Mistral has announced the acquisition of Emmi AI, emphasizing its dedication to advancing AI research and enterprise solutions for industrial engineering. The integration of Emmi’s expertise into Mistral’s operations aims to enable engineers to develop next-generation products more efficiently and maintain consistent performance improvements for their clients. Mistral is focusing on foundational Physics AI for industries that shape the physical world, including aerospace, automotive, semiconductors, and energy. The company highlighted several published breakthroughs that support this initiative. On December 1, 2025, Mistral released a dataset of computational fluid dynamics (CFD) simulations for 3D wings in the transonic regime, comprising volumetric and surface-level fields for approximately 30,000 samples with unique geometries and inflow conditions. In November 2025, a paper titled 'Fluid Intelligence' explored the integration of AI foundation models into CFD, driven by GPU advancements. Another paper from October 17, 2025, introduced two new datasets for AB-UPT, combining high-quality data with state-of-the-art neural surrogates. On October 8, 2025, a study on GyroSwin presented 5D surrogates for gyrokinetic plasma turbulence simulations, addressing challenges in nuclear fusion. The AB-UPT model, introduced in February 2025, handles raw geometry without remeshing on a single GPU, managing 9 million surface and 140 million volume cells. In November 2024, NeuralDEM was introduced as the first end-to-end deep learning surrogate for large-scale multi-physics processes, enabling real-time simulations of industrial processes like fluidised bed reactors. The UPT framework, launched in February 2024, supports efficient scaling of neural operators across diverse spatio-temporal problems, accommodating both grid and particle simulations. *Source: [mistral](https://mistral.ai/news/physics-ai-research)*