NVIDIA Jetson T5000 Module – High‑Performance Embedded AI and Robotics Compute Module with 2070 FP4 TFLOPS Blackwell GPU, 14‑Core Arm Neoverse CPU, 128 GB LPDDR5X Memory, Multi‑Sensor Processing Support

The NVIDIA Jetson T5000 Module is a high‑performance embedded AI compute module built for the most demanding robotics, autonomous machines, and physical AI applications. Based on the NVIDIA Jetson Thor platform, this module delivers up to 2070 FP4 TFLOPS of AI compute, enabling sophisticated machine learning inference, real‑time sensor fusion, and autonomous decision‑making at the edge. This level of performance is significantly higher than previous generations, making it ideal for advanced robotics, industrial automation, and multimodal perception tasks. At its core, the Jetson T5000 Module features a 14‑core Arm Neoverse‑V3AE 64‑bit CPU and a powerful NVIDIA Blackwell architecture GPU with 2560 CUDA cores and 96 Tensor Cores, enabling seamless execution of large AI models and intensive parallel workloads. The integrated PVA v3 vision accelerator further enhances sensor processing performance, allowing developers to efficiently handle high‑speed camera and vision streams without bottlenecks. With 128 GB of LPDDR5X memory and a 256‑bit interface providing 273 GB/s of bandwidth, the module supports memory‑intensive applications and large‑scale model inference directly on the edge. Storage is accommodated through NVMe support via PCIe and external SSD options through USB 3.2, giving system designers flexibility in balancing storage capacity and performance. Connectivity options include multiple USB ports, PCIe Gen5 lanes, and support for high‑speed networking such as 4× 25 GbE, making the Jetson T5000 Module suitable for integrating into complex multi‑sensor systems and robotics platforms. Extensive video encode and decode support allows for real‑time streaming of multiple high‑resolution video feeds. Packaged in a compact module form factor with an integrated thermal transfer plate and heatpipe, the Jetson T5000 is designed for custom carrier board integration and industrial hardware builds. Whether for humanoid robotics, autonomous vehicles, industrial robotics, or advanced research projects, this module delivers the compute power and flexibility needed to accelerate innovation at the edge.