Disney Characters Coming to Life at NVIDIA GTC 2026

2026-03-05 · Source: NVIDIA · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems · Depth: Intermediate, quick

Summary

Disney is employing physical AI, specifically the open-source, GPU-accelerated Newton framework and its proprietary Kamino solver, to develop artist-centric reinforcement learning for robotic characters. This initiative focuses on training motion diffusion models and policies to achieve expressive motion, pushing the boundaries of hardware capabilities. Additionally, the project integrates onboard perception for autonomous navigation, enabling unique and engaging guest interactions through authentic character movements. Moritz Baecher will present these advancements at NVIDIA GTC on Tuesday, March 17, at 9:00 AM PT, detailing how physical AI accelerates the transition of animated characters into real-world robotic forms within months.

Key takeaway

For robotics engineers developing interactive character systems, you should investigate integrating artist-centric reinforcement learning frameworks like Newton and Kamino. Focusing on motion diffusion models and onboard perception can significantly enhance expressive motion and autonomous interaction capabilities, accelerating deployment timelines for complex robotic characters.

Key insights

Disney uses physical AI, Newton, and Kamino to create expressive, autonomous robotic characters.

Principles

• Artist-centric reinforcement learning is key.
• Push hardware limits for expressive motion.

Method

Combines open-source Newton framework with Disney's Kamino solver for reinforcement learning, trains motion diffusion models, and integrates onboard perception for autonomy.

In practice

• Utilize GPU-accelerated frameworks.
• Develop motion diffusion models.
• Implement onboard perception for navigation.

Topics

• Physical AI
• Reinforcement Learning
• Motion Diffusion Models
• Robotic Characters
• Autonomous Navigation

Best for: Machine Learning Engineer, Computer Vision Engineer, AI Scientist, AI Engineer, Robotics Engineer, Creative Technologist

Related on AIssential

• [P] I open-sourced a synth framework for creating physics-simulated humanoids in Unity with MuJoCo -- train them with on-device RL and interact in VR — This system enables physics-based humanoid creation and on-device reinforcement learning for mixed reality.
• What Amazon’s Astro Taught Me About Giving Robots a Soul — Character design for embodied AI must be intentional, integrated early, and adapt to context to create truly "aware" products.
• TAI #190: Genie 3 World Model Goes Public — Genie 3 enables real-time interactive world generation, advancing both game development prototyping and embodied AI research.
• Game, Set, Bot: Sony AI’s ‘Ace’ Serves Up a Defeat to Table Tennis Pros — Gaming environments serve as safe, scalable testbeds for developing and benchmarking advanced physical AI agents.
• DiscoForcing: A Unified Framework for Real-Time Audio-Driven Character Control with Diffusion Forcing — DiscoForcing enables real-time, stable audio-driven character animation by combining causal encoding with diffusion forcing and history-guided sampling.
• 🤖Physically-Plausible Human🤖 👉PhysMoDPO is a novel direct preference optimization... — PhysMoDPO uses direct preference optimization to generate physically plausible humanoid motions.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by NVIDIA.