Today is a big day. Pre-orders for VENTUNO Q are open, and we couldn’t be more excited to share this with you. Our most advanced platform to date was built with one clear ambition: to give developers, makers, educators, and engineers a board that doesn’t just perceive the world, but interacts with it. A board where AI takes action.
“For twenty-one years, Arduino has taken complex technology and made it simple enough for anyone to build something real. VENTUNO Q is that same mission, applied to the most consequential technology of our time. With VENTUNO Q, we are giving every developer, educator and innovator the tools to build machines that don’t just think, but do.” – Fabio Violante, VP & GM, Arduino, Qualcomm Technologies, Inc.
What makes VENTUNO Q different
Most edge AI boards stop at perception. VENTUNO Q goes further. Its dual-brain architecture pairs a Qualcomm Dragonwing™ IQ8 processor – delivering up to 40 dense TOPS of AI performance – with a dedicated STM32H5 real-time microcontroller, so the same board that runs a local LLM can also control a motor, communicate via CAN bus protocol, or trigger a physical response in milliseconds. Perception, decision, and action, unified on a single board, entirely offline.
It ships with 16 GB LPDDR5 RAM and 64 GB of eMMC storage – expandable via an M.2 connector for NVMe Gen.4 – giving it the headroom to run concurrent AI inference, complex multitasking, and local data storage without compromise. It comes with pre-loaded Ubuntu, distributed by Canonical, and bundled with Ubuntu Pro license that delivers enterprise-grade security updates. Arduino® App Lab is your fastest path to a working prototype, but VENTUNO Q maintains full compatibility with VS Code, PyCharm, Jupyter, Docker, and native Linux tools. The Arduino Core runs on Zephyr RTOS on the MCU side for deterministic, real-time hardware control. No proprietary toolchain. No lock-in. Just a powerful, open platform you can build on your own terms.
VENTUNO Q is equipped for demanding real-world deployments. On the connectivity side, it features Wi-Fi 6 tri-band (2.4/5/6 GHz) with onboard antenna, Bluetooth 5.3, and 2.5 Gb Ethernet. Video output is available via HDMI, DisplayPort over USB-C Alt mode, and MIPI DSI. For cameras, three MIPI CSI connectors are available alongside USB camera support. Audio I/O, USB-C with host/device and power role switching, two USB 3.0 Type-A ports, and additional USB 3.0 and power options on the JOMEGA header round out the interface set. For industrial applications, VENTUNO Q provides CAN-FD connectivity across multiple headers – including a CAN-FD PHY on a dedicated screw terminal – alongside I2C/I3C, SPI, PWM, and UART.
ROS 2 compatibility is built in, making VENTUNO Q a natural fit for robotics development from day one. On the hardware compatibility side, VENTUNO Q works out of the box with Arduino® UNO™ shields and carriers, Arduino® Modulino™ nodes via the onboard Qwiic connector, and Raspberry Pi HATs. That’s how it ensures your existing ecosystem of sensors, actuators, and accessories carries forward without modification.
Build your way with Arduino® App Lab
Whether you want to get started immediately, bring your own models, or train something custom, VENTUNO Q is the board for you. Out of the box, a growing library of NPU-optimized models is ready to run via Arduino App Lab powered by Qualcomm AI Hub – the current library includes: 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; and MediaPipe gesture recognition. Additional models and bricks are added continuously.
You can also upload a GGUF-format model from Hugging Face directly into Arduino App Lab, or train your own with the integrated Edge Impulse Studio.
Scale from prototype to production
One of the things we’re most excited about is the path VENTUNO Q opens beyond the prototype stage. Under the Works with Arduino™ program, SECO and Toradex – two leading providers of production-grade computing solutions – are among the first partners to offer production-certified SOMs based on the same Dragonwing IQ8 architecture. That means that the apps you prototype on VENTUNO Q with Arduino App Lab, you can scale without re-architecting your hardware stack. Your software, your models, your application logic – all carry forward.
Physical AI is more accessible than ever
Arduino App Lab already features a library of 100+ examples available for VENTUNO Q, with more added regularly to always give you new starting points to explore physical AI out of the box. At launch, you already have everything you need to quickly develop applications with:
- Large language models (LLM)
- Vision-language models (VLM)
- Automatic speech recognition (ASR)
- Text-to-speech (TTS)
- Gesture recognition
And there’s more! Arduino App Lab 0.10.0 brings a wave of new features and improvements, including a new Agentic Mode experience.
