Nvidia’s New Robot Computers Bring Human-Like AI to Industry
349Nvidia has introduced a new device designed to make robots smarter and more capable of acting on their own. These computers, called Jetson Thor, are built to go into high-performance robots that need to understand their environment, make decisions, and act quickly. They’re powered by advanced AI technology that’s already available today, especially generative AI, which helps machines learn and adapt.
Jensen Huang, Nvidia’s CEO, envisions a future where billions of robots work alongside humans, creating huge opportunities for the company’s revenue. To help make that happen, Nvidia revealed the Thor robotics computers, which are meant to be used in robots that can imitate human behavior. These devices are meant to give robots a kind of “brain” that can hear, recognize patterns, and make decisions in real time.
What Makes Nvidia’s Thor Special
The Thor boards are packed with Nvidia’s latest Blackwell graphics processors. These processors are what power the newest AI models, allowing robots to perform complex tasks without needing to connect to the cloud. Normally, AI models run on powerful servers far away, but with Thor, the AI can run directly inside the robot. This means robots can respond instantly, which is crucial when they need to act in fractions of a second.
Deepu Talla, Nvidia’s vice president of robotics and edge AI, explained that doing AI processing locally at the edge reduces delays. If robots rely on the cloud, they could face hundreds of milliseconds of lag, which is too slow for many tasks like manufacturing or heavy industry. Thor’s design focuses on being fast and efficient while consuming less power than data center GPUs.
Balancing Power and Performance
The Thor developer board offers about 2,070 teraflops of performance, which is impressive but still much less than Nvidia’s top data center GPUs. For comparison, Blackwell GPUs can deliver around 20 petaflops, making them thousands of times faster. However, these high-powered GPUs need a lot more electricity—up to 20 times more—making them unsuitable for battery-powered robots.
Robots using Thor are expected to operate in demanding environments like factories, oil rigs, and mines. They need to run on batteries, so power efficiency is key. Thor’s computers draw between 40 and 130 watts, roughly double the power used by typical laptops, which is a challenge but still manageable for industrial robots.
Nvidia doesn’t see robots just as humanoids. Instead, they believe general-purpose robots that can reason and handle different tasks are more important. These robots will be able to learn and adapt to new situations, making them useful across many industries.
AI Advancements with Thor and Beyond
Nvidia has been making waves with AI-powered robots recently. For example, they announced Cosmos Reason, a vision AI model that helps robots better understand their surroundings by analyzing video input. This kind of reasoning allows robots to make smarter decisions, and Thor hardware will be the platform that brings these AI capabilities to life.
Adding reasoning to AI models is a big step forward. It lets robots generalize their skills and tackle new problems that were previously too complex. Talla explained that this approach opens up new possibilities for solving difficult challenges in real-world applications.
The new Nvidia hardware, including the $3,499 AGX Thor system, features 14 ARM CPU cores, 128GB of RAM, and fast network ports. Developers can also build with Nvidia’s T5000 and T4000 boards, which have lower power requirements. The T5000 uses 130 watts, suitable for larger robots with bigger batteries, while the T4000 uses just 70 watts for smaller applications.
As Nvidia continues to push the boundaries of AI robotics, these advancements could reshape industries like manufacturing, energy, and mining. Robots powered by Thor could soon perform complex tasks with human-like reasoning, all while operating efficiently on batteries and in real-time.
In the coming years, expect to see more intelligent robots working alongside humans, driven by Nvidia’s cutting-edge AI hardware and software innovations.
What do you think?
It is nice to know your opinion. Leave a comment.