NVIDIA, Google DeepMind, Disney unveil Newton

By combining cutting-edge simulation capabilities with NVIDIA’s accelerated computing technology,  to push the boundaries of what robots can achieve, from precision industrial tasks to expressive entertainment characters.

Building a smarter, faster robotics ecosystem

At its core, Newton is built on NVIDIA Warp, a GPU-accelerated computing framework that enables high-performance simulations. By leveraging Warp’s ability to execute parallel processing on NVIDIA GPUs, roboticists can now train models faster and more efficiently. This means real-time simulations that allow robots to refine their movements, optimise interactions, and adapt to complex environments with greater accuracy.

Moreover, Newton is fully open source, fostering a collaborative ecosystem where researchers and developers can experiment, modify, and contribute to the evolution of robotic simulation. With Newton’s extensibility, developers can tailor simulations to unique use cases, from intricate robotic grasping tasks to large-scale autonomous navigation.

MuJoCo-Warp: speeding up simulation

One of Newton’s defining features is its integration with MuJoCo-Warp, a newly introduced open-source robotics simulator from Google DeepMind. By accelerating MuJoCo’s physics engine with Warp, developers can experience up to a 100x speedup in certain robotic manipulation tasks. This means that complex contact-rich interactions – such as a robotic hand manipulating a tool – can now be simulated with unprecedented efficiency.

By combining MuJoCo-Warp with Newton’s flexible simulation environment, roboticists can significantly reduce development time and computational costs, making AI-powered robotics more accessible across industries.

Differentiable physics: a new era of learning

Traditional physics engines simulate the world in a forward manner, where an input results in an output. However, Newton introduces differentiable physics, a technique that enables robots to learn by optimising system parameters through backpropagation. This allows for more effective training by fine-tuning how robots interact with their surroundings, much like how AI models adjust weights in deep learning.

This capability will be crucial in developing autonomous systems that can refine their actions through self-improvement, leading to more intelligent and adaptable robots in fields such as manufacturing, logistics, and healthcare.

Courtesy of Walt Disney Imagineering

Disney Research and the future of entertainment robots

Beyond industrial and research applications, Newton is set to transform entertainment robotics. Disney Research is integrating the engine into its robotic character platform, bringing expressive, interactive robots to life. The Star Wars-inspired BDX droids, introduced at NVIDIA’s GTC keynote, are an early example of how Newton can enhance realism and engagement in robotic characters.

“The BDX droids are just the beginning. We’re committed to bringing more characters to life in ways the world hasn’t seen before, and this collaboration with Disney Research, NVIDIA, and Google DeepMind is a key part of that vision,” said Kyle Laughlin, SVP, Walt Disney Imagineering Research and Development.

“This collaboration will allow us to create a new generation of robotic characters that are more expressive and engaging than ever before – and connect with our guests in ways that only Disney can.”

As Disney continues to push the boundaries of robotic storytelling, Newton’s high-fidelity simulations will play a key role in developing robots that can express emotions, interact dynamically, and captivate audiences in ways previously unimaginable.

A unified robotics pipeline with OpenUSD

Newton is also built on OpenUSD (Universal Scene Description), a powerful framework that allows seamless data aggregation and simulation. Alongside partners such as Google DeepMind, Intrinsic, and NVIDIA, Disney Research is working to define an OpenUSD asset structure for robotics, unifying workflows across different robotic platforms.

This structured approach ensures that robotic simulations can be easily shared and replicated, accelerating progress in the field by providing a common language for roboticists worldwide.

The road ahead for Newton

As Newton continues to evolve, it is expected to become a cornerstone in the advancement of humanoid robotics. The first version is set for release later this year, marking the beginning of a new era where AI-driven, high-speed robotic simulation becomes the norm.