Nvidia Unveils Groot N1: A Groundbreaking Foundation Model for Humanoid Robotics

Nvidia announces Groot N1, a foundation model for humanoid robotics

Nvidia is releasing what it's calling an AI foundation model for humanoid robotics. Announced at GTC 2025 in San Jose, the model, dubbed Groot N1, is a "generalist" model, trained on both synthetic and real data. Nvidia claims that Groot N1 features a "dual system architecture" for "thinking fast and slow," inspired by human cognitive processes. Groot N1 is an evolution of Nvidia's Project Groot, which the company launched at its GTC conference last year. Project Groot was geared toward on industrial use cases, but Groot N1 broadens the focus to humanoid robots in a range of different form factors. Groot N1's slow thinking system lets a robot perceive and reason about its environment and instructions, and then plan the right actions to take, according to Nvidia. As for the fast thinking system, it translates the aforementioned plan into robotic actions, including actions that involve manipulating objects over multiple steps. Groot N1 is available in open source, Nvidia said. Alongside the model, Nvidia is releasing simulation frameworks and blueprints for generating synthetic training data.

Nvidia debuts Groot N1, a foundation model for humanoid robotics | TechCrunch

Nvidia is releasing what it's calling an AI foundation model for humanoid robotics. Announced at GTC 2025 in San Jose, the model, dubbed Groot N1, is a "generalist" model, trained on both synthetic and real data. Nvidia claims that Groot N1 features a "dual system architecture" for "thinking fast and slow," inspired by human cognitive processes. Groot N1 is an evolution of Nvidia's Project Groot, which the company launched at its GTC conference last year. Project Groot was geared toward on industrial use cases, but Groot N1 broadens the focus to humanoid robots in a range of different form factors. Groot N1's slow thinking system lets a robot perceive and reason about its environment and instructions, and then plan the right actions to take, according to Nvidia. As for the fast thinking system, it translates the aforementioned plan into robotic actions, including actions that involve manipulating objects over multiple steps. Groot N1 is available in open source. Alongside the model, Nvidia is releasing simulation frameworks and blueprints for generating synthetic training data. "The age of generalist robotics is here," Nvidia CEO Jensen Huang said in a statement. Humanoid robots have attracted a lot of publicity in recent years. Companies like X1 and Figure are attempting to create general-purpose robots that move more or less like humans. The challenges are formidable, but these companies claim that technology has reached the point where mass-produced humanoid robotic systems are a realistic near-term goal.

Nvidia's Role in the Next Generation of Humanoid Robots

One often-repeated phrase at this year's Nvidia GTC was "Physical AI," the kind of AI that allows robots to operate effectively in the physical world. To get there, Nvidia offers its Isaac GR00T N1, an open-source foundation model that is pre-trained with some of the basics and broken down into a fast system for taking action and a slow system for planning action. For teaching robots new tasks, lots of varied data is needed. Tele-operation and video demonstrations are important but usually not enough. Synthetic data generated from training in simulation helps fill the gap, and this is where Nvidia's Omniverse with Cosmos comes in. "Omniverse is a platform that brings together data between different sources and allows you to build a digital twin that's physically accurate," says Akhil Docca, senior project marketing manager for Omniverse. "Cosmos is important because it allows not only the augmentation of data for photorealism, it also allows you to create an exponentially large [amount] of data." This was illustrated during Nvidia CEO Jensen Huang's keynote speech when a digital twin of a robot was shown in candy-like bright colors in Omniverse. The same image was shown cycling through several different iterations with lifelike shading and texture added by Cosmos using simple text prompts. To make that synthetic data as applicable to the real world as possible, a detailed representation of the laws of physics is required. Nvidia partnered with DeepMind and Disney to develop the open-source physics engine Newton. During his keynote, Huang brought out a remote-controlled BDX droid to celebrate the announcement. Much of this work on humanoid robots is framed as an answer to looming anxiety about labor shortages. In his keynote, Huang predicted a worldwide shortage of 50 million workers by the end of the decade and envisioned a future where we might pay robots "$50,000 per year" to work. To get more clarity on Huang's comments, I contacted Nvidia, which told me the predicted shortage of 50 million workers by the end of the decade was based on a combination of estimates from various sources, industries and locations. The sources Nvidia provided included estimates regarding manufacturing, renewable energy, trucking, nursing, hospitality workers in the United States and Europe, domestic workers in US and Europe, and construction workers in the US and Europe. A representative from Nvidia also clarified the company's anticipated pricing model for humanoid robots, saying "It's more like subscribing to full self-driving services for an autonomous vehicle. The customer buys the car and then pays for the services it wants on top. We believe robots will be the same. Buy the robot and pay for services on top." To see these technologies in action, check out the video in this article.

