NVIDIA Jetson T3000 brings Blackwell AI compute to smaller robots
NVIDIA has announced two smaller Jetson Thor computers for robots, industrial machines and other edge AI systems. The Jetson T3000 and T2000 are scheduled to become available in the first quarter of 2027.
The T3000 uses an NVIDIA Blackwell GPU and an eight-core Arm Neoverse processor. It offers 865 FP4 TFLOPS of AI performance, 32GB of LPDDR5X memory, 273GB/s of bandwidth and 25 GbE connectivity.
A smaller version of Jetson AGX Thor
NVIDIA says the T3000 is roughly half the size and uses about half the power of the existing T5000 module. Despite the reduction, the company claims it can deliver similar inference performance in workloads such as language, vision and robotics models.
Source: NVIDIA
The T2000 is a lower-end option with 400 FP4 TFLOPS and 16GB of memory. NVIDIA expects it to be used in visual AI systems, mobile robots, industrial arms and other machines that do not require the performance of the T3000.
Source: NVIDIA
Developers will be able to emulate the T3000 later this month using JetPack 7.2.1 and the current Jetson AGX Thor developer kit. T2000 emulation will arrive later, while physical T3000 and T2000 modules are planned for Q1 2027.
Source: NVIDIA
