Arduino VENTUNO Q

An edge-AI single-board computer that can see, understand, and physically respond to the world around it. Perception, decision-making and actuation all happen on the board, with no cloud dependency and no data leaving the device.

Two processors with different purposes share the work over a high-speed RPC bridge. A Qualcomm Dragonwing IQ8 (IQ-8275) with an 8-core ARM CPU, an Adreno GPU, Spectra 692 ISP, and a Hexagon NPU rated at up to 40 dense TOPS, all on one SoC, runs your neural networks, vision pipelines and language models. Alongside it, an STM32H5F5 Cortex-M33 runs the Arduino core on Zephyr for sub-millisecond, deterministic control of motors, robots and industrial I/O. On-device intelligence and real-time actuation usually take two separate systems; here they share one 160 × 100 mm board.

Everything is developed in Arduino App Lab, where Arduino sketches, Python scripts, Linux applications and AI models sit side by side. Use the board as a standalone computer with a monitor, keyboard and mouse on an Ubuntu or Debian desktop, or connect from your existing laptop over USB-C or Ethernet.

It works with hardware you may already own: Arduino UNO shields on the UNO headers, Raspberry Pi HATs on the standard 40-pin GPIO header, and Qwiic, PiicoDev, STEMMA QT and Modulino I2C modules through the solder-free Qwiic connector. Plan for around 25 W of power draw: through USB-C PD, a barrel jack or the screw terminals.

We recommend a 12 V Power Supply, along with a compatible IEC adapter, or, once set up, a 65+ W USB-C supply and a high-quality USB-C cable.

Features

• Runs language, vision-language and speech models fully offline, with a curated library that includes Qwen LLM and VLM, Whisper transcription and Melo TTS

• Recognises gestures, tracks objects and detects poses on-device using ready-to-run MediaPipe, YOLO-X and PoseNet models

• Gets complex systems prototyped quickly with App Lab Bricks: reusable blocks for IoT endpoints, audio and image recognition, database queries and more

• Keeps actuator control deterministic, with native ROS 2 support and GPIO, PWM and CAN-FD handled by the dedicated real-time microcontroller

• Connects to industrial and vehicle networks through five CAN-FD channels, one with an onboard transceiver on a screw terminal

• Moves data quickly over 2.5 Gb Ethernet, Wi-Fi 6 and Bluetooth 5.3, with onboard antennas

• Handles several cameras and displays at once via MIPI-CSI, USB cameras, HDMI or MIPI-DSI, and DisplayPort Alt Mode over USB-C

• Expands storage beyond the 64 GB eMMC with an M.2 2230 slot for an NVMe Gen 4 SSD

• Supports custom carrier designs, with high-speed JMEDIA, JOMEGA and JMISC headers for cameras, displays, audio and extra USB and CAN

Arduino VENTUNO Q Specifications

Resources

• VENTUNO Q Tutorials (Arduino Docs)

• User Manual

• First Setup Guide

• Official 3D-Printable Enclosure Case (Printables): prints in PLA, PETG or ABS and assembles with 10 × M3 × 6 mm flathead screws