Nvidia says 'the age of generalist robotics is here'

Andrew Liszewski is a senior reporter who's been covering and reviewing the latest gadgets and tech since 2011, but has loved all things electronic since he was a kid. Nvidia has announced that Isaac GR00T N1 -- the company's open-source, pretrained but customizable foundation model that's designed to expedite the development and capabilities of humanoid robots -- is now available. "The age of generalist robotics is here," says Nvidia founder and CEO, Jensen Huang. "With Nvidia Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI." During his GTC 2025 keynote today, Huang demonstrated 1X's NEO Gamma humanoid robot performing autonomous tidying jobs using a post-trained policy built on the GR00T N1 model. "The future of humanoids is about adaptability and learning," says 1X Technologies CEO Bernt Børnich. "NVIDIA's GR00T N1 model provides a major breakthrough for robot reasoning and skills. With a minimal amount of post-training data, we were able to fully deploy on NEO Gamma -- furthering our mission of creating robots that are not tools, but companions that can assist humans in meaningful, immeasurable ways." You might recall seeing this freakishly lifelike bot a few weeks ago in Nothing's teaser for its latest phone. We didn't post it because it looked like another human in a robot suit -- thanks, Elon. Other companies developing humanoid robots who have had early access to the GR00T N1 model include Boston Dynamics, the creators of Atlas; Agility Robotics; Mentee Robotics; and Neura Robotics. Originally announced as Project GR00T a year ago, the GR00T N1 foundation model utilizes a dual-system architecture inspired by human cognition. System 1, as Nvidia calls it, is described as a "fast-thinking action model" that behaves similarly to human reflexes and intuition. It was trained on data collected through human demonstrations and synthetic data generated by Nvidia's Omniverse platform. System 2, which is powered by a vision language model, is a "slow-thinking model" that "reasons about its environment and the instructions it has received to plan actions." Those plans are passed along to System 1, which translates them into "precise, continuous robot movements" that include grasping, moving objects with one or two arms, as well as more complex multistep tasks that involve combinations of basic skills. While the GR00T N1 foundation model is pretrained with generalized humanoid reasoning and skills, developers can customize its behavior and capabilities for specific needs by post-training it with data gathered from human demonstrations or simulations. Nvidia has made GR00T N1 training data and task evaluation scenarios available for download through Hugging Face and GitHub.

NVIDIA unveil world's first open humanoid robot foundation model

The launch also included the Isaac GR00T Blueprint for synthetic data generation and Newton, a physics engine developed with Google DeepMind and Disney Research. The tech giant also introduced Cosmos World Foundation Models (WFMs), offering developers unprecedented control over AI-driven world generation and reasoning. The announcements mark a significant step in NVIDIA's mission to bridge the gap between AI and the physical world, opening up new possibilities across multiple industries. One of the most anticipated reveals, NVIDIA Isaac GR00T N1, is the first in a series of fully customizable models that NVIDIA will pre-train and offer to robotics developers worldwide -- helping industries tackle global labor shortages, which exceed 50 million workers. The model features a dual-system architecture inspired by human cognition. System 1 operates as a rapid-response mechanism, executing instinctive actions, while System 2 takes a deliberate, analytical approach to decision-making. Using a vision-language model, System 2 interprets its environment and instructions before System 1 translates those insights into precise movements.

Nvidia will supercharge humanoid robot development with Isaac GR00T N1 foundation model for human-like reasoning

