Arduino's VENTUNO Q Puts a 40-TOPS Edge AI Brain on the Robot Itself

Arduino S.r.l., a subsidiary of Qualcomm Technologies Inc., has opened pre-orders for the VENTUNO Q, an edge AI platform that marks what the company calls its first deliberate move into robotics hardware — despite having more than 20 years of experience working with robotics developers.

"This is the first intentional step for a product that is meant to be used as a single point of control of a modern robotics platform," said Marcello Majonchi, Chief Product Officer of Arduino at Qualcomm.

Two Brains, One Board

The VENTUNO Q uses a dual-brain architecture: a Qualcomm Dragonwing IQ8 Series processor delivering up to 40 dense TOPS of AI performance, paired with a real-time STM32H5 microcontroller. The board carries 16 GB LPDDR5 memory, 64 GB eMMC, and expandable M.2 NVMe storage. It ships with pre-loaded Ubuntu, and developers can work in VS Code, native Linux tools, or their existing workflows. On the microcontroller side, Arduino Core on Zephyr RTOS handles deterministic, real-time control.

The design philosophy is accessibility rather than novelty.

"We have this amazing computing power. We have this ecosystem that is just waiting for something with an easier way for people to be able to translate their vision into reality," Majonchi said.

Why the Edge Matters on the Factory Floor

For manufacturing professionals, the board's offline capability is the headline. Robots deployed in warehouses or offshore manufacturing sites can't count on a reliable high-speed internet connection — and cloud-hosted models don't deliver real-time response.

"One of the blockers of the robotics industry to get into the AI era was the fact that models, until just a couple of years back, were stuck on the cloud," Majonchi said. "Edge computing is the only way to really fulfill many of the constraints of real-world deployments of these robots that might end up in warehouses or manufacturing sites that are offshore."

App Lab: Assembly Over Architecture

Arduino App Lab offers a curated library of NPU-optimized models, and developers can bring their own GGUF-format models. Majonchi compared the workflow to building with bricks.

"You just need to assemble those bricks and recreate connecting code, which is orders of magnitude simpler than what you would do in a normal platform out there," he said.

The intent is abstraction, not reinvention: containers, ROS, and AI models brought into a single approachable framework.

From Prototype to Production

The "Works with Arduino" program targets the prototype-to-production gap. SECO and Toradex are among the first partners supporting the pathway, and the VENTUNO Q is designed for commercial deployment in autonomous mobile robots, predictive maintenance, and real-time quality inspection.

"Arduino is the ultimate prototyping platform," Majonchi said. "At the same time, there are some specificities of a production setup that are slightly different from a prototyping setup."

Meanwhile, Qualcomm is acquiring PickNik Inc., curator of the MoveIt open-source framework, with plans to integrate MoveIt with Arduino VENTUNO Q and Qualcomm Dragonfly — a signal that the path from AI model to motion planning is next on the roadmap.