QIHUA X733 ARM Board, 8-Core A76+A55 AI Core Module with 3 Tops NPU, Up to 16GB LPDDR5, 4K Video, Gigabit Ethernet, USB3.0, HDMI, MIPI CSI/DSI, Supports Android 15 & Ubuntu 24

• Quick Start
• Tips

Introduction

Meet the QIHUA X733 – a pocket-size yet powerhouse core module built around Allwinner’s latest 8-K capable SoC. Whether you are crafting an AI vision gateway, a 4-K signage player or an industrial controller, the X733 delivers desktop-class compute, 3-TOPS on-chip AI and cinema-grade multimedia in a 69 × 41 mm SO-DIMM package that draws < 8 W typical.
Out of the box you get mainline Linux 5.15, Android 15 and Ubuntu 24 images, all tested and hosted on QIHUA’s public GitHub. A generous foot-print of LPDDR5, eMMC, USB 3.0, PCIe 3.0, Gigabit Ethernet and dual MIPI lanes means you can connect NVMe SSDs, 4-K HDMI monitors, MIPI-DSI touch screens, GMSL or CSI cameras, RS-485 sensors and more—without glue logic.
Developers keep full control: every GPIO is documented, drivers are open-source, and QIHUA’s FAE team offers 24-hour email support plus OEM/ODM services for volume projects. Replace your ageing A53 boards with a simple plug-and-play upgrade and future-proof your product roadmap today.

Features

1. [8-Core AI Power] Dual Cortex-A76 @ 2 GHz + Hexa Cortex-A55, paired with a 3 TOPS NPU—handles real-time object detection, LLM inference, 4-K video analytics and more without external accelerators.
2. [Up-to 16 GB LPDDR5] Plug-in 69 × 41 mm SO-DIMM core board ships with 4, 8 or 16 GB LPDDR5 (2 000 MHz+) and 64 GB eMMC 5.1, expandable to 256 GB; eliminates memory bottlenecks in edge-AI and multi-display projects.
3. [Multi 4-K Display & Camera] HDMI 2.0b 4-K@60 Hz, dual MIPI-DSI up to 4-K@45 Hz, dual-channel LVDS 1080p, eDP 1.3; drive two screens independently. 4+4+2-lane MIPI-CSI + parallel CSI for 8 MP@30 fps machine-vision cameras.
4. [Rich I/O & Connectivity] Gigabit Ethernet, USB 3.0 + 2.0 OTG, PCIe 3.0, RS-485, 5× UART, SPI, I²C, PWM, ADC, DAC, Wi-Fi 6 + BT 5.2 on board—ready for industrial HMI, robotics, smart kiosks.
5. [Open Source SDKs] Official images for Android 15, Ubuntu 24, Debian 11, Buildroot + Qt5. Full kernel & U-Boot source, schematic, PCB files and step-by-step manuals supplied; customisation service available for volume buyers.
6. [Drop-In Upgrade] Core board is pin-compatible with QIHUA A523/A527/T527/T736 families; swap and scale your current designs in minutes while reusing the same carrier PCB and accessories.

Shipping List

Core board + baseboard + power supply + programming cable + debugging cable + Wi-Fi and Bluetooth antenna

Applications

Edge AI vision, 4K digital signage, industrial HMI, smart retail kiosks, robotics control, smart camera hubs, face-payment terminals, classroom IoT, streaming media boxes.

Specification

• CPU: Dual Cortex-A76 2 GHz + Hexa Cortex-A55 1.8 GHz
• NPU: 3 TOPS, INT8/FP16, supports TensorFlow, ONNX, PyTorch
• RAM: 4 GB / 8 GB / 16 GB LPDDR5 (2 000 MHz+)
• Storage: 16 GB / 32 GB / 64 GB eMMC 5.1 (expandable to 256 GB) + SDIO 3.0
• GPU: PowerVR BXM-4-64, OpenGL ES 3.2, Vulkan 1.3, OpenCL 2.2
• Video Decode: 8-K@24 fps H.265/VP9/AVS2; 4-K@30 fps H.264
• Video Encode: 4-K@30 fps H.264/H.265
• Display: HDMI 2.0b 4-K@60 Hz, 2× MIPI-DSI 4-K@45 Hz, dual LVDS 1080p, eDP 1.3
• Camera: 4+4+2-lane MIPI-CSI + 16-bit parallel CSI, 8 MP@30 fps
• Connectivity: Gigabit Ethernet, Wi-Fi 6 802.11ax, BT 5.2, PCIe 3.0, USB 3.0 & 2.0 OTG
• OS: Android 15, Ubuntu 24, Debian 11, Buildroot + Qt5 (source & image provided)
• Power: 5 V / 12 V DC-in, 8 W typical, PMIC on board
• Dimensions: 69 mm × 41 mm × 4 mm, 260-pin SO-DIMM edge connector
• Temp: Commercial 0 – 70 °C; industrial -20 – 85 °C option on request

Detailed Descriptions