Nvidia announced a portfolio of technologies to supercharge humanoid robot development, including Nvidia Isaac GR00T N1, the world's first open, fully customizable foundation model for generalized humanoid reasoning and skills. The other technologies include simulation frameworks and blueprints such as the Nvidia Isaac GR00T Blueprint for generating synthetic data, as well as Newton, an open-source physics engine -- under development with Google DeepMind and Disney Research -- purpose-built for developing robots. Available now, GR00T N1 is the first of a family of fully customizable models that Nvidia will pretrain and release to worldwide robotics developers -- accelerating the transformation of industries challenged by global labor shortages estimated at more than 50 million people. "The age of generalist robotics is here," said Jensen Huang, founder and CEO of Nvidia, in a statement. "With Nvidia Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI." The company unveiled the news during Huang's keynote speech at the GTC 2025 event. "This could be the biggest industry of all," Huang said. GR00T N1 Advances Humanoid Developer Community The GR00T N1 foundation model features a dual-system architecture, inspired by principles of human cognition. "System 1" is a fast-thinking action model, mirroring human reflexes or intuition. "System 2" is a slow-thinking model for deliberate, methodical decision-making. Powered by a vision language model, System 2 reasons about its environment and the instructions it has received to plan actions. System 1 then translates these plans into precise, continuous robot movements. System 1 is trained on human demonstration data and a massive amount of synthetic data generated by the Nvidia Omniverse platform. GR00T N1 can easily generalize across common tasks -- such as grasping, moving objects with one or both arms, and transferring items from one arm to another -- or perform multistep tasks that require long context and combinations of general skills. These capabilities can be applied across use cases such as material handling, packaging and inspection. Developers and researchers can post-train GR00T N1 with real or synthetic data for their specific humanoid robot or task. In his GTC keynote, Huang demonstrated 1X's humanoid robot autonomously performing domestic tidying tasks using a post-trained policy built on GR00T N1. The robot's autonomous capabilities are the result of an AI training collaboration between 1X and Nvidia. "The future of humanoids is about adaptability and learning," said Bernt Børnich, CEO of 1X Technologies, in a statement. "Nvidia's GR00T N1 model provides a major breakthrough for robot reasoning and skills. With a minimal amount of post-training data, we were able to fully deploy on NEO Gamma -- furthering our mission of creating robots that are not tools, but companions that can assist humans in meaningful, immeasurable ways." Among the additional leading humanoid developers worldwide with early access to GR00T N1 are Agility Robotics, Boston Dynamics, Mentee Robotics and Neura Robotics. Nvidia, Google DeepMind and Disney Research Focus on Physics Nvidia announced a collaboration with Google DeepMind and Disney Research to develop Newton, an open-source physics engine that lets robots learn how to handle complex tasks with greater precision. Built on the Nvidia Warp framework, Newton will be optimized for robot learning and compatible with simulation frameworks such as Google DeepMind's MuJoCo and Nvidia Isaac Lab. Additionally, the three companies plan to enable Newton to use Disney's physics engine. Google DeepMind and Nvidia are collaborating to develop MuJoCo-Warp, which is expected to accelerate robotics machine learning workloads by more than 70 times and will be available to developers through Google DeepMind's MJX open-source library, as well as through Newton. Disney Research will be one of the first to use Newton to advance its robotic character platform that powers next-generation entertainment robots, such as the expressive Star Wars-inspired BDX droids that joined Huang on stage during his GTC keynote. "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, senior vice president at Walt Disney Imagineering Research & Development, in a statement. "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." Nvidia and Disney Research, along with Intrinsic, announced an additional collaboration to build OpenUSD pipelines and best practices for robotics data workflows. More Data to Advance Robotics Post-Training Large, diverse, high-quality datasets are critical for robot development but costly to capture. For humanoids, real-world human demonstration data is limited by a person's 24-hour day. Announced today, the Nvidia Isaac GR00T Blueprint for synthetic manipulation motion generation helps address this challenge. Built on Omniverse and Nvidia Cosmos Transfer world foundation models, the blueprint lets developers generate exponentially large amounts of synthetic motion data for manipulation tasks from a small number of human demonstrations. Using the first components available for the blueprint, Nvidia generated 780,000 synthetic trajectories -- the equivalent of 6,500 hours, or nine continuous months, of human demonstration data -- in just 11 hours. Then, combining the synthetic data with real data, Nvidia improved GR00T N1's performance by 40%, compared with using only real data. To further equip the developer community with valuable training data, Nvidia is releasing the GR00T N1 dataset as part of a larger open-source physical AI dataset -- also announced at GTC and now available on Hugging Face. Availability Nvidia GR00T N1 training data and task evaluation scenarios are now available for download from Hugging Face and GitHub. The Nvidia Isaac GR00T Blueprint for synthetic manipulation motion generation is also now available as an interactive demo on build.nvidia.com or to download from GitHub. The Nvidia DGX Spark personal AI supercomputer, also announced today at GTC, provides developers a turnkey system to expand GR00T N1's capabilities for new robots, tasks and environments without extensive custom programming. The Newton physics engine is expected to be available later this year. At GTC 2025, Nvidia will hold Humanoid Developer Day sessions, including: ● "An Introduction to Building Humanoid Robots" for a deep dive into Nvidia Isaac GR00T; ● "Insights Into Disney's Robotic Character Platform" to learn how Disney Research redefines entertainment robotics with BDX droids; ● "Announcing Mujoco-Warp and Newton: How Google DeepMind and Nvidia are Supercharging Robotics Development" for a deeper look into these new technologies and how Google deploys AI models to train AI-powered humanoids for real-world tasks.

The age of generalist robotics is here" - Nvidia's latest GROOT AI model just took us another step closer to fully humanoid robots

