X Series · system
PX22 Modular AI PC
Triple-layer stacking with GPU, CPU and dual I/O expansion. 128×128×97 mm, ~1.7 kg.
Get a QuoteNVIDIA GTX 1650 / RTX 3050 / 3050 Ti, Intel A380, AMD RX550GPU
128×128×97 mmSize
~1.7 kgWeight
GPU module + up to 2× I/O modules, or dual-CPUExpansion
01
Specifications
| Compute | CPU module | 8th/11th/12th Gen Intel Core, swappable (L1 layer) |
|---|---|---|
| Dual CPU | Dual-system option: L1 compute module + L2 switch module + L3 compute module | |
| Graphics | GPU module | NVIDIA GTX 1650 / RTX 3050 / RTX 3050 Ti, Intel A380, AMD RX550 |
| Multi-display | GTX 1650 (896 CUDA cores, 4GB VRAM) + 11th Gen Core drives 6–8× 4K displays | |
| Expansion | I/O expansion | Up to two I/O module layers (L2–L3) |
| Chassis | Triple-layer stacking, 128×128×97 mm, ~1.7 kg |
02
Variants
| Model | Layer 2 | Layer 3 | Note |
|---|---|---|---|
| PX22_GPU | GPU 模块 | GPU 模块 | 独立显卡性能版，支持 CUDA 训练与推理 |
| PX22_IO | IO 扩展模块 | IO 扩展模块 | 支持两层 IO 扩展 |
| PX22_2CPU | 切换模块 | 计算模块 | 双 CPU 系统 |
03
Compatible Modules
04
Use Cases
- Digital Signage — Drive up to 10 displays with 6× Mini DP I/O modules and discrete GPU options.
- GPU Workstation — Discrete GPU modules (GTX 1650 and up) bring CUDA training and inference to a 1.7 kg chassis.
The X Series stacks three layers: a compute module plus two expansion layers for GPU, I/O or a second CPU system. With a GTX 1650 module (896 CUDA cores, 4GB VRAM) and an 11th Gen Core CPU it drives 6–8 4K displays and supports CUDA for training and inference.