Foundries.io is showcasing its Qualcomm Dragonwings demo at Embedded World 2025, highlighting advancements in open-source development, IoT device management, and AI workloads. Dragonwings represents Qualcomm’s latest branding for its IoT, automotive, and industrial processors, aiming to expand beyond mobile markets. One of the standout demonstrations features the HP OmniBook X, powered by the Snapdragon X Elite processor, running a locally trained large language model. This model, based on Foundries.io’s documentation, operates entirely on the device’s neural processing unit (NPU), showcasing Qualcomm’s edge AI capabilities. More importantly, Foundries.io remains at the core of device management, enabling secure updates and AI deployment across Qualcomm’s platforms. Learn more at https://foundries.io.
—
Synaptics is my Embedded World 2025 video sponsor, check out my Synaptics videos here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvhAbQoe9YN4c84SqXxIY3fQ
—
A major breakthrough at Embedded World 2025 is Qualcomm’s commitment to open development. The company is now upstreaming its software directly on GitHub, accepting third-party contributions for the first time. This includes the RB3 Gen 2 platform running the 416 RC5 kernel without proprietary patches, making it a fully open-source Linux system. The platform supports Debian and Yocto builds, with source code available for developers to modify and contribute. Qualcomm is also ensuring broad support for its upcoming processors, reinforcing its commitment to the open-source ecosystem.
Another key demonstration focuses on Foundries.io’s integration with Qualcomm’s AI ecosystem, specifically with Edge Impulse, which Qualcomm recently acquired. Using the Foundries.io platform, developers can seamlessly create, train, and deploy AI models for edge applications. The RB3 Gen 2 board, running Qualcomm Linux Release 1.3, features an Edge Impulse IMDK AI app that processes high-speed camera feeds and applies AI-driven object detection. This marks a shift towards democratizing AI deployment on embedded systems, reducing barriers to entry for developers.
The Snapdragon X Elite processor is gaining traction within the Linux community, with Canonical leading efforts to ensure robust Ubuntu support. Multiple X Elite-powered laptops, including models from Lenovo, HP, and Microsoft, are receiving dedicated optimizations. The installation process includes GRUB bootloader support, allowing dual-boot configurations with Windows. Currently, Wi-Fi, touchscreen, and graphics acceleration work smoothly, with upcoming patches expected to enable audio and camera support. This is a crucial step toward making ARM-based laptops competitive with traditional x86 systems.
Dragonwings is also pushing the boundaries of edge AI processing with the IQ9 series, particularly the 9075 model, which features dual NPUs for enhanced inference capabilities. These processors cater to demanding AI workloads, including real-time video analysis for industrial and security applications. Qualcomm’s focus on upstreaming its AI frameworks ensures that developers can build robust solutions without relying on vendor-specific toolchains. The goal is to provide long-term software support while maximizing hardware capabilities.
Qualcomm’s deep involvement in AI is evident through its recent optimizations for deep learning models on Snapdragon hardware. While some aspects remain proprietary, Snapdragon X Elite laptops have demonstrated exceptional AI performance, making them viable alternatives to traditional AI accelerators. Edge Impulse’s integration into Foundries.io’s platform further simplifies AI deployment, enabling seamless model quantization and optimization for embedded environments. This move is expected to accelerate AI adoption across various industries.
Developers interested in robotics will find Qualcomm’s Dragonwings ecosystem particularly appealing. The RB3 Gen 2 platform is designed for autonomous applications, including humanoid robots and edge AI processing. Qualcomm also offers a fully open-source ROS stack, available on GitHub, to streamline robotics development. The company aims to make these developer-friendly boards more accessible, addressing past availability challenges. This openness could lead to a surge in robotic applications powered by Qualcomm hardware.
Foundries.io’s growth within Qualcomm highlights the increasing importance of open-source solutions for secure device management. Originally a Linaro spinoff, Foundries.io continues to support open-source initiatives while expanding its role within Qualcomm. The company’s focus on upstreaming software ensures long-term support without reliance on vendor-specific BSPs. By aligning with Qualcomm’s open-source strategy, Foundries.io is positioning itself as a key player in secure edge computing and embedded Linux development.
all my #ew25 videos https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga
