NVIDIA has released the next generation of its Jetson system, which is powered by the Blackwell architecture and is claimed to bring the next era of humanoid robots.
NVIDIA's Newest Jetson Thor Platform Brings In Massive Upgrades In Compute Capabilities, But With a Hefty Cost
Team Green's CEO Jensen Huang has specifically expressed 'Physical AI' as the next big thing for the company, not only in revenue prospects, but also in how it will enable robots to become much more capable in independent reasoning, driven by AI. According to a new blog post, NVIDIA has launched the Jetson Thor modules, which bring in much higher compute capabilities than Jetson Orin, and more importantly, have opened up the way for humanoid robots.
Diving a bit into the technical details, the Jetson Thor is equipped with the new Blackwell-based Jetson T5000 module, offering 14 ARM Neoverse-V3AE CPU cores and a 2560-core GPU. Regarding generational improvements in performance, NVIDIA claims that the latest platform brings in 7.5x more AI compute, 3.1x more CPU performance, and 2x more memory, allowing it to run genAI frameworks and reasoning models. Ultimately, the intention here is to give robots a capable brain to serve humans.
The Jetson Thor system delivers up to 2,070 FP4 teraflops of AI compute, all within a 130W power usage, allowing for more widespread adoption across humanoid robot projects. Here's a quick rundown on the specifications:
| Category | Specification |
|---|---|
| Module | Jetson T5000 |
| GPU | 2560-core NVIDIA Blackwell architecture GPU with 96 fifth-gen Tensor Cores |
| CPU | 14-core Arm Neoverse V3AE 64-bit CPU |
| Memory | 128 GB 256-bit LPDDR5X 273 GB/s |
| Networking | 1x 5GbE RJ45 connector, 1x QSFP28 (4x 25 GbE) |
| I/O | QSFP connector, HDMI port, DisplayPort, 2x USB-A 3.2, 2x USB-C 3.1, Gigabit Ethernet, 2x 13-pin CAN header, Microfit power jack |
| Storage | 1 TB NVMe M.2 Key M slot |
| Power | 40 W – 130 W |
NVIDIA has disclosed that Agility Robotics and Boston Dynamics, two of the most reputable names in robotics, have started deploying Jetson Thor systems in their products. This means that we might see capable robots from them soon. Now, with such computing power onboard, there's indeed a tradeoff in the form of the cost to acquire such systems. NVIDIA has priced the Jetson AGX Thor developer kit at $3,499, which is a hefty figure when you consider the overall costs in building humanoid robots.
Interestingly, Jensen Huang personally signed many of the initial units of Jetson Thor sent to NVIDIA customers, showing his commitment to this particular segment of the AI industry. More importantly, NVIDIA's key partner Foxconn is set to introduce humanoid robots in the market by year-end, which shows that physical AI could pan out to be the next big thing after AI training and inferencing.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
