A revolutionary ‘robot brain’ chip that dramatically reduces humanoid robot prototype development time has been launched. At the relatively accessible price of $3,499, it can simultaneously run large language models and vision AI, marking a new turning point in robotic automation.
Automation Impact NVIDIA’s Jetson AGX Thor delivers 7.5x faster processing speed and 3.5x improved energy efficiency compared to the previous Jetson Orin, revolutionarily shortening robot development processes. With 128GB high-capacity memory enabling simultaneous processing of multiple AI models, developers can reduce prototype creation time by over 70%. This enables integrated operation from vision-language-action models to large language models required for humanoid robot development.
Work Efficiency The Thor chip based on Blackwell GPU performs real-time AI inference and sensor data processing in parallel, showing over 90% latency reduction compared to existing distributed processing systems. Robot operation accuracy improves to 99.7%, maintaining stable performance even in complex work environments. Particularly in multimodal AI processing integrating autonomous driving, object recognition, and natural language processing, it achieves 5x faster response speed compared to existing solutions.
Safety Enhancement Thor chip’s real-time environment recognition and predictive AI functionality prevents over 85% of potential safety accidents during robot operations. High-performance vision systems and sensor fusion technology enable real-time collision risk detection between workers and robots, with emergency stopping capability within 0.1 seconds. Additionally, predictive maintenance functions detect robot system failures in advance, preventing unexpected work interruptions.
Cost Analysis The initial investment of $3,499 for the developer kit can reduce existing robot development costs by 40-60%. For mass production, units are supplied at $2,999 each, significantly reducing cost burden for robot manufacturers. Energy efficiency improvements reduce operating costs from an average of $2,000 to $500 monthly, with total cost of ownership (TCO) savings exceeding $1.3 million over three years of use.
Deployment Cases Global robotics companies including Amazon, Meta, and Boston Dynamics have begun developing next-generation humanoid robots utilizing Jetson chips. Domestically, Samsung Electronics and LG Electronics are reviewing adoption of these chipsets for household robots and industrial collaborative robots, with commercialized robot products expected to launch in earnest from next year.
