NIM for DoMINO Automotive Aero
Type: Microservice Tags: NVIDIA, NIM, DoMINO, automotive aerodynamics, PhysicsNeMo, CFD, surrogate model, simulation, vehicle design Related: NVIDIA-NIM, PhysicsNeMo, NVIDIA-Modulus, NIM-for-Earth-2-CorrDiff, NIM-for-Earth-2-FourCastNet, NVIDIA-Drive-Platform, NVIDIA-DRIVE-AGX-Thor, NVIDIA-AI-Enterprise, TensorRT, Triton-Inference-Server, NVIDIA-CUDA, NGC Sources: https://docs.nvidia.com/nim/physicsnemo/domino-automotive-aero/latest/overview.html, https://docs.nvidia.com/nim/physicsnemo/domino-automotive-aero/latest/index.html, https://docs.nvidia.com/nim/physicsnemo/domino-automotive-aero/latest/prerequisites.html Last Updated: 2026-04-29
Summary
NIM for DoMINO Automotive Aero is NVIDIA’s PhysicsNeMo NIM for automotive external aerodynamics prediction. Current NVIDIA docs describe DoMINO as a Decomposable Multi-Scale-Iterative Neural Operator surrogate model that predicts aerodynamic solution fields from point-cloud geometry, reducing simulation time for CFD and vehicle design workflows.
Detail
Purpose
Automotive aerodynamic analysis is expensive when every design iteration requires high-fidelity CFD. DoMINO Automotive Aero NIM provides a fast API-backed surrogate model for engineers evaluating external vehicle aerodynamics during design optimization.
Current scope
- Learns local geometry encodings from point-cloud representations.
- Predicts PDE solution fields on discrete points using dynamically constructed local computational stencils.
- Captures short- and long-range geometric features for vehicle surface and surrounding flow-domain predictions.
- Provides surface and volume field evaluation, velocity inlet configuration, stencil-size controls, confidence intervals, and custom volume point-cloud workflows in current docs.
- Supports multi-GPU and flexible batched inference in current advanced docs.
- Current prerequisites note CUDA compute capability 7.5+ hardware and NVIDIA driver/container runtime requirements.
NVIDIA context
DoMINO Automotive Aero is the simulation/physics NIM counterpart to Earth-2 weather NIMs. It belongs with PhysicsNeMo and NVIDIA-Modulus for physics-ML training, and with NVIDIA-Drive-Platform for vehicle design and automotive engineering context.
Connections
- PhysicsNeMo - framework context for DoMINO and other physics-ML models.
- NVIDIA-Modulus - broader NVIDIA physics-ML lineage.
- NIM-for-Earth-2-CorrDiff and NIM-for-Earth-2-FourCastNet - other physics/science NIMs in the wiki graph.
- NVIDIA-Drive-Platform and NVIDIA-DRIVE-AGX-Thor - automotive AI and vehicle development context.
- NVIDIA-AI-Enterprise - production support and deployment context.
- TensorRT and Triton-Inference-Server - optimized inference and serving context for NIM deployments.
- NVIDIA-CUDA and NGC - GPU software and container distribution context.
Source Excerpts
- NVIDIA docs describe DoMINO as a robust, accurate, scalable surrogate model for automotive aerodynamics.
- The current docs state that DoMINO learns from point-cloud geometry and predicts solution fields on vehicle surfaces and in the surrounding flow domain.