AI BriefingNVIDIA Launches New Jetson Thor Computers to Advance Robotics and Edge AI
Key point
NVIDIA has launched the Jetson T3000 and T2000 modules based on the Thor architecture, targeting the robotics and edge AI markets.
Details
NVIDIA unveiled new modules based on the Thor architecture, the Jetson T3000 and T2000, to support the mainstream adoption of general-purpose robots and autonomous devices. These modules serve as high-efficiency AI supercomputers capable of running foundation models in edge environments.
The Jetson T3000 reduces size and power consumption by half compared to the existing T5000, while delivering similar inference performance for multimodal workloads such as LLMs and VLMs. Key specifications are as follows.
- Equipped with an NVIDIA Blackwell GPU
- 8-core Neoverse Arm CPU
- 32GB LPDDR5X memory with 273GB/s bandwidth
- 865 FP4 teraflops of AI compute performance
The Jetson T2000 offers 400 FP4 teraflops of compute performance and 16GB of memory, providing an entry-level solution for developers of visual AI agents and autonomous mobile robots (AMRs).
In addition, NVIDIA is transforming developer productivity through Jetson agent skills. This feature automates memory optimization and system configuration, helping developers run high-performance workloads even on lower memory specifications. Through this, companies such as UBTech have achieved significant reductions in memory usage, cutting hardware costs.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.