VeriSilicon(SH:688521) Zhong Zheng Wang·2025-12-14 09:11Core Viewpoint - Chip Origin Co., Ltd. is advancing its strategic acquisitions by investing in TianSui XinYuan and acquiring control of ZhuDian Semiconductor, while terminating the acquisition of XinLai ZhiRong to focus on its core business and RISC-V ecosystem development [1][7]. Group 1: Acquisition and Investment - Chip Origin plans to invest in TianSui XinYuan alongside prominent investors, contributing 350 million yuan to a total capital increase of 940 million yuan, which will make it the largest shareholder with a 40% stake [2][3]. - The acquisition of ZhuDian Semiconductor is aimed at enhancing Chip Origin's competitive edge in AI ASIC markets by combining its image processing technology with ZhuDian's expertise in video processing [4][5]. Group 2: Termination of Acquisition - The company has decided to terminate the acquisition of 97.0070% of XinLai ZhiRong's shares due to misalignment of key demands from the target company's management with market conditions and shareholder interests [7]. - The termination of this acquisition is not expected to adversely affect the company's normal business operations or shareholder interests [7]. Group 3: RISC-V Ecosystem Development - Chip Origin has been actively promoting the RISC-V ecosystem for over seven years, leading initiatives such as the establishment of the China RISC-V Industry Alliance, which has grown to 204 member units [9]. - The company has developed multiple chip design platforms based on RISC-V architecture, which are being adopted by various clients, thus facilitating the industrialization of RISC-V technology [9]. Group 4: Technological Advancements - The company has launched the Coral NPU IP in collaboration with Google, designed for low-power edge applications, which is based on the RISC-V architecture and aims to enhance the edge AI ecosystem [10]. - Chip Origin is developing verification chips based on the Coral NPU IP for applications in AI/AR glasses and smart home devices, accelerating the deployment of large language models at the edge [10].