Arduino® VENTUNO™ Q — Detailed spec
| Category | Specification | Details |
|---|---|---|
| Product | Model | Arduino® VENTUNO™ Q (ABX00181) |
| Platform | Positioning | Most powerful Arduino platform to date; edge AI computer purpose-built for robotics and industrial edge AI |
| Architecture | Design | Dual-brain: MPU (AI brain) + MCU (action brain) communicating via RPC bridge |
| Processor (MPU) | Chipset | Qualcomm Dragonwing™ IQ8 (IQ-8275) |
| Processor (MPU) | CPU | Octa-core Qualcomm® Kryo™ Gen 6 |
| Processor (MPU) | GPU | Qualcomm® Adreno™ 623 |
| Processor (MPU) | NPU | Qualcomm® Hexagon™ Tensor AI Processor, up to 40 dense TOPS |
| Processor (MPU) | ISP | Qualcomm Spectra 690 ISP (692 per some sources) |
| Microcontroller (MCU) | Chipset | STM32H5F5 |
| Microcontroller (MCU) | Core | Arm® Cortex® M33 at 250 MHz |
| Microcontroller (MCU) | Flash | 4 MB |
| Microcontroller (MCU) | RAM | 1.5 MB |
| Microcontroller (MCU) | RTOS | Arduino Core on Zephyr RTOS; sub-millisecond deterministic control of motors, CAN-FD, PWM, GPIO |
| Memory | RAM | 16 GB LPDDR5 |
| Storage | Internal | 64 GB industrial-grade eMMC |
| Storage | Expansion | M.2 connector for NVMe Gen.4 external storage |
| Operating System | MPU OS | Ubuntu (pre-loaded); Debian coming soon; Docker and apt pre-installed |
| Operating System | MCU OS | Arduino Core on Zephyr RTOS |
| Software | IDE options | Arduino App Lab (pre-loaded); VS Code; your own IDE; native Linux tools |
| Software | App Lab | Unified IDE across Linux and RTOS; Arduino Sketches, Python scripts, containerized AI models from one interface |
| Software | Pre-loaded AI models | Qwen 3 4B LLM; Qwen 2.5 7B and 3 4B VLM; Gemma 4 E2B and E4B; Whisper ASR; Melo and Piper TTS; YoloX small object detection; MediaPipe gesture recognition |
| Software | Bring-your-own-model | Upload GGUF-format models from Hugging Face or Qualcomm AI Hub via llama.cpp Brick in Arduino App Lab |
| Software | Model training | Integrated Edge Impulse Studio; train and quantize custom models for the IQ8 NPU, deploy into App Lab with one click |
| AI | Capabilities | Vision models, LLMs, multi-modal AI inference on-device/offline; object/human detection, anomaly detection, image classification, sound recognition, keyword spotting |
| Connectivity | Wireless | Tri-band Wi-Fi 6; Bluetooth 5.3 |
| Connectivity | Ethernet | 2.5 Gb Ethernet |
| Connectivity | CAN | CAN-FD with real-time control |
| Camera | Inputs | Triple MIPI-CSI camera inputs for 360° vision |
| Robotics | ROS 2 | Compatible |
| Expansion | Compatibility | Arduino UNO Shields; UNO Carriers; Raspberry Pi HATs; QWIIC |
| Production | Path to production | Same Dragonwing IQ8 architecture available as production-certified SOMs from certified partners; software stack, AI models, and app logic carry forward without hardware re-architecture |
| Typical applications | Use cases | Autonomous mobile robots, drones, pick-and-place arms with vision-guided manipulation and Visual SLAM; industrial edge AI for predictive maintenance, process automation, energy monitoring; smart-city traffic monitoring; connected-home responsive environments; automated quality inspection with local VLMs; education and research for computer vision, generative AI, embedded systems |
| Included accessories | Bundle | Free power supply and USB-C cable included with pre-order |
| Warranty | Coverage | Product warranty info available via Arduino's official warranty page |
Takeaway: If you want a single board that can run real AI models (LLMs, vision, speech) offline while still doing precise, low-latency motor and CAN control for robots or industrial gear, VENTUNO Q is the whole setup in one box — no juggling multiple devices. Its 40-TOPS NPU, 16 GB RAM, and Ubuntu stack make it a serious dev platform, and because the same Dragonwing IQ8 silicon is available as production SOMs, what you prototype can go straight to market without a redesign.