Infineon Technologies AG and embedUR Systems have launched the DEEPCRAFT™ AI Hub, a specialized online platform aimed at fostering intelligent edge development for Infineon’s PSOC™ Edge microcontroller unit (MCU) and neural processing unit (NPU) family. Officially announced on October 30, 2025, this innovative hub, accessible at deepcraft.infineon.com, is crafted to equip developers with pre-optimized AI models, practical implementation resources, and valuable benchmark data.
As demand for AI capabilities at the edge surges, the industry faces a significant skills gap in deploying and optimizing AI applications on resource-conscious embedded devices. The DEEPCRAFT™ AI Hub addresses this challenge by providing a suite of readily available AI models designed for immediate use on Infineon platforms, thus removing the need for extensive deployment expertise.
Key Highlights of the DEEPCRAFT™ AI Hub:
- Pre-optimized AI Models: The hub features a collection of twenty AI models derived from open-source architectures and optimized for the PSOC™ Edge E84 device, supported by the Arm® Ethos-U55 NPU.
- Simplified Deployment: Developers can leverage these models without the complexities traditionally associated with AI implementation, streamlining the process from concept to production.
- Benchmarking and Performance Metrics: Each model has been rigorously tested for both performance and energy efficiency, ensuring reliable execution in real-world applications.
“Infineon is dedicated to enhancing IoT edge devices by merging low-power consumption with cutting-edge AI performance,” stated Sarah Hemmer, Managing Director at Imagimob, an Infineon subsidiary. The hub also includes expert-developed models specific to the PSOC™ Edge platform, such as YOLOv8-Nano, MoveNet Lightning, and MobileNetV2, applicable across various domains from object detection to audio denoising.
John Marconi, VP of Technology at embedUR, emphasized the collaborative effort in simplifying AI deployment. “Our goal with the DEEPCRAFT™ AI Hub is to alleviate integration challenges, enabling developers to concentrate on crafting impactful intelligent products.”
With over two decades of experience in embedded systems, embedUR continues to push the boundaries of Edge AI technologies, delivering solutions that translate silicon capabilities into actionable intelligence. For more information, visit www.embedur.ai and www.modelnova.ai.
This partnership illustrates a significant step towards democratizing AI development at the edge, promising to expedite the evolution of innovative, low-power applications.