Nvidia has taken the world a step closer to smart, humanoid robots with the launch of its latest advanced AI model. At its Nvidia GTC 2025 event, the company revealed Isaac GROOT N1, which it says is, "the world's first open Humanoid Robot foundation model", alongside several other important development tools. Nvidia says its tools, which are available now, will make developing smarter and more functional robots easier than ever, along with allowing them to have more humanoid reasoning and skills - which doesn't sound terrifying at all. Isaac GROOT N1 "The age of generalist robotics is here," said Jensen Huang, founder and CEO of NVIDIA. "With NVIDIA Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI." The company says its robotics work can help fill a shortfall of more than 50 million caused by a global labor shortage. Nvidia says Isaac GROOT N1, which can be trained on real or synthetic data, can "easily" master tasks such as grasping, moving objects with either a single or multiple arms, and moving items from one arm to the other - but can also carry out multi-step tasks which combine a number of general skills. The model is built across a dual-system architecture inspired by the principles of human cognition, with "System 1" is a fast-thinking action model, mirroring human reflexes or intuition, whereas "System 2" is a slow-thinking model for "deliberate, methodical decision-making." Powered by a vision language model, System 2 is able to consider and analyze its environment, and the instructions it was given, to plan actions - which are then translated by System 1 into precise, continuous robot movements. Among the other tools being released are a range of simulation frameworks and blueprints such as the NVIDIA Isaac GR00T Blueprint for generating synthetic data, which help generate large, detailed synthetic data sets needed for robot development which would be prohibitively expensive to gather in real life. There is also Newton, an open source physics engine, created alongside Google DeepMind and Disney Research, which Nvidia says is purpose-built for developing robots. Huang was joined on stage by Star Wars-inspired BDX droids during his GTC keynote, showing the possibilities of the technology in theme parks or other entertainment locations. Nvidia first launched Project GROOT ("Generalist Robot 00 Technology") at GTC 2024, primarily focusing on industrial use cases, which could learn and become smarter by watching human behaviour, understanding natural language and emulating movements, allowing them to quickly learn coordination, dexterity and other skills in order to navigate, adapt and interact with the real world.

NVIDIA Announces Isaac GR00T N1 -- the World's First Open Humanoid Robot Foundation Model -- and Simulation Frameworks to Speed Robot Development

GTC -- NVIDIA today announced a portfolio of technologies to supercharge humanoid robot development, including NVIDIA Isaac GR00T N1, the world's first open, fully customizable foundation model for generalized humanoid reasoning and skills. The other technologies include simulation frameworks and blueprints such as the NVIDIA Isaac GR00T Blueprint for generating synthetic data, as well as Newton, an open-source physics engine -- under development with Google DeepMind and Disney Research -- purpose-built for developing robots. Available now, GR00T N1 is the first of a family of fully customizable models that NVIDIA will pretrain and release to worldwide robotics developers -- accelerating the transformation of industries challenged by global labor shortages estimated at more than 50 million people. "The age of generalist robotics is here," said Jensen Huang, founder and CEO of NVIDIA. "With NVIDIA Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI." GR00T N1 Advances Humanoid Developer Community The GR00T N1 foundation model features a dual-system architecture, inspired by principles of human cognition. "System 1" is a fast-thinking action model, mirroring human reflexes or intuition. "System 2" is a slow-thinking model for deliberate, methodical decision-making. Powered by a vision language model, System 2 reasons about its environment and the instructions it has received to plan actions. System 1 then translates these plans into precise, continuous robot movements. System 1 is trained on human demonstration data and a massive amount of synthetic data generated by the NVIDIA Omniverse™ platform. GR00T N1 can easily generalize across common tasks -- such as grasping, moving objects with one or both arms, and transferring items from one arm to another -- or perform multistep tasks that require long context and combinations of general skills. These capabilities can be applied across use cases such as material handling, packaging and inspection. Developers and researchers can post-train GR00T N1 with real or synthetic data for their specific humanoid robot or task. In his GTC keynote, Huang demonstrated 1X's humanoid robot autonomously performing domestic tidying tasks using a post-trained policy built on GR00T N1. The robot's autonomous capabilities are the result of an AI training collaboration between 1X and NVIDIA. "The future of humanoids is about adaptability and learning," said Bernt Børnich, CEO of 1X Technologies. "While we develop our own models, NVIDIA's GR00T N1 provides a significant boost to robot reasoning and skills. With minimal post-training data, we fully deployed on NEO Gamma -- advancing our mission of creating robots that are not just tools, but companions capable of assisting humans in meaningful, immeasurable ways." Among the additional leading humanoid developers worldwide with early access to GR00T N1 are Agility Robotics, Boston Dynamics, Mentee Robotics and NEURA Robotics. NVIDIA, Google DeepMind and Disney Research Focus on Physics NVIDIA announced a collaboration with Google DeepMind and Disney Research to develop Newton, an open-source physics engine that lets robots learn how to handle complex tasks with greater precision. Built on the NVIDIA Warp framework, Newton will be optimized for robot learning and compatible with simulation frameworks such as Google DeepMind's MuJoCo and NVIDIA Isaac™ Lab. Additionally, the three companies plan to enable Newton to use Disney's physics engine. Google DeepMind and NVIDIA are collaborating to develop MuJoCo-Warp, which is expected to accelerate robotics machine learning workloads by more than 70x and will be available to developers through Google DeepMind's MJX open-source library, as well as through Newton. Disney Research will be one of the first to use Newton to advance its robotic character platform that powers next-generation entertainment robots, such as the expressive Star Wars-inspired BDX droids that joined Huang on stage during his GTC keynote. "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, senior vice president at Walt Disney Imagineering Research & 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." NVIDIA and Disney Research, along with Intrinsic, announced an additional collaboration to build OpenUSD pipelines and best practices for robotics data workflows. More Data to Advance Robotics Post-Training Large, diverse, high-quality datasets are critical for robot development but costly to capture. For humanoids, real-world human demonstration data is limited by a person's 24-hour day. Announced today, the NVIDIA Isaac GR00T Blueprint for synthetic manipulation motion generation helps address this challenge. Built on Omniverse and NVIDIA Cosmos Transfer world foundation models, the blueprint lets developers generate exponentially large amounts of synthetic motion data for manipulation tasks from a small number of human demonstrations. Using the first components available for the blueprint, NVIDIA generated 780,000 synthetic trajectories -- the equivalent of 6,500 hours, or nine continuous months, of human demonstration data -- in just 11 hours. Then, combining the synthetic data with real data, NVIDIA improved GR00T N1's performance by 40%, compared with using only real data. To further equip the developer community with valuable training data, NVIDIA is releasing the GR00T N1 dataset as part of a larger open-source physical AI dataset -- also announced at GTC and now available on Hugging Face. Availability NVIDIA GR00T N1 training data and task evaluation scenarios are now available for download from Hugging Face and GitHub. The NVIDIA Isaac GR00T Blueprint for synthetic manipulation motion generation is also now available as an interactive demo on build.nvidia.com or to download from GitHub. The NVIDIA DGX Spark personal AI supercomputer, also announced today at GTC, provides developers a turnkey system to expand GR00T N1's capabilities for new robots, tasks and environments without extensive custom programming. The Newton physics engine is expected to be available later this year.

