If AI is reshaping the world, then the next step is to ‘unleash its power’.
On August 25, NVIDIA officially launched its next-generation robotic computing platform—Jetson Thor. This platform is highly anticipated by Huang Renxun and several industry leaders, hailed as the ‘new brain for robots‘.
Explosive Performance
Jetson Thor is based on the latest Blackwell architecture, with an FP4 computing power of up to 2070 TFLOPS, which is a 7.5 times improvement over the previous generation Jetson Orin, with energy efficiency increased by 3.5 times and memory doubled.
This allows robots to run multiple large models locally, processing high-frequency data from cameras, LiDAR, and other sensors in real-time, enabling ultra-fast responses, such as answering your commands within 200 milliseconds and generating over 25 words per second, truly achieving natural and smooth human-robot dialogue.
Multi-Scenario Adaptability
Jetson Thor can now adapt to a wide range of scenarios, such as surgical assistance robots, intelligent logistics tractors, industrial robotic arms, and visual AI agents.
Currently, well-known international companies such as Agility Robotics, Boston Dynamics (which has integrated Thor into Atlas), Figure, and Amazon Robotics are among the first users.
Domestically, companies like UBTECH, Yushu Technology, Galaxy General, Zhiyuan Robotics, and United Imaging Healthcare have also taken the lead in application. Yushu Technology CEO Wang Xingxing stated: ‘Jetson Thor has brought a tremendous leap in computing power.’
Price Announcement: Developer Kit Starting at $3499
The starting price for the Jetson AGX Thor developer kit is $3499, which can be reduced to $2999 per unit for bulk purchases (over 1000 units).
Available globally from August 25 (domestic pricing has not yet been announced).
‘Physical AI’ is the Next Wave
Huang Renxun has repeatedly emphasized that AI has undergone three generations of evolution:
1️⃣ Conscious AI (can see, can hear)
2️⃣ Generative AI (can speak, can draw)
3️⃣ Agent AI (can plan, can decide)
Next Stop: Physical AI (can act in the real world)
Physical AI refers to AI systems that can understand physical rules, interact with the environment in real-time, and perform complex operations, typically applied in robotics, autonomous vehicles, and other physical entities.
NVIDIA predicts that Physical AI will unlock a trillion-dollar market in the future.
The launch of Jetson Thor is a solid step by NVIDIA towards the era of Physical AI. Stronger edge computing power, lower latency, and more reliable autonomous responses. What we are witnessing may be the eve of robots truly integrating into the human world.
The future is here, are you ready?
(Note: Image source from the internet, please delete if infringed)