（来源：VCA创投社） 随着AI大模型应用需求激增，计算机需要进行大量的矩阵乘法、加法运算，参数量和计算量巨大。当前处理器计算能力快速发展，但数据传输和 存储的速度明显跟不上，制约了整体计算效率的提升。 RRAM（阻变存储器）作为"存算一体"的重要技术路径，能够将存储和计算融为一体。RRAM不需要消耗大量能量来维持存储状态，整体功耗 低，同时RRAM器件可调制到多个不同电阻状态，单个存储单元能存储多位数据，有望进一步提高存储密度。 这一技术特点使RRAM成为突破算力效能边界、引领下一代计算架构变革的关键方向。 铭芯启睿虽然成立于2024年5月27日，成立时间不长，但其背后是中国科学院刘明团队二十多年的技术积累。公司知识产权来源于该团队的200多 项发明专利及芯片设计IP。 AI算力需求正经历爆发式增长，传统计算架构的"内存墙"瓶颈日益凸显。在这一背景下，存算一体技术被视为下一代计算架构变革的关键方向。 2026年1月5日，智能存算芯片研发商铭芯启睿宣布完成超亿元人民币Pre-A轮融资。 本轮融资由国开科创、联想创投领投，中芯聚源、顺禧基金、恒裕投资跟投，老股东中科创星、小米战投持续加码。这是铭芯启睿自2025 ...

Core Viewpoint - The article discusses the innovative RRAM technology developed by "Mingxin Qirui," which integrates storage and computing to significantly enhance AI computing efficiency. The company recently completed nearly 100 million yuan in angel financing, led by Jin Qiu Fund, with participation from major strategic and financial investors like Lenovo Ventures and Xiaomi Investment [1][4]. Group 1: Company Overview - "Mingxin Qirui" was established in May 2024 and focuses on developing new RRAM storage and AI computing technologies to overcome traditional computing architecture limitations [1]. - The company has a strong foundation in intellectual property, with over 200 patents and chip design IP derived from the research team at the Chinese Academy of Sciences, which has over 20 years of systematic research in semiconductor storage [3]. Group 2: Technology Advantages - RRAM technology allows for the integration of storage and computation, addressing the "memory wall" bottleneck in traditional computing architectures, especially in AI applications that require extensive matrix operations [1][2]. - RRAM consumes significantly less energy compared to traditional storage, has the potential for higher storage density due to its ability to modulate multiple resistance states, and is expected to play a crucial role in AI computing [2]. Group 3: Market Position and Partnerships - The company is rapidly advancing its productization and commercialization efforts, having signed contracts worth several million yuan for embedded IP and strategic cooperation agreements for independent RRAM chips [4]. - Major industry players, including TSMC, Samsung, Micron, and SK Hynix, are also exploring RRAM technology, indicating a competitive landscape [3]. Group 4: Investment Insights - Investors view RRAM as a promising technology for AI computing, with Jin Qiu Fund highlighting its low power consumption, fast read/write capabilities, and high density as key advantages [5][6]. - Lenovo Ventures emphasizes the potential of RRAM to significantly reduce AI operational costs and enhance computing efficiency, aligning with the evolving demands of the AI landscape [6].