Physics AI research that’s shaping the industry.

Mistral AI’s acquisition of Emmi AI underscores its focus on Physics AI, aiming to enhance industrial engineering through AI-driven solutions. The integration of Emmi’s expertise supports engineers in accelerating product development and improving operational performance at scale. The initiative targets sectors such as aerospace, automotive, semiconductors, and energy, where AI can drive significant advancements.

A new dataset of CFD simulations for 3D wings in the transonic regime addresses gaps in existing aerospace datasets, which primarily cover 2D airfoils. The dataset includes volumetric and surface-level fields for approximately 30,000 samples with unique geometry and inflow conditions, providing a more comprehensive foundation for aerospace research.

The paper bridges the gap between machine learning and Computational Fluid Dynamics (CFD) by breaking down industrial-scale CFD simulations into core components. This work highlights the transformative impact of GPUs and AI on CFD, enabling more efficient and scalable simulations for industrial applications.

Research into nuclear fusion addresses plasma turbulence, a critical challenge for viable fusion power. The Anchored-Branched Universal Physics Transformer (AB-UPT) introduces new datasets and neural surrogates, supporting real-time simulation of industrial processes like fluidised bed reactors and large-scale multi-physics scenarios.