Pre-order now to get free accessories included!
Head to the Arduino Store to secure your VENTUNO Q today at its introductory price! Until September 30, 2026 (and while supplies last), you’ll see two freebies added automatically to your cart: the new Arduino® USB-C Power Supply (65W), and the Arduino® USB-C cable (24 pin). These are official accessories that together are designed to support the board in sustaining demanding AI workloads and connected peripherals without throttling.
VENTUNO Q is also available through our official distribution partners: DigiKey, Farnell, Mouser, Robu.in, and RS.
For full technical specs, use cases, and everything you need to know before ordering, read the full press note and visit the VENTUNO Q product page.
We built this for you, and we can’t wait to see what you build with it.
Qualcomm branded products are products of Qualcomm Technologies, Inc. and/or its subsidiaries. Arduino and VENTUNO are trademarks or registered trademarks of Arduino S.r.l.
Facts Only
* The VENTUNO Q runs on the Arduino® VENTUNO™ Q board.
* It utilizes a dual-brain architecture: a Qualcomm Dragonwing™ IQ8 processor and an STM32H5 real-time microcontroller.
* The system provides up to 40 dense TOPS of AI performance from the processor.
* The board includes 16 GB LPDDR5 RAM and 64 GB eMMC storage, with M.2 support for NVMe Gen.4 expansion.
* It ships with pre-loaded Ubuntu and an Ubuntu Pro license.
* The platform supports the Arduino® App Lab, VS Code, PyCharm, Jupyter, Docker, and native Linux tools.
* Connectivity includes Wi-Fi 6 tri-band, Bluetooth 5.3, 2.5 Gb Ethernet, HDMI, DisplayPort over USB-C Alt mode, and MIPI CSI connectors for cameras.
* Industrial connectivity includes CAN-FD across multiple headers alongside I2C/I3C, SPI, PWM, and UART.
* ROS 2 compatibility is built into the system.
* The board is compatible with Arduino® UNO™ shields, Arduino® Modulino™ nodes, and Raspberry Pi HATs.
* Arduino App Lab includes pre-loaded models such as Qwen 3 4B LLM, YoloX small object detection, Whisper ASR, and MediaPipe gesture recognition.
* Pre-orders include free accessories: a 65W USB-C Power Supply and a 24 pin USB-C cable.
Executive Summary
Full Take
The narrative positions Physical AI as the necessary evolution beyond simple perception toward embodied action, leveraging an existing developer ecosystem (Arduino) to integrate cutting-edge AI hardware. The core mechanism is the successful merging of high-performance parallel processing (Dragonwing IQ8) with deterministic, real-time control (STM32H5), which addresses a critical gap in current edge AI systems: the transition from thought to physical execution. This integration strategy—unifying perception, decision, and action on a single board—suggests a pattern of convergence where high-level machine learning models are not treated as purely software abstractions but are directly instantiated as operational control systems.
The emphasis on open standards (Ubuntu, Zephyr RTOS) and broad compatibility with established tools (VS Code, Docker) serves to mitigate vendor lock-in while simultaneously anchoring the technology within a known development context. The pathway from prototype (App Lab) to production (Works with Arduino™ program) suggests an intent to capture value across the entire technology stack, moving beyond niche hobbyist projects toward scalable industrial deployments via partners like SECO and Toradex.
The promotional structure skillfully frames the product not just as a piece of hardware, but as an enabler for building functional physical systems, utilizing excitement over pre-orders and immediate access to diverse AI models (LLMs, VLM, ASR). This creates a strong pull by providing immediate, tangible starting points (100+ App Lab examples) while simultaneously signaling future scalability. The pattern observed here is Authority Game combined with Fear Appeal: leveraging the credibility of Arduino and Qualcomm to validate the claim that this complex integration is achievable now, urging immediate adoption rather than cautious, layered assessment of the real-world computational demands implied by running large models alongside hard real-time tasks offline.
Bridge Questions: If full production scaling relies on external partners like SECO and Toradex, what specific intellectual property or modularity concessions are made regarding the Dragonwing IQ8 architecture to ensure truly independent innovation within the developer community? How does the promised "offline" operation impact safety-critical applications where real-time failure modes must be mathematically guaranteed, rather than merely controlled deterministically? What long-term governance structure is in place for continuously integrating evolving LLM and model formats into the strict constraints of the Zephyr RTOS environment without introducing systemic security vulnerabilities?