Nvidia has introduced a new module designed to enhance the capabilities of humanoid robots and self-driving cars. This advanced robot brain, called Jetson Thor, is engineered to process larger amounts of information more efficiently and with lower energy consumption.
Announced in a press release on Monday, Jetson Thor is a supercomputer tailored for real-time AI computations in both humanoid robots and smart machines. Powered by the latest Blackwell GPUs, it boasts more than seven times the AI compute power and twice the memory of its predecessor, Jetson Orin, achieving over three times the speed and efficiency.
This increased power enables faster sensor data processing and improved visual reasoning, allowing humanoid robots to better perceive their environment, move, and make autonomous decisions. “Jetson Thor addresses one of the biggest challenges in robotics: enabling real-time, intelligent interactions with people and the physical world,” the company stated.
Nvidia anticipates that this significant performance upgrade will attract engineers, with early adopters including Amazon, Meta, Caterpillar, and Agility Robotics, which produces humanoid robots for various industries. The model is also under consideration by John Deere and OpenAI, and research labs at Stanford, Carnegie Mellon, and the University of Zurich are exploring its use in medical research.
The developer kit, Jetson AGX Thor, which includes the Jetson T5000 module, reference carrier board, power supply, and an active heatsink with a fan, is now available for purchase on Nvidia’s website for $3,499.
Additionally, Nvidia has announced the upcoming Nvidia Drive AGX Thor, a developer kit utilizing similar technology for autonomous vehicles, with pre-orders currently available and deliveries expected to start in September.
Nvidia’s Ambitious Robotics Strategy
While AI chips are Nvidia’s core business, the company is making significant investments in robotics and autonomous vehicles. CEO Jensen Huang remarked, “This is going to be the decade of AV [autonomous vehicles], robotics, and autonomous machines.”
At the company’s annual shareholders meeting, Huang expressed confidence in the scalability of the robotics industry. He believes that robotics, combined with AI, represents a “multitrillion-dollar growth opportunity.”
ICYMI: Bawumia Pushes Back Against Ethnic and Religious Attacks on His Candidature
Earlier this year, Nvidia also released a range of AI models called Cosmos designed to train humanoid robots.
Despite advancements, the technology remains limited and has yet to achieve widespread adoption. For example, during a recent robotics showcase in China, many robots faced technical issues, with one even colliding with a bystander.
Upcoming Week for Nvidia
Nvidia’s announcement comes at a crucial time, as the company prepares to report its fiscal second-quarter earnings on Wednesday. The market is abuzz with speculation, especially given the volatile policy landscape and questions surrounding the economic impact of widespread AI adoption.
The company has faced challenges selling AI chips in China amid escalating trade tensions, a significant market for Nvidia. Investors are also concerned about a recent MIT report indicating that fewer than one in ten AI pilot programs have led to actual revenue gains.
Last month, Nvidia reached a market value of $4 trillion, becoming the first public company to achieve this milestone. Now, the firm faces the challenge of demonstrating that its valuation is not merely based on AI hype.
SOURCE: GIZMODO
