
Critical Link announces Best in Show awards for two of the company’s latest SOMs.
Embedded Computing Design presented Best in Show awards for two Critical Link SOMs at Embedded World North America 2025. The MitySOM-QC6490, based on Qualcomm’s Dragonwing QCS6490 processor, and the MitySOM-A5E family of products, with Altera’s Agilex 5 E-Series FPGA.
The MitySOM-QC6490 delivers multi-core CPU, GPU, and VPU processors and a host of memory options. The SOM includes on board power supply, and a range of high-speed interfaces all in a 45mm x 45mm package. Developers can leverage the Qualcomm AI Hub, a toolkit for building speech recognition, image classification, text generation, and many other common AI applications.
Next up is Critical Link’s Agilex 5 FPGA family of products, with three powerful offerings: a full-featured system on module, a smaller SOM to address size or budget constrained applications, and a single board computer for ultimate flexibility and speed. The MitySOM-A5E family offers high performance, lower power, and small form factors at the edge.
In addition to designing world-class products, Critical Link delivers solutions customers can count on for long term availability, quality, and design maintenance. We provide a level of support that is unmatched in the industry. And we have fully US-based engineering and production teams. Learn more about what sets Critical Link apart: https://www.criticallink.com/the-critical-difference/.
Visit www.criticallink.com to explore our available products and engineering services. Reach out to us directly anytime at info@criticallink.com to discuss how we can support your next project.

Add Critical Link booth #3065 to your Embedded World must-see list and check out a live demonstration of AI at the edge using our latest family of Qualcomm-based system on modules!




Meeting and greeting RIT’s best and brightest at the 
We’re excited to get to St Louis this week for another 
Thank you to the 
Join Omar Rahim and Critical Link in Michigan on September 18! Learn about leveraging Asymmetric Multiprocessing and 

