Mistral and Nvidia expand partnership into engineering simulation and AI testing
Mistral and Nvidia plan faster engineering simulations, shared physics datasets and tests of AI agents used in engineering and chip design.

Mistral announced an engineering collaboration with Nvidia on October 21. The companies aim to speed up simulations and connect AI development with engineering design.
The work combines Mistral Large 4 and the company's physics modeling tools with Nvidia's CUDA-X and Warp software. The partners also plan to connect Mistral models with Omniverse to generate 3D scenes and simulation assets.
The collaboration extends an existing relationship that includes the jointly developed Mistral NeMo model. Its engineering focus includes aerospace, automotive manufacturing and semiconductors, where simulations help engineers evaluate designs before building physical products.
The companies will develop public tests and openly available datasets for physics models. Separate evaluations will measure whether AI agents can operate engineering tools and finish tasks correctly, alongside their speed and computing costs. The announcement did not specify release dates for those resources.
The day's biggest AI news in one short email, every morning. Plus an instant heads-up when a reset or a big release lands.
More on NVIDIA
- Oct 21Nvidia expands local AI lineup with 64GB DGX Spark and Windows DGX StationOfficial
- Oct 21Google Cloud schedules Nvidia Vera Rubin A5X availability for NovemberOfficial
- Oct 21Nvidia details Kumo Forecast's historical retrieval and explanation toolsOfficial
- Oct 21Nvidia expands Agent Toolkit to connect AI agents with business dataOfficial
- Oct 21Nvidia expands open models for European AI developmentOfficial
- Oct 21Jensen Huang urges Europe to build its own AI capacity at GTC BerlinOfficial