Manufacturer reference image. Included components and condition follow the selected item.
NETWORK ACCESSORY
NVIDIA Jetson Orin NX 16GB Production Module
- Type
- Jetson Orin NX 16 GB production system-on-module
Before you order
- Confirm the exact SKU, part number, specification option and quantity.
- Confirm the host platform, port, electrical interface and firmware requirements.
- Confirm availability, lead time, destination and applicable service terms before payment.
| CPU | 8-core 64-bit Arm Cortex-A78AE |
|---|---|
| DLA | 2 × NVDLA v2 |
| GPU | NVIDIA Ampere; 1,024 CUDA cores; 32 Tensor Cores |
| Memory | 16 GB, 128-bit LPDDR5 |
| CPU cache | 2 MB L2; 4 MB L3 |
| AI compute | Up to 157 sparse INT8 TOPS in Super mode |
| Part Number | 900-13767-0000-001 |
| Product type | Jetson Orin NX 16 GB production system-on-module |
| PCIe interface | PCIe Gen 4; 1 ×4 plus 3 ×1 configuration listed |
| Camera interface | 8 MIPI CSI-2 lanes; up to 4 cameras or 8 via virtual channels |
| External storage | NVMe supported; external system storage required |
| Memory bandwidth | 102.4 GB/s |
| Module connector | 260-pin SO-DIMM |
| Fixed power modes | 10 W, 15 W, 25 W; 40 W with a supported Super configuration |
| Module dimensions | 69.6 × 45 mm |
| Vision accelerator | 1 × PVA v2 |
| Networking interface | 1 × Gigabit Ethernet capability |
| Maximum CPU frequency | Up to 2 GHz |
| Maximum GPU frequency | 1,173 MHz in the published Super configuration |
| Manufacturer lifecycle | Available through January 2032 |
| Production integration | Carrier, suitable cooling and power, external storage and flashed software required |
| Super-mode integration | Carrier HV rail and thermal design supporting 40 W are required |
| Manufacturer part number | 900-13767-0000-001 |
Product questions & answers
What distinguishes 900-13767-0000-001 from the 8 GB Orin NX?
This exact part is the 16 GB production module. It has eight CPU cores and two NVDLA engines, compared with six CPU cores and one NVDLA engine on the 8 GB model. Choose it when measured memory use, CPU work or supported DLA concurrency justify the additional resources.
Does this module deliver 157 TOPS immediately after installation?
157 TOPS is NVIDIA’s overall sparse INT8 Super-mode figure. Enabling that configuration requires a compatible software release and flash configuration, plus suitable carrier power and cooling. Actual latency depends on model operators, precision, input size and simultaneous tasks.
Which power modes should I design around?
The r39.2.1 guide lists fixed 10 W, 15 W and 25 W modes, plus 40 W for the Super configuration. MAXN_SUPER is an experimental unconstrained mode rather than a guarantee of maximum sustained performance. Include carrier, sensors and storage when sizing the system supply.
Does the 16 GB module include an SSD or a microSD slot?
The 16 GB figure is LPDDR5 memory capacity. NVIDIA lists external NVMe support for Orin NX; the microSD slot shown on an Orin Nano developer-kit module is a different product feature. Plan compatible external storage and flash the production software image.
Can I reuse a Xavier NX carrier because the board is the same size?
NVIDIA describes Orin NX and Xavier NX as form-factor compatible, but not pin-compatible. A carrier can be designed around their shared interfaces; an existing board still needs a pinout, power, lane-mapping and software review. Matching size alone is not an interchangeability guarantee.
Can the 16 GB module run several perception models at once?
It is a reasonable candidate for concurrent pipelines, but memory capacity does not establish a guaranteed model count. Profile model weights, runtime buffers, camera queues and CPU activity together. Measure sustained latency and peak memory in the selected power mode before committing to the final configuration.