Workflow
商道创投网·会员动态|芯动力科技·完成近亿元B2轮融资
Sou Hu Cai Jing·2025-08-24 16:33

Group 1 - The core point of the article is that Zhuhai Chip Power Technology Co., Ltd. has recently completed nearly 100 million yuan in Series B financing exclusively invested by Feitu Venture Capital [2] - Chip Power Technology was founded in 2017 and has R&D centers in Zhuhai, Shenzhen, Xi'an, and Silicon Valley. The company has developed a reconfigurable parallel processing architecture (RPP) that combines the versatility of GPUs with the energy efficiency of NPUs, targeting the edge AI inference market with high-performance, low-power chip and accelerator card solutions [3] - The funds from the latest financing will primarily be used for ramping up the mass production of RPP chips, tape-out of the second generation 7nm high-efficiency products, and large-scale deployment for benchmark customers in edge computing scenarios, ensuring rapid implementation of software and hardware platforms across various fields such as security, industrial vision, and robotics [4] Group 2 - The reason for the investment is that the RPP architecture breaks the traditional boundaries between GPUs and NPUs, with a measured energy efficiency ratio leading competitors by over three times. The team has a strong academic background from Tsinghua University and engineering experience from AMD and NVIDIA, with a clear commercialization rhythm and multiple POC orders from leading clients, indicating potential for significant growth [5] - The investment perspective highlights that the Ministry of Industry and Information Technology's "Action Plan for High-Quality Development of Computing Power Infrastructure" has just been launched, and Chip Power is actively responding with its edge AI chips. Feitu Venture Capital is fulfilling its responsibility with industry resources, and the founding team, after seven years of development, balances innovation with pragmatism, warranting long-term attention. The platform also emphasizes the need for rational expansion in the industry to avoid homogenization and internal competition, contributing to the domestic computing power ecosystem [6]