Article illustration 1

A One‑Stop AI Edge Platform

On October 23, 2025 Canonical announced that Ubuntu now runs natively on the Thundercomm RUBIK Pi 3, a lightweight developer board built around Qualcomm’s Dragonwing™ QCS6490 processor. The board, designed specifically for AI workloads, packs a 12 TOPS ML accelerator, an 8‑core GPU, 8 GB LPDDR4x RAM, 128 GB UFS 2.2 storage, and integrated Wi‑Fi/Bluetooth—all while drawing less than 6.5 W.

The new Ubuntu image is not just a generic distribution; it’s a tailored experience. It comes pre‑loaded with the Qualcomm AI Hub’s pre‑optimized vision, audio, and NLP models, the Edge Impulse MLOps platform for training and deployment, and the Qualcomm VSCode IDE extensions that simplify device setup. Containers are also fully supported, giving developers immediate access to hardware acceleration for NPUs, GPUs, and VZPUs.

“By delivering an optimized Ubuntu image preloaded on the powerful RUBIK Pi 3, we’re offering an integrated, securely designed, and supported foundation for AI developers,” said Cindy Goldberg, VP of Silicon Alliances at Canonical. “Our partnership with Qualcomm and Thundercomm allows developers to move from a concept to a deployed solution with speed and confidence.”

Why This Matters for Edge AI

Edge AI is a race against latency and power constraints. The RUBIK Pi 3’s low‑power profile means it can run continuously in battery‑powered or remote deployments, while the 12 TOPS accelerator delivers the compute needed for real‑time inference. With Ubuntu’s LTS support, developers gain a stable, long‑term foundation that reduces the risk of supply‑chain vulnerabilities and obsolescence.

From a tooling perspective, the integration of Edge Impulse and the Qualcomm SDKs means that a single board can serve as a prototype, training hub, and production gateway. Developers no longer need to juggle multiple boards or OS images; they can spin up a container, pull a pre‑optimized model from the AI Hub, and deploy it with a single docker run command.

The Bigger Picture

This move echoes a broader industry trend: major open‑source vendors partnering with silicon makers to deliver turnkey AI solutions. By aligning Ubuntu’s software stack with Qualcomm’s hardware, Canonical and Thundercomm are lowering the barrier to entry for startups and research labs that need a reliable, low‑power edge platform.

The RUBIK Pi 3’s feature set—especially its integrated Wi‑Fi/Bluetooth and 8‑core GPU—also positions it as a compelling platform for robotics, autonomous vehicles, and IoT gateways. The board’s support for ROS/ROS2 via the Qualcomm Intelligent Robotics Product SDK further broadens its appeal.

“At Thundercomm, we’re committed to lowering the barriers to AI innovation,” said Ali Mesri, Sr Vice President of Business Development. “With the optimized Ubuntu image on RUBIK Pi 3, developers now have a unified platform that combines Qualcomm’s performance, Canonical’s stability, and Thundercomm’s deep system integration—enabling faster, more reliable AI deployment from concept to production.”

Getting Started

Developers can download the optimized Ubuntu image directly from the Thundercomm website and flash it to the board using standard tools like dd or balenaEtcher. Once booted, the pre‑installed VSCode extensions and container runtimes are ready to use, and the board’s hardware acceleration can be accessed through familiar APIs.

Article illustration 2

For more details, visit Canonical’s blog post on the announcement: https://canonical.com/blog/rubik-pi-3-thundercomm-canonical.