchers, developers and industrial teams a shared basis for comparing accuracy and performance and validating automated CAD-to-simulation workflows. Mistral and NVIDIA will also create evaluations for agentic AI in computer-aided engineering (CAE) and electronic design automation (EDA). Those benchmarks will measure whether agents can use engineering tools, complete multistep workflows and produce technically correct results, as well as execution time and computational cost. The companies say the work will establish consistent, reproducible methods for assessing engineering AI.

The collaboration extends an existing Mistral-NVIDIA partnership in open models, domain-specific AI and production infrastructure. The companies previously co-developed Mistral NeMo and are working through the NVIDIA Nemotron Coalition on a frontier base model, with coalition members contributing data, evaluations and domain expertise. Mistral recently introduced a public preview of Mistral Large 4, with approximately 1 trillion total parameters and 49 billion active parameters. The multimodal model was trained on 3,800 NVIDIA Grace Blackwell GPUs in Mistral-operated data centers in Europe; its weights are planned for release at the end of October.

Mistral is bringing additional NVIDIA-powered AI infrastructure online as part of a compute roadmap targeting 1 GW of capacity by 2030. The companies say this expansion connects European infrastructure and research with open models, engineering capabilities and customer applications, supporting enterprises that build AI around their own expertise and serve markets worldwide.

Mistral and NVIDIA are collaborating on AI-native engineering workflows that combine Mistral Large 4, data generation and physics modeling pipelines with NVIDIA accelerated computing. Using NVIDIA CUDA-X libraries and NVIDIA Warp, the work aims to speed up engineering simulations and connect simulation, AI model development and design in closed-loop workflows. The companies also plan to integrate Mistral models with NVIDIA Omniverse libraries for GPU-accelerated 3D scene generation and SimReady asset creation.

The partnership will develop public benchmarks and open datasets to evaluate physics models, giving researchers, developers and industrial teams a shared basis for comparing accuracy and performance and validating automated CAD-to-simulation workflows. Mistral and NVIDIA will also create evaluations for agentic AI in computer-aided engineering (CAE) and electronic design automation (EDA). Those benchmarks will measure whether agents can use engineering tools, complete multistep workflows and produce technically correct results, as well as execution time and computational cost. The companies say the work will establish consistent, reproducible methods for assessing engineering AI.

The collaboration extends an existing Mistral-NVIDIA partnership in open models, domain-specific AI and production infrastructure. The companies previously co-developed Mistral NeMo and are working through the NVIDIA Nemotron Coalition on a frontier base model, with coalition members contributing data, evaluations and domain expertise. Mistral recently introduced a public preview of Mistral Large 4, with approximately 1 trillion total parameters and 49 billion active parameters. The multimodal model was trained on 3,800 NVIDIA Grace Blackwell GPUs in Mistral-operated data centers in Europe; its weights are planned for release at the end of October.

Mistral is bringing additional NVIDIA-powered AI infrastructure online as part of a compute roadmap targeting 1 GW of capacity by 2030. The companies say this expansion connects European infrastructure and research with open models, engineering capabilities and customer applications, supporting enterprises that build AI around their own expertise and serve markets worldwide.