Nvidia's new model brings humanoid robotics a step closer to reality

Nvidia has unveiled a new model which could form the foundation of future humanoid robotics projects, representing a step forward toward a reality of generalist robots. The Nvidia Isaac Groot N1 model is open source and fully customizable for various robotics uses, inspired by the way that humans think. It consists of two systems, one of which is fast-reacting, similar to human reflexes, while the other is slower-thinking for careful decision making. Recommended Videos The slower system can perform reasoning, such as assessing its environment and planning out how to perform a commanded action. Then the faster system can turn that plan quickly into a series of movements for the robot to perform. The model can also generalize, so it can learn from handling one object about how to handle other kinds of objects too. Get your weekly teardown of the tech behind PC gaming ReSpec Subscribe Check your inbox! Privacy Policy The model will be available worldwide to robotics developers, who can build on it to meet their own particular robotics needs. A humanoid robot from 1X trained on Groot N1 was used to demonstrate the capabilities, autonomously performing a tidying up task for the home. "The future of humanoids is about adaptability and learning," said Bernt Børnich, CEO of 1X Technologies. "While we develop our own models, NVIDIA's GR00T N1 provides a significant boost to robot reasoning and skills. With minimal post-training data, we fully deployed on NEO Gamma -- advancing our mission of creating robots that are not just tools, but companions capable of assisting humans in meaningful, immeasurable ways." Other well-known names in robotics including Boston Dynamics and Agility Robotics have already gained access to Groot N1 for their development, and Nvidia says it is working on a whole family of related pre-trained robotics models. "The age of generalist robotics is here," said Jensen Huang, founder and CEO of NVIDIA. "With NVIDIA Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI."

