Malaysia
Singapore
Philippines
Stay connected with BizTech Community—follow us on Instagram and Facebook for the latest news and reviews delivered straight to you.
-
Nvidia, Google, and Disney launch GR00T N1, the first open humanoid foundation model, blending vision, language, and motion for generalist robotics.
-
GR00T N1 mimics human cognition, using dual systems for reflexive movement and reasoning; it generalizes across tasks like grasping and tidying.
-
Newton physics engine, co-developed with Disney and DeepMind, powers precise robotic simulations—soon to be used in Disney’s expressive droids.
-
Synthetic data via Isaac GR00T Blueprint generated 780K training motions in 11 hours, improving model performance by 40%.
-
Early adopters include Agility and Boston Dynamics, targeting global labor gaps in manufacturing, logistics, and more.
-
Resources now live on Hugging Face and GitHub, with a live demo of synthetic motion generation available for developers.
-
Future impact spans beyond entertainment, with implications for healthcare, defense, and industry-wide robotics innovation.
When you imagine Nvidia, your mind might leap to sleek GPUs powering epic gaming marathons or AI crunching data at lightning speed—not robots pirouetting across a stage.
Yet, at GTC 2025, Nvidia flipped the script, unveiling a futuristic trio with Google DeepMind and Disney that’s set to redefine more than just tech specs.
This isn’t about faster chips or sharper graphics; it’s a bold step toward a world where robots don’t just compute—they captivate.
Teased amidst the buzz of San Jose’s SAP Center, this partnership promises to weave innovation into industries from factory floors to fairy-tale parks, igniting a new chapter in robotics that’s as unexpected as it is thrilling.
This trio is set to turbocharge robotics innovation, from factory floors to Disney’s theme parks, with tools like the Newton physics engine and synthetic data frameworks. The message was clear: the age of adaptable, intelligent robots has arrived.
Table of Contents
GR00T N1: The Humanoid Game-Changer
Isaac GR00T N1 isn’t just another robotics platform—it’s a leap toward generalist robots that think and act like humans.
Built with a dual-system architecture inspired by human cognition, GR00T N1 combines a fast, reflexive “System 1” for precise movements and a deliberate “System 2” powered by a vision-language model for reasoning and planning.
Whether grasping objects, transferring items between arms, or tidying a room, GR00T N1 can easily generalize across tasks.
Huang showcased this with 1X’s NEO Gamma robot, which autonomously tidied a space using a GR00T N1-trained policy—proof of its real-world potential.
Available now to developers via Hugging Face and GitHub, GR00T N1 is the first in a family of models Nvidia will release, fully customizable with real or synthetic data.
Early adopters like Agility Robotics and Boston Dynamics are already onboard, signaling its promise to tackle labor shortages—estimated at 50 million workers globally—across industries like manufacturing and logistics.
“The next frontier in AI is here,” Huang declared, envisioning a world where robots don’t just work but adapt.
Google, Disney, and Nvidia, a Robotics Dream Team
Nvidia CEO Jenson Huang delivered his annual keynote address at GTC 2025
The collaboration with Google DeepMind and Disney Research elevates this vision. Together, they’re developing Newton, an open-source physics engine built on Nvidia’s Warp framework, designed to simulate robotic interactions with unparalleled accuracy.
Compatible with DeepMind’s MuJoCo and Nvidia’s Isaac Lab, Newton will let robots master complex tasks—think manipulating cloth or navigating uneven terrain.
Google’s contribution includes MuJoCo-Warp, which accelerates machine learning by over 70x, while Disney adds its magic touch to the mix.
Disney Research plans to deploy Newton in its next-generation entertainment robots, starting with the BDX droids—Star Wars-inspired bots that waddled onto the GTC stage alongside Huang.
Dubbed “Blue,” one droid showcased real-time simulation powered by Nvidia tech, hinting at a future where robotic characters roam Disney parks.
“This is just the beginning,” said Kyle Laughlin, Senior VP at Walt Disney Imagineering. “We’re crafting robots that connect with guests in ways only Disney can—expressive, engaging, and alive.”
Synthetic Data: The Secret Sauce
Training robots demands vast datasets, but human demonstrations are limited.
Enter the Isaac GR00T Blueprint, a synthetic data generator built on Nvidia’s Omniverse.
In just 11 hours, Nvidia created 780,000 motion trajectories—equivalent to nine months of human effort—boosting GR00T N1’s performance by 40% when paired with real data.
This tool, now available as a demo, empowers developers to scale training exponentially, making robotics more accessible and efficient.
Beyond Entertainment: A Broader Impact
While Disney’s droids steal the spotlight, GR00T N1 and Newton have far-reaching potential. From healthcare to defense, these technologies could redefine how robots assist humanity.
Nvidia’s DGX Spark, a personal AI supercomputer also unveiled at GTC, further simplifies customization, letting developers tailor GR00T N1 to new tasks without heavy coding.
Newton, expected later this year, could even find its way into industrial or military applications, raising questions about robotics’ future beyond theme parks.
Availability and Future Development
Based on their statement, NVIDIA has made the GR00T N1 training datasets and task evaluation scenarios accessible for download on Hugging Face and GitHub.
Furthermore, developers can explore the NVIDIA Isaac GR00T Blueprint for generating synthetic motion data through an interactive demo at build.nvidia.com or grab it from GitHub.
As robotics advances at breakneck speed, NVIDIA’s robust ecosystem marks a pivotal step, broadening access to cutting-edge humanoid tech and unlocking fresh opportunities across fields like manufacturing, healthcare, entertainment, and more.