Nvidia is releasing what it calls AI basic model for humanoid robotics.
The model, known as Groot N1, was announced at GTC 2025 in San Jose, is a “generalist” model that is trained on both synthetic and real data. NVIDIA claims that Groot N1 has a “dual system architecture” for “fast thinking and slow” and is inspired by human cognitive processes.
Groot N1 is the evolution of the NVIDIA project Groot, which the company launched last year at its GTC conference. The Groot project is specifically aimed at industrial use cases, but Groot N1 will expand its focus to a range of different forms of human robots.
NVIDIA says the Groot N1’s slow thinking system allows the robot to perceive and justify its environment and instructions, and then plan to take the right action. As for the quick thinking system, it translates the above plan into robotic actions, including actions involving manipulating objects through multiple steps.
Groot N1 is available in open source. In addition to models, NVIDIA is releasing simulation frameworks and blueprints for generating synthetic training data.
“The era of generalist robots is here,” NVIDIA CEO Jensen Huang said in a statement.
In recent years, humanoid robots have attracted a lot of publicity. The company likes it x1 and number Trying to create universal robots that are more or less like humans. The challenge is huge, but these companies claim that technology has reached the near-term goal of mass-producing humanoid robot systems.
this many disappointment In recent history of robotics, this will be easier said than done.
