Yuanyong Technology develops 3D-stacked edge-inference processors built around its OmniFold architecture and a distributed vertical-dataflow NPU. Its first chip is moving toward tape-out for AI devices, industrial control and other local-inference systems, with embodied intelligence identified as a target market.
About
Yuanyong Technology is developing a dedicated edge-inference processor using its OmniFold three-dimensional stacking architecture. A distributed vertical-dataflow NPU is designed to combine compute, memory capacity and bandwidth within a locally deployable chip, with support for more than 500 neural-network inference operators.
The company was formed in Suzhou in September 2025 by a team with experience in chip architecture, AI algorithms and semiconductor supply chains. Its first processor is approaching tape-out after joint development with an industrial partner and is aimed at AI-agent devices, industrial controls and other systems that require low-latency local model execution.