Core Insights - Nvidia has launched its latest robotics chip, Jetson AGX Thor, which is marketed as the "new brain" for robots, featuring a 7.5 times speed increase and equipped with 128GB of memory to run large language and vision models [2][5] - The initial development kits are priced at $3,499 and will begin shipping next month, with bulk purchases of over 1,000 units available at $2,999 each [4] Group 1 - The CEO of Nvidia, Jensen Huang, stated that robotics represents the company's largest growth opportunity outside of artificial intelligence, which has already tripled overall sales in the past two years [5] - Nvidia's Vice President, Deepu Talla, emphasized that the company does not manufacture robots or cars but empowers the entire industry through infrastructure computing and related software [5] - The Jetson Thor chip is based on the Blackwell graphics processor, which is currently used for AI and gaming chips [5] Group 2 - The Jetson Thor chip's capabilities include running generative AI models, which are crucial for humanoid robots, and it supports large AI models due to its 128GB memory [5] - Companies such as Agility Robotics, Amazon, Meta, and Boston Dynamics are already utilizing Nvidia's Jetson chips, and Nvidia has also invested in robotics companies like Field AI [5] - Despite the rapid growth of the robotics business, it still accounts for only about 1% of Nvidia's overall revenue [5] Group 3 - Nvidia has recently merged its automotive and robotics businesses into the same segment, reporting quarterly sales of $567 million in May, a 72% year-over-year increase [6] - The Jetson Thor chip can also be applied to autonomous vehicles, particularly models from Chinese brands, while Nvidia's automotive chips, known as Drive AGX, run on a specialized operating system optimized for automotive use [6]