Nvidia announces new AI models for smarter, more adaptable robots

Nvidia announces new AI models for smarter, more adaptable robots Nvidia Corp. today unveiled new technologies and artificial intelligence models that help developers build smarter robots, including humanoids, and self-driving vehicles faster by automating the complex modeling and data work. These included Nvidia Isaac GR00T N1, the world's first open-source, fully customizable generalist foundation model for humanoid robot brains, the GR00T blueprint for generating synthetic data and Newton, an open-source physics engine purpose-built for developing robots. The company also announced a major release of new Nvidia Cosmos world foundation models, with customizable reasoning for physical AI development and world generation. Physical AI refers to the integration of artificial intelligence into physical systems such as robots and other machines to allow them to perceive and react to the real world. It combines AI algorithms with physical hardware that can sense, act and adapt to changing conditions. Emerging modern-day examples include robotics systems and autonomous vehicles. "The age of generalist robotics is here," said Jensen Huang, founder and chief executive of Nvidia. "With Nvidia Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI." As humanoid robot adoption becomes an increasing trend, developers will be challenged to train AI in tasks and test them at scale. GR00T N1 can easily generalize across common robot tasks - such as grasping, moving objects with one or both arms and transferring items from one arm to another - it can also perform multistep tasks that combine steps. It's also designed to be easily post-trained with real or synthetic data for industry-specific skills. In his keynote at this year's Nvidia GTC, Huang demonstrated a humanoid robot from the startup 1X tidying up using GR00T N1 AI training policies. Nvidia GR00T N1 training data and task evaluation datasets are available open-source for download from Hugging Face and GitHub. In collaboration with Google DeepMind, Google LLC's AI research lab, and Disney Research, Nvidia developed Newton, an open-source physics engine that allows robots to learn how to handle complex tasks with greater precision. Built on the Nvidia Warp framework, DeepMind and Nvidia are collaborating to develop the simulation framework MuJoCo-Warp, which is expected to accelerate robotics development by more than 70x through DeepMind's MJX open-source library and Newton. Robotics requires massive amounts of data, but is extremely costly to capture. For humanoids, real-world data capture must be done my demonstration from actual people. To aid with this, Nvidia announced the Isaac GR00T Blueprint for synthetic manipulation motion generation. This blueprint Nvidia generated 780,000 synthetic trajectories - the company said this is equivalent to 6,500 hours, or nine continuous months, of human demonstration data - in just 11 hours. AI models for physical AI training and reasoning World foundation models, or WFMs, empower developers and engineers to create virtual training grounds where robots learn to navigate real-world challenges through simulated environments so they can be trained across various scenarios. As part of today's announcements, Nvidia announced the additions of a family of new AI models within the company's Cosmos WFMs, introducing breakthroughs for synthetic data generation, high-fidelity world generation and multimodal reasoning. "Just as large language models revolutionized generative and agentic AI, Cosmos world foundation models are a breakthrough for physical AI," said Huang. "Cosmos introduces an open and fully customizable reasoning model for physical AI and unlocks opportunities for step-function advances in robotics and the physical industries." Cosmos Transfer WFMs can produce vast amounts of synthetic data for robotic training by ingesting structured video inputs such as segmentation maps, depth maps, lidar scans, pose estimations and trajectories to generate photoreal video outputs. The model streamlines perception AI training by allowing developers to produce large-scale photorealistic videos for large-scale datasets. "Cosmos offers us an opportunity to scale our photorealistic training data beyond what we can feasibly collect in the real world," said Pras Velagapudi, chief technology officer of Agility Robotics Inc., an American humanoid robotics company. Foretellix Ltd., a verification and validation solution for self-driving car developers, uses Cosmos Transfer to create variations of physically based sensor data at scale using the Nvidia Omniverse Blueprint for autonomous vehicle simulation. It can produce numerous different conditions including heavy weather, bad driving conditions, twilight and other road hazards for the driving robot to train with. Cosmos Predict, announced during the CES trade show in January, can generate virtual worlds from text, images and video. The newest models will allow developers to input a first frame and a last frame of video and predict the frames in-between, generating all of the intermediate actions and motions that took place. It is purpose-built for post-training physical AI models to provide them the skills to understand what happened during the video. Robotics developers are already using Cosmos Predict and Transfer, including 1X to train its new humanoid robot Neo Gamma and robot brain developer Skild AI Inc. is using Transfer to generate synthetic training data. Cosmos Reason provides a fully customizable WFM that uses chain-of-thought reasoning to understand video data and predict the outcomes of interactions through natural language processing. For example, it can understand what might happen next from watching a person stepping onto a crosswalk or a box falling from a shelf. Developers can distill the Reason model to enhance existing world foundation models or create new vision language action models for robotics. It can also be used to improve physical AI data annotation and curation, which can help produce accurate datasets for understanding the world. It can also be used to post-train high-level planners for multi-step processes needed orchestrate physical AI to complete tasks.

GR00T N1: Nvidia's next leap in AI-powered robotics

Nvidia announced a suite of technologies to enhance humanoid robot development, including the Nvidia Isaac GR00T N1, the first open, fully customizable foundation model designed for human-like reasoning and skills. This announcement took place during Jensen Huang's keynote at the GTC 2025 event. Nvidia GR00T N1 and its capabilities The Nvidia Isaac GR00T N1 foundation model introduces a dual-system architecture inspired by human cognition. "System 1" acts as a fast-thinking action model emulating human reflexes and intuition, while "System 2" serves as a slow-thinking model for methodical decision-making. Powered by a vision language model, System 2 examines the environment and instructions to devise actions, which System 1 then executes as precise robot movements. GR00T N1 has the ability to generalize across various tasks, including grasping and moving objects, as well as performing multistep tasks that require contextual understanding and the combination of general skills. These capabilities are applicable in fields such as material handling, packaging, and inspection. Researchers can further train GR00T N1 using real or synthetic data tailored to specific humanoid robots or tasks. In his keynote, Huang showcased 1X's humanoid robot autonomously completing domestic tidying tasks based on a policy built on GR00T N1. "The future of humanoids is about adaptability and learning," stated Bernt Børnich, CEO of 1X Technologies. He emphasized that Nvidia's GR00T N1 model is a significant advancement in robot reasoning and skills, enabling the development of robots as companions rather than mere tools. AI at warp speed: Nvidia's new GB300 superchip arrives this year Collaboration on physics and data generation Nvidia also partnered with Google DeepMind and Disney Research to develop Newton, an open-source physics engine designed to help robots perform complex tasks with precision. Built on the Nvidia Warp framework, Newton is optimized for robot learning and will work with simulation frameworks such as Google DeepMind's MuJoCo and Nvidia Isaac Lab. Additionally, the collaboration aims to implement Disney's physics engine into Newton. Google DeepMind and Nvidia are working on MuJoCo-Warp, expected to enhance robotics machine learning workloads by over 70 times, and this technology will be accessible through Google DeepMind's MJX open-source library as well as via Newton. Disney Research plans to utilize Newton to enhance its robotic character platform, which includes next-generation entertainment robots like the Star Wars-inspired BDX droids, which were highlighted during Huang's presentation. "This collaboration will allow us to create a new generation of robotic characters that are more expressive and engaging," remarked Kyle Laughlin, senior vice president at Walt Disney Imagineering Research & Development. Nvidia introduced the Isaac GR00T Blueprint for synthetic manipulation motion generation, aimed at addressing the need for large, diverse datasets for robot development. This blueprint allows developers to create vast amounts of synthetic motion data from limited human demonstrations, generating 780,000 synthetic trajectories in a short timeframe. The combination of synthetic and real data improved GR00T N1's performance by 40%. Nvidia is releasing the GR00T N1 dataset as part of an open-source physical AI dataset, now available on Hugging Face. Nvidia GR00T N1 training data and task evaluation scenarios can now be downloaded from Hugging Face and GitHub. The Nvidia DGX Spark personal AI supercomputer, announced during the event, provides developers with a system to enhance GR00T N1's capabilities across new robots, tasks, and environments. The Newton physics engine is anticipated to be released later this year. During GTC 2025, Nvidia will host Humanoid Developer Day sessions focused on various aspects of humanoid robotics, including in-depth discussions on the Nvidia Isaac GR00T, insights into Disney's Robotic Character Platform, and announcements regarding Mujoco-Warp and Newton.

