Newton Physics
Newton Physics Engine is an open-source, extensible physics engine built on NVIDIA Warp and OpenUSD, developed by NVIDIA, Google DeepMind, and Disney Research, and managed by the Linux Foundation, to advance robot learning and development.
How Newton Physics Works
Physics plays a crucial role in robotic simulation, providing the foundation for accurate virtual representations of robot behavior and interactions within realistic environments.
Accurately simulating robotic behavior relies on following the fundamental laws of physics, including the conservation of mass and momentum, rigid and soft body dynamics, contact and friction, and actuator and sensor modeling. These principles are used to predict how physical multi-body systems, including robots, will behave in various scenarios and environments.
Built on NVIDIA Warp, a developer framework for building and accelerating simulation and spatial computing, Newton enables robots to acquire and refine physical intelligence in a safe, virtual environment and is compatible with robot learning frameworks such as MuJoCo Playground and NVIDIA Isaac™ Lab.
Built on OpenUSD
OpenUSD’s flexible data model and composition engine aggregates data needed for describing robots and their surrounding environments.
NVIDIA Warp
Newton is powered by NVIDIA Warp, an open-source developer framework for building and accelerating data generation and spatial computing in Python.
NVIDIA Isaac Lab
Newton is compatible with NVIDIA Isaac Lab, an open-source, unified framework for robot learning designed to help train robot policies, delivering high-fidelity physics simulation.
MuJoCo Warp
MuJoCo Warp is a GPU-optimized version of the MuJoCo physics simulator, designed for NVIDIA hardware and a key solver in Newton.
Key Features and Benefits
GPU Accelerated
Newton uses NVIDIA Warp to run high-performance simulations on GPUs, giving roboticists CUDA-level speed without low-level coding—cutting simulations from days to minutes.
Extensible
Highly extensible, enabling rich multiphysics simulation via pluggable custom solvers
Differentiable
Newton’s differentiable physics accelerates training, design optimization, and system identification—enabling efficient learning of complex, optimal robot motions
Open Source
Open source by design, it enables rapid adoption, collaboration, and continuous improvement across the robotics community
Get Started With Newton Physics
Quadruped Robot Locomotion and Multiphysics Simulation Using Newton in NVIDIA Isaac Lab
Read the TutorialMore Resources
Get started with Newton Physics today.