DEEPX produces low power AI accelerators for edge devices. According to Avnet Silica, the new signing “significantly bolsters” its AI offering and addresses growing customer demand for low power, AI for edge devices.
The portfolio includes the DX-M1 AI processor, which delivers 25 TOPS and the DX-V3, a specialised AI vision SoC, performing at 13 TOPS but consuming just 5W, and capable of supporting up to 12 camera streams. According to DEEPX, the semiconductors achieve up to 20 times better performance per Watt than traditional solutions while maintaining GPU-level AI accuracy.
There is also the DX-MI AI accelerator and the DX-HI V-NPU PCIe card.
All are designed for use in embedded applications such as smart city infrastructure, autonomous mobile robots (AMRs), machine vision, smart factories, and autonomous vehicles.
“DEEPX delivers the essential NPU architecture and efficiency required to deploy sophisticated models in power- and thermal-sensitive environments, significantly enhancing the options across our existing linecard,” commented Michaël Uyttersprot, manager System Solutions, AI/ML & Vision System Solutions at Avnet Silica.
In addition to power management, embedded system and industrial IoT expertise, Avnet Silica offers engineering support. Tim Park, strategy marketing director at DEEPX, sees the partnership as “a significant catalyst” to extend the company’s reach across global markets. “Together, our combined supply chain reliability and proven technologies will empower more customers to adopt AI as a practical and scalable infrastructure,” he added.
