Recently, Anhui Zhiyuan Intelligent Technology Co., Ltd. (hereinafter referred to as: Zhiyuan Intelligent) successfully completed a new round of financing. The investor in this round of financing is Yuanyu Fund.
Zhiyuan Intelligence was founded in December 2025. The company is headquartered in Jiujiang District, Wuhu City, Anhui Province.
It is an exclusive industrial AI equipment diagnosis underlying software service provider for the domestic pan-semiconductor track. It is the first start-up in the industry to launch a large multi-dimensional fault diagnosis model dedicated to semiconductor equipment.
The company has differentiated its entry into the pan-semiconductor manufacturing production line operation and maintenance subdivision track, avoiding the homogeneous competition among major general industrial Internet manufacturers, and focusing on the rigid demand scenario for predictive maintenance of high-value equipment in wafer, packaging, and panel production lines.
The founder of Zhiyuan Intelligence is Jin Weiming. He has been deeply involved in the field of industrial automation for more than ten years and has rich relevant experience.
He controls 5 smart manufacturing-related companies under his name, covering the entire chain of automation engineering, smart equipment, and industrial digitalization.
In the early days, he was deeply involved in the operation of industrial automation integration projects, and successively founded Wuhu Chicheng Intelligent Engineering Co., Ltd., Wuhu Houde Intelligent Technology Co., Ltd., and Wuhu Mingzhi Automation Technology Co., Ltd.
In 2025, Zhiyuan Intelligence was founded, and it reached a cooperation with the laboratory of Shanghai Jiao Tong University on the transformation of technological achievements, and launched a large-scale industrialization project for AI diagnosis of semiconductor equipment.
All products of Zhiyuan Intelligent are developed based on the self-developed industrial AI algorithm base and are divided into three major parts: semiconductor equipment diagnostic product line, pan-semiconductor warehousing intelligent product line, and supporting sensing hardware.
Among them, the Tingzhijipan Semiconductor Equipment Intelligent Diagnosis AI large model platform is the industry’s first vertical industrial large model for semiconductor special equipment training. It is equipped with lightweight edge data acquisition terminals and semiconductor-specific acoustic/vibration sensor hardware packages, and supports two delivery models: local privatized deployment and cloud SaaS.
The warehousing AI agent collaborative scheduling system is developed for wafer factories and semiconductor material warehousing scenarios, integrating machine vision, multi-machine collaborative SLAM navigation, and autonomous task planning technology.
The supporting standardized hardware suite is a self-developed lightweight industrial data acquisition terminal, a semiconductor-specific micro acoustic sensor, and a high-frequency vibration sensing acquisition module.
In addition, Zhiyuan Intelligence also provides customized technical services, including fine-tuning of production line-specific models, digital transformation of old production lines, planning and deployment of sensor points, and customized development project services for digital operation and maintenance platforms.
Zhiyuan Intelligent’s core technologies are all self-developed. Among them, the pan-semiconductor multi-source heterogeneous sensing data fusion large model technology integrates more than ten types of heterogeneous sensing signals such as acoustics, vibration, temperature, voltage, and process parameters. It does not require a large number of manual labeling of fault samples. It uses a semi-supervised training mode to quickly adapt to different types of semiconductor equipment and output integrated results of fault prediction, root cause analysis, and operation and maintenance recommendations.
Supporting subdivided core technology modules include weak fault characteristic signal noise reduction algorithm, semi-supervised industrial time series data training framework, dust-free workshop multi-robot collaborative agent scheduling algorithm and edge-end privatized deployment security architecture.
It is speculated that in 2026, the scale of the domestic pan-semiconductor (wafer manufacturing, packaging and testing, panels, photovoltaic cells) intelligent manufacturing market will exceed 120 billion yuan, and there is broad room for domestic substitution.
With the advantages of self-developed technology and policy support, Zhiyuan Intelligent is bound to quickly become the leading domestic manufacturer in the pan-semiconductor equipment diagnostic AI segment in the future.