NVIDIA Isaac GR00T N1 Humanoid Reasoning & Skills Model and Robot Introduced by NVIDIA CEO

NVIDIA has unveiled the Isaac GR00T N1, the world's first open, fully customizable foundation model for generalized humanoid reasoning and skills. This advanced system is designed to address the pressing global labor shortage by combining robotics and artificial intelligence (AI) in a way that promises to reshape industries. By using NVIDIA's innovative ecosystem, the Isaac GR00T N1 aims to streamline the development, training, and deployment of robotic systems, setting a new standard in humanoid robotics. This innovation highlights the growing role of robotics in solving real-world challenges and advancing technological progress. Imagine a world where robots don't just assemble cars or vacuum floors but actively collaborate with humans to solve complex problems, tackle labor shortages, and transform entire industries. It might sound like science fiction, but NVIDIA's latest innovation, the Isaac GR00T N1, is bringing us closer to this reality. By the end of the decade, a staggering 50 million workers are expected to be missing from the global workforce, leaving industries scrambling to fill critical gaps. This isn't just a statistic -- it's a challenge that could affect everything from how products are made to how patients are cared for. Isaac GR00T N1 Robot Enter the Isaac GR00T N1, a foundation model designed not only to perform intricate tasks but also to collaborate seamlessly with humans. With NVIDIA's robust ecosystem of AI tools and technologies backing it, this innovation offers a glimpse into a future where robotics doesn't just replace human effort but enhances it, paving the way for smarter, more efficient solutions across industries. Tackling the Global Labor Shortage The world faces a looming labor crisis, with a projected shortage of 50 million workers by 2030. This shortage is driven by factors such as aging populations, declining birth rates, and increasing demand for skilled labor in key industries. To address this challenge, robotics is emerging as a critical solution. Humanoid robots trained with the foundation model are specifically designed to fill labor gaps by performing complex tasks, collaborating seamlessly with human workers, and enhancing productivity across essential sectors. Industries such as manufacturing, logistics, and healthcare stand to benefit significantly from this technology. In manufacturing, robots can handle repetitive or hazardous tasks, reducing workplace injuries and improving efficiency. In logistics, they can optimize supply chain operations by automating sorting, packing, and delivery processes. In healthcare, humanoid robots can assist with patient care, medical procedures, and administrative tasks, alleviating the burden on overworked professionals. By addressing these needs, the Isaac GR00T N1 represents a pivotal step toward mitigating the global labor shortage. Isaac GR00T N1: A Cognitive Breakthrough At the core of the Isaac GR00T N1 lies a sophisticated dual-system architecture inspired by human cognition. This innovative design integrates two distinct systems that work together to enable advanced functionality: This dual-system approach allows the robot to perform a wide range of tasks, from manipulating objects with precision to executing multi-step operations. Additionally, the Isaac GR00T N1 is designed to collaborate effectively with human operators, making it a versatile tool for addressing diverse industrial challenges. Its ability to combine cognitive reasoning with physical dexterity positions it as a fantastic asset in the field of robotics. Here are more guides from our previous articles and guides related to Humanoid robotics that you may find helpful. NVIDIA's Robotics Ecosystem: The Backbone of Innovation The Isaac GR00T N1 is supported by NVIDIA's robust robotics development ecosystem, which integrates several advanced technologies to enhance its capabilities. These technologies form the foundation for the robot's development, training, and deployment: These tools work in harmony to streamline the development process, from virtual training to real-world implementation. By using this ecosystem, NVIDIA ensures that the Isaac GR00T N1 is equipped with the tools needed to operate efficiently and accurately in complex environments. Training with Synthetic Data The Isaac GR00T N1 relies heavily on synthetic data generation for training, a method that overcomes the limitations of traditional real-world data collection. Synthetic data allows the robot to learn and adapt in controlled environments, accelerating the training process and improving outcomes. Two key approaches are used in this training: Digital twins -- virtual replicas of physical systems -- play a crucial role in this process. These replicas enable developers to test and refine the robot's capabilities in a risk-free environment before deployment. By using synthetic data and digital twins, the Isaac GR00T N1 achieves a high level of reliability and adaptability, making sure its effectiveness in real-world applications. Overcoming Scalability Challenges Scaling humanoid robotics like the Isaac GR00T N1 presents several technical challenges that must be addressed to enable widespread adoption. Key challenges include: NVIDIA addresses these challenges by integrating generative AI to enhance data diversity and employing verifiable physics rewards to ensure realistic robot behavior. These advancements streamline the training and deployment processes, making it possible to scale humanoid robotics efficiently and effectively. Open-Sourcing the Isaac GR00T N1 To accelerate innovation in humanoid robotics, NVIDIA has announced that the Isaac GR00T N1 will be open sourced. This decision invites researchers, developers, and organizations worldwide to collaborate, share insights, and contribute to the advancement of robotics technology. By fostering a global community, NVIDIA aims to drive rapid progress in this field, allowing the development of new applications and solutions that address critical challenges. The Future of Robotics and AI As robotics and physical AI continue to evolve, they are poised to play a central role in shaping the global economy. NVIDIA's commitment to advancing AI infrastructure -- spanning cloud computing, enterprise solutions, and robotics applications -- underscores its dedication to driving progress in this fantastic field. The Isaac GR00T N1 foundation model represents a significant milestone in humanoid robotics, offering a glimpse into a future where robots enhance human capabilities, solve pressing challenges, and redefine the boundaries of innovation.

