The Jetson TX2 NX is a compact system-on-module (SoM) from NVIDIA — designed for edge AI, embedded computing, robotics, vision systems, and other applications requiring GPU acceleration + low power.
• Part number 900-13636-0010-000 corresponds to the 4-GB version.
• The module aims to strike a balance: better performance than entry-level modules (like TX2 Nano / older boards) while keeping footprint, power draw, and cost low.
⸻
Key Specifications & 6-core ARM 64-bit CPU complex: dual-core NVIDIA Denver 2 + quad-core ARM Cortex-A57
CPU Clock Speed
Up to ≈ 2.0 GHz (depending on power mode / configuration)
GPU
NVIDIA Pascal architecture — 256 CUDA cores
AI / GPU Performance
~ 1.33 TFLOPS (FP16, as often rated)
Memory
4 GB LPDDR4 (128-bit) — ~51.2 GB/s bandwidth
Storage
16 GB eMMC 5.1 onboard
Power Envelope
Configurable — roughly 7.5 W to 15 W depending on workload and power mode
I/O & Expansion
- PCIe Gen2: supports 1× x1 + 1× x2 (total throughput ~30 GT/s) - Camera input: up to 5 cameras (12 via virtual channels) over 12-lane MIPI CSI-2 (D-PHY) interface - Video output / display support: capable of video encode/decode, display outputs (depending on carrier board) - Networking: Gigabit Ethernet support (10/100/1000 BASE-T) via onboard MAC + external PHY interface
The TX2 NX is well-suited for:
• Embedded AI applications at the edge — computer vision, object detection, neural-net inference, sensor fusion
• Robotics, drones, autonomous machines where compute + GPU power is needed but size/power are constrained
• Smart cameras / surveillance devices, where you need onboard processing rather than relying on cloud
• IoT + industrial automation: real-time processing, data collection + inference + control
• Prototyping or products needing cost-effective AI/GPU acceleration in a small, efficient form-factor
In short: edge AI + vision + embedded compute, where power draw and form factor matter as much as performance.
• TX2 NX gives you a compact, powerful enough AI platform for edge compute + vision + control logic — great if you want on-device processing without big power or space overhead.
• It’s a balanced choice: good compute + GPU + I/O => you get flexibility.
• For niche or custom systems, this module is a smart base, especially if you design around a custom carrier board that matches your connectivity, sensor and power requirements.
Condition: 100% Original
Package: 100% Original Package
Orders can be shipped from one or more of our locations overseas depending on availability.
Import duties (custom fee), taxes and charges if applicable, are not included in item price or shipping charges. The buyer should be responsible for these charges.
Damaged Box or Damaged Product:If you receive damaged or incorrect item(s), contact usno later than 5 business days of receipt of your shipment to make the necessary corrections. If there is extensive damage to the box, refuse delivery and contact a Customer Support immediately
Refusal of Delivery:If a package is refused for delivery, outside of arriving damaged or in a damaged box, the customer will be responsible for actual shipping charges incurred by Yudr3m. Once we receive the package back in our warehouse, we will issue a credit to the original form of payment for the purchase price less the actual shipping cost of shipping the package to you.
Incorrect Shipping Address Provided:Packages that are returned because of an incorrect address provided by the customer will be subject to additional shipping charge. This shipping charge will apply even if you were not originally charged shipping under a free shipping promotion. Requests for packages to be reshipped will be subject to a reshipping fee. If the package is returned and customer has not contacted a customer support the order will then be refunded back to the original form of payment.
Limitation of Liability:The Client acknowledges and agrees that Yudr3m will not be liable for any losses or damages, whether indirect, incidental, special or consequential, in profits, goods or services, irrespective of whether or not the Client has been advised or otherwise might have anticipated the possibility of such loss or damage the buyer should be responsible for these charges.
