At GTC 2025 in San Jose, Nvidia took a major step toward advancing humanoid robotics with the launch of Groot N1, an AI foundation model designed to power general-purpose robots.
Built as a “generalist” AI system, Groot N1 combines synthetic and real-world data to train robots that think, reason, and perform tasks with greater precision. Nvidia revealed that the model features a unique dual-system architecture, inspired by how humans process information—balancing quick reflexes with thoughtful reasoning.
“The age of generalist robotics is here,” declared Nvidia CEO Jensen Huang, underscoring the company’s ambitious vision for the future of robotics.
Groot N1 builds on Nvidia’s earlier Project Groot, which initially focused on industrial automation. With this new evolution, Nvidia expands its scope, targeting humanoid robots designed to operate in various environments and handle more complex tasks.
The model’s “slow-thinking” system allows robots to analyze their surroundings, process instructions, and map out the best course of action. Meanwhile, its “fast-thinking” system takes over to execute those plans through precise movements—even when tasks involve multiple steps or manipulating objects.
Nvidia isn’t stopping at just the model. It has also open-sourced Groot N1, giving developers access to simulation frameworks and detailed blueprints for generating synthetic training data. This move aims to accelerate innovation in humanoid AI and robotics development worldwide.
Interest in humanoid robots has soared, with startups like X1 and Figure racing to build general-purpose machines that mimic human motion. While the journey remains challenging, these companies believe recent technological leaps make mass-produced humanoid robots a near-term reality.
Still, the road ahead won’t be easy. The robotics industry has seen plenty of hype followed by setbacks. Yet, Nvidia’s entry—with Groot N1’s powerful capabilities—signals growing confidence that this time, humanoid robots might finally deliver on their long-promised potential.