Nvidia at GTC 2025: Introducing Groot N1, the Future of Generalist Robotics

In his keynote, Nvidia CEO Jensen Huang announced, "The age of generalist robotics is here." He emphasized how Groot N1, along with new data-generation and , will promote humanoid robotics innovation. It is set to revolutionize AI-driven automation. Nvidia's Groot N1 is a pre-trained AI foundation model that can be further post-trained by developers to meet particular robotic applications. The AI was trained on available datasets and synthetically created data, allowing for better learning and adaptability of robots in real-world environments.

NVIDIA unveils Isaac GR00T N1, advancing humanoid robot market By Investing.com

Investing.com -- NVIDIA (NASDAQ:NVDA) has launched the Isaac GR00T N1 foundation model, an open-source model designed for general-purpose humanoid robots, according to TrendForce. The unveiling took place at GTC 2025, and the model's features include comprehensive datasets, multi-modal inputs, and adaptable training capabilities. These features allow researchers to customize the GR00T N1 using both real and synthetic data to suit specific environments and applications. The introduction of GR00T N1 is anticipated to stimulate early market adoption of humanoid robots by speeding up AI model training and reducing the time-to-market. The total market value of the industry is projected to near US$4 billion by 2028. Figure AI and Google (NASDAQ:GOOGL) have also stepped into the humanoid robotics sector with their respective models, Helix and Gemini Robotics. These models, like GR00T N1, combine rapid motion planning with cognitive decision-making to improve robotic autonomy. Nevertheless, NVIDIA's superior dataset quality and deployment flexibility provide it with a competitive advantage over its competitors. Jensen Huang, CEO of NVIDIA, has voiced strong confidence in humanoid robots, viewing them as a significant catalyst for next-generation AI computing power. NVIDIA has established a comprehensive technology ecosystem in this field, offering embedded AI computing solutions such as the Jetson Orin NX and Nano for mid-to-small humanoid robots, and the AGX Orin for high-performance applications that can deliver up to 275 TOPS. NVIDIA Omniverse plays a crucial role in training simulations and digital twins, and it's heavily relied upon by leading Western humanoid robot developers. Major Chinese firms, including UBTECH, Unitree, and Fourier Intelligence, also use Omniverse for AI model training. NVIDIA expanded its robotics ecosystem at GTC 2025 with the introduction of the open-source physics engine Newton, the synthetic data generation tool GR00T Blueprint, and the GR00T N1 foundation model, reinforcing its dominance in the humanoid robotics sector. While GR00T N1 is open-source, AI model parameters are often closely tied to specific hardware configurations. TrendForce predicts that humanoid robot developers using GR00T N1 will likely choose NVIDIA's hardware solutions to maximize efficiency. This approach, while offering an open-source model, maintains a closed ecosystem, building a protective barrier around NVIDIA's dominance in humanoid robotics, similar to how its CUDA technology and GPU chips have cemented its leadership in the AI industry.