| Category | Specification |
|---|
| Product Name | NVIDIA Jetson AGX Orin 32GB Module (900‑13701‑0040‑000) |
| AI Performance | Up to 200 TOPS (INT8) |
| GPU | NVIDIA Ampere architecture with 1792 CUDA cores & 56 Tensor Cores |
| GPU Max Frequency | ~930 MHz – 939 MHz |
| CPU | 8‑core Arm® Cortex‑A78AE v8.2 64‑bit CPU |
| CPU Max Frequency | Up to 2.2 GHz |
| Deep Learning Accelerators | 2× NVDLA v2.0 |
| Vision Accelerator | PVA v2.0 |
| Memory | 32 GB 256‑bit LPDDR5 (~204.8 GB/s) |
| On‑Module Storage | 64 GB eMMC 5.1 |
| Video Encode | Up to 1× 4K60, 3× 4K30, 6× 1080p60, 12× 1080p30 |
| Video Decode | Up to 1× 8K30, 2× 4K60, 4× 4K30, 9× 1080p60, 18× 1080p30 |
| Camera Support | Up to 6 cameras via 16 MIPI CSI‑2 lanes |
| PCIe | Multiple PCIe Gen4 lanes (e.g., up to 2× x8 + 1× x4 + 2× x1) |
| USB | USB 3.2 Gen2 (3×) + USB 2.0 (4×) |
| Networking | 1× GbE + 1× 10GbE (module‑carrier dependent) |
| Display Output | 1× 8K60 multi‑mode DP 1.4a / eDP 1.4a / HDMI 2.1 |
| Other I/O | UART, SPI, I2C, CAN, GPIO, audio interfaces |
| Power Consumption | Configurable 15 W – 40 W |
| Form Factor | 100 × 87 mm, 699‑pin Molex Mirror Mezz connector |
| Thermal Solution | Integrated thermal transfer plate |
| Typical Applications | Edge AI, robotics, autonomous machines, industrial AI |
Product Description
The NVIDIA Jetson AGX Orin 32GB Module (900‑13701‑0040‑000) is a flagship AI compute module in NVIDIA’s embedded Jetson family, delivering up to 200 TOPS of AI performance while offering flexible power scaling between 15 W and 40 W. Built on the Ampere architecture, it includes a 1792‑core GPU with 56 Tensor Cores and an 8‑core Arm® Cortex‑A78AE CPU, providing exceptional processing power for deep learning inference, robotics perception, autonomous navigation, and complex multi‑sensor fusion tasks at the edge.
The module integrates 32 GB of high‑speed LPDDR5 memory and 64 GB of onboard eMMC storage, enabling fast data access and smooth handling of large models and datasets. It also features dedicated deep learning (NVDLA) and vision (PVA) accelerators to efficiently offload AI workloads. Video encode and decode engines support multiple high‑resolution streams, and extensive camera interfaces via MIPI CSI‑2 lanes allow sophisticated vision systems to be deployed.
With high‑speed PCIe Gen4 lanes, USB 3.2 connectivity, and multi‑gigabit Ethernet support, this module is well suited for embedded systems requiring robust IO and networking. Its compact 100 × 87 mm form factor and integrated thermal transfer plate simplify integration into carrier boards or industrial AI computers for applications such as autonomous robots, industrial automation, smart cities, and vehicle‑to‑everything (V2X) systems.
-750x250.png "Datacom Technologies (BD)"){kind=logo}
