Key points
- Nvidia’s new Jetson AGX Thor robotics chip module is now available for $3,499 (developer kit).
- The chip boasts 7.5x the speed of its predecessor, enabling advanced AI capabilities.
- Thor supports generative AI models, including large language and visual models. It features 128GB of memory, crucial for running large AI models.
- Major robotics companies like Boston Dynamics and Amazon are already utilizing the chip.
Nvidia has officially launched its latest high-performance computing module designed specifically for robotics, the Jetson AGX Thor. Available now as a $3,499 developer kit, with shipments commencing next month, this “robot brain” promises a significant leap forward in robotics capabilities.
Nvidia positions Thor as a key component in enabling the next generation of intelligent robots, emphasizing its role in powering advancements across the industry rather than directly manufacturing robots itself.
The Jetson AGX Thor, built upon Nvidia’s Blackwell graphics processor, offers a substantial performance improvement. Nvidia claims it’s 7.5 times faster than its preceding generation, a critical upgrade for running complex generative AI models.
These models, including large language models and sophisticated visual interpretation models, are essential for creating advanced, responsive robots capable of interacting with the real world. The inclusion of 128GB of memory further enhances its ability to handle the demands of these large AI models.
Prominent players in the robotics field are already leveraging the chip’s capabilities. Companies such as Agility Robotics, Amazon, Meta, and Boston Dynamics are currently utilizing Nvidia’s Jetson chips in their projects.
Nvidia’s commitment to the robotics sector is evident through its investments in companies like Field AI and the recent restructuring of its business units, combining automotive and robotics divisions into a single reporting entity. This newly combined unit reported a significant 72% year-over-year increase in quarterly sales, reaching $567 million in May.
While robotics currently accounts for only about 1% of Nvidia’s overall revenue, the company views it as a major growth area, second only to artificial intelligence. This strategic focus is driven by the belief that Nvidia’s technology provides the fundamental infrastructure and software enabling advancements across the entire robotics industry.
The Jetson AGX Thor, with its powerful AI capabilities and large memory capacity, represents a significant step towards realizing this vision. Furthermore, the chip’s adaptability extends to the automotive sector, with particular application in self-driving cars, especially among Chinese manufacturers.
