NVX-N6 Vision — STM32N6 AI Development Board

The NVX-N6 Vision is our flagship AI development board — the most capable STM32 platform we build. Built around the STM32N6, it combines a Cortex-M55 core running at 800MHz with a dedicated hardware Neural Processing Unit (NPU) capable of delivering up to 600 GOPS at INT8.

This is not a software AI accelerator. This is a hardware NPU. That distinction matters: where a standard Cortex-M7 might take 500ms to run a YOLOv8n inference, the STM32N6 NPU does the same in under 40ms.

KEY SPECIFICATIONS:
– MCU: STM32N6 — Cortex-M55 at 800MHz + Hardware NPU
– NPU Performance: Up to 600 GOPS (INT8)
– AI Frameworks: STM32Cube.AI, TensorFlow Lite Micro, ONNX
– Internal SRAM: Up to 4MB
– Internal Flash: Up to 4MB
– External RAM: LPDDR4 / PSRAM via XSPI interface
– External Flash: Up to 512MB OctoSPI NOR Flash
– Display: MIPI-DSI output + LTDC + ChromART hardware graphics accelerator
– Camera: DCMI parallel + MIPI-CSI2 camera input
– USB: USB HS 2.0
– Connectivity: SPI, I2C, UART, FDCAN, Ethernet
– Debug: SWD/JTAG header

AI USE CASES:
– Object detection (YOLOv8n at 25fps on-device)
– Image classification (MobileNetV2, EfficientNet)
– Face detection and recognition
– Gesture recognition
– Anomaly detection
– Keyword spotting

IN THE BOX:
– NVX-N6 Vision board (hand-assembled and individually tested)
– Pin headers
– Pinout diagram PDF download
– Quick start card with AI deployment guide

DOCUMENTATION:
Full pinout, schematic, STM32Cube.AI deployment guide, and example AI models at github.com/nvixeon

This board is designed and built in Ghana and ships worldwide via tracked courier.