“NVIDIA Introduces Open-Source Foundation Model for Humanoid Robots”

Market research company TrendForce reports that NVIDIA has introduced the Isaac GR00T N1 foundation model for general-purpose humanoid robots at GTC 2025. Featuring comprehensive datasets, multi-modal inputs, and adaptable training capabilities, the open-source model enables researchers to fine-tune GR00T N1 with real and synthetic data for specific environments and applications. TrendForce expects GR00T N1 to drive early market adoption of humanoid robots by accelerating AI model training and reducing time-to-market, with the industry’s total market value projected to approach US$4 billion by 2028.

Figure AI and Google have entered the humanoid robotics sector with Helix and Gemini Robotics. Similar to GRooT N1, these models incorporate rapid motion planning and cognitive decision-making for improved autonomy. However, NVIDIA contends that its dataset quality and deployment flexibility are superior.

NVIDIA CEO Jensen Huang has conveyed strong confidence in humanoid robots and sees them as a major driver of next-generation AI computing power. The company has built 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 AGX Orin for high-performance applications—delivering up to 275 TOPS.

NVIDIA Omniverse plays a vital role in training simulations and digital twins, with leading Western humanoid robot developers depending on the platform. Major Chinese players, including UBTECH, Unitree, and Fourier Intelligence, also utilise Omniverse for AI model training. At GTC 2025, NVIDIA further expanded its robotics ecosystem 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 humanoid robotics.

While GR00T N1 is open-source, AI model parameters are often closely linked to specific hardware configurations. TrendForce anticipates that developers of humanoid robots using GR00T N1 will likely opt for NVIDIA’s hardware to enhance efficiency. This strategy, even with the provision of an open-source model, upholds a closed ecosystem, thereby creating a “moat” around NVIDIA’s dominance in humanoid robotics, similar to how its CUDA technology and GPU chips have reinforced its leadership in the AI industry.

Source: Trendforce