Newton

Newton is a GPU-accelerated physics simulation engine built upon NVIDIA Warp, specifically targeting roboticists and simulation researchers.

Newton extends and generalizes Warp's (deprecated) warp.sim module, and integrates MuJoCo Warp as its primary backend. Newton emphasizes GPU-based computation, OpenUSD support, differentiability, and user-defined extensibility, facilitating rapid iteration and scalable robotics simulation.

Newton is a Linux Foundation project that is community-built and maintained. Code is licensed under Apache-2.0. Documentation is licensed under CC-BY-4.0. Additional and third-party license texts are available in newton/licenses.

Newton was initiated by Disney Research, Google DeepMind, and NVIDIA.

Requirements

• Python 3.10+ (3.10 support is deprecated; removal planned for Newton 1.9; use 3.11+)
• OS: Linux (x86-64, aarch64), Windows (x86-64), or macOS (CPU only)
• GPU: Fresh installs currently select Warp 1.18, which requires an NVIDIA Turing GPU (compute capability 7.5+) and driver R580+ for CUDA acceleration. On an R545–R579 driver or an older supported GPU, install with warp-lang==1.17.0 (see the installation guide). No local CUDA Toolkit is required; macOS runs on CPU.

For detailed system requirements, see the installation guide. For tested configurations and Newton's versioning and deprecation policies, see the compatibility guide.

Quickstart

pip install "newton[examples]"
python -m newton.examples

To install from source with uv, see the installation guide.

Community Projects

Explore community solvers, integrations, and applications through the newton-physics topic on GitHub.

To share your own project, add the newton-physics topic to your public repository's About section. See GitHub's guide to repository topics for instructions.

Projects opt in by adding the topic; inclusion does not imply endorsement or support by the Newton maintainers. External solvers and extensions are maintained separately from Newton. Check each project's documentation for maturity, compatibility, and support.

Examples

Before running the examples below, install Newton with the examples extra:

pip install "newton[examples]"

If you run the examples from a source checkout with uv, use uv run --extra examples -m newton.examples <example_name> instead of the python -m newton.examples <example_name> commands below.

Joint impedance heterogeneous

python -m newton.examples controller_joint_impedance_heterogeneous

Operational space hybrid force and motion

python -m newton.examples controller_operational_space_hybrid_force_motion

MuJoCo Franka VBD cable ADMM

python -m newton.examples mujoco_franka_vbd_cable_admm_solver

Example Options

All examples share a common set of command-line options. The most commonly used ones are:

Argument	Description	Default
--viewer	Viewer to use: gl, usd, rtx, rerun, viser, or null.	gl
--device	Compute device to use, e.g., cpu, cuda:0.	None (default Warp device)
--num-frames	Total number of frames to simulate.	100
--output-path	Path to the output USD file (used by the usd viewer).	output.usd

For the complete, always-current list of options — plus any arguments a specific example adds — run:

python -m newton.examples <example_name> --help

Example Usage

# List available examples
python -m newton.examples --list

# Run with the USD viewer and save to my_output.usd
python -m newton.examples basic_viewer --viewer usd --output-path my_output.usd

# Run on a selected device
python -m newton.examples basic_urdf --device cuda:0

# Combine options
python -m newton.examples basic_viewer --viewer gl --num-frames 500 --device cpu

Research using Newton

See PUBLICATIONS.md for research using Newton, inclusion criteria, and instructions for adding your publication.

Contributing and Development

See CONTRIBUTING.md for ways to contribute and the pull-request process. The governance contribution guidelines cover legal requirements, project roles, and approval authority. Use the development guide for environment setup and development workflows. Code and public API changes must follow the source code guidelines; reviewers may use the suggested review guidelines.

Support and Community Discussion

For questions, please consult the Newton documentation first before creating a discussion in the main repository.

Code of Conduct

By participating in this community, you agree to abide by the Linux Foundation Code of Conduct.

Project Governance, Legal, and Members

Please see the newton-governance repository for more information about project governance.