- VENTUNO Q runs fully autonomous AI agents completely offline without external servers
- The Dragonwing processor delivers up to forty dense TOPS of AI compute
- Robotics applications include vision-guided arms and autonomous machines navigating complex environments
Qualcomm and Arduino have launched Arduino VENTUNO Q, a single-board computer designed for robotics, generative AI, and edge computing able to operate fully offline.
The board uses the Qualcomm Dragonwing IQ8 Series processor and a dedicated STM32H5 microcontroller for deterministic control, enabling systems to perceive, decide, and act on the same device.
Qualcomm says the neural processing unit delivers up to 40 dense TOPS of compute, supporting simultaneous inference and complex processing locally, while 16GB of RAM and 64GB expandable storage handle demanding multitasking.
A new board built for edge AI workloads
Fabio Violante, VP and GM at Arduino, said the platform enables “systems that don’t just interpret the world, they interact with it,” allowing machines to operate without reliance on cloud connectivity.
The architecture integrates AI acceleration with real-time microcontroller logic, uniting perception, decision, and actuation on a single board.
VENTUNO Q can run fully autonomous AI agents for tasks including offline voice assistants, gesture-responsive smart mirrors, and interactive kiosks at transport hubs, healthcare desks, or tourist centers.
The system supports robotics applications, including pick-and-place robotic arms guided by vision and service robots capable of following individuals across dynamic spaces.
It also enables autonomous machines to navigate complex environments using Visual SLAM combined with path optimization.
Edge AI vision systems are also possible, allowing proactive security monitoring, traffic observation, and automated quality inspections with local visual language models.
All of these functions are handled on the board, eliminating the need to transmit data to external servers.
VENTUNO Q runs Ubuntu and Debian Linux on its main processor and Arduino Core on Zephyr OS for real-time control.
Arduino App Lab supports Python scripts, Arduino sketches, and ready-to-use AI models, including gesture recognition and object tracking.
It also supports local LLMs powered by Qualcomm AI Hub, while Edge Impulse Studio allows training custom models.
Industrial I/Os, multiple MIPI CSI camera connectors, audio, displays, and 2.5Gb Ethernet provide comprehensive hardware compatibility.
“With VENTUNO Q, AI can finally move from the cloud into the physical world. This platform makes it possible to build machines that perceive, decide, and act, all on a single board,” said Fabio Violante, VP & GM, Arduino, Qualcomm Technologies, Inc.
“Our goal is to make advanced robotics and edge AI accessible to every developer, educator, and innovator. VENTUNO Q is the natural evolution of Arduino’s mission, and a major step toward bringing real world intelligence to everyone.”
VENTUNO Q is compatible with Arduino UNO shields, Modulino nodes, Qwiic sensors, and Raspberry Pi HAT expansions.
It will be available in Q2 2026 through the Arduino Store and other authorized resellers such as DigiKey, Farnell, Mouser Electronics, and RS Components, although the board’s practical influence on existing platforms like the Raspberry Pi remains uncertain.
