中国芯片技术近期取得多项突破性进展，涵盖设计、制造、材料及架构创新等领域，标志着国产半导体 产业链从"单点突破"迈向"体系化创新"。以下是核心进展分类梳理： 一、颠覆性算力芯片：突破物理瓶颈 高精度模拟计算芯片（北京大学） 全球首款24位精度模拟矩阵芯片：基于阻变存储器（RRAM），通过动态误差校准算法将传统模拟计算 精度从8位提升至24位，误差率低于0.1%。 性能表现： 求解128×128矩阵方程时，计算吞吐量达顶级GPU的1000倍以上，能效提升超100倍。 应用于6G通信基站信号处理，仅需3次迭代即可恢复高清图像，误码率与32位数字计算相当。 战略意义：攻克模拟计算"精度低、难扩展"的百年难题，为AI大模型训练、边缘计算提供新路径。 存算一体芯片（清华大学） 全球首颗集成存储、计算与片上学习的忆阻器芯片，能效较传统ASIC提升75倍，突破冯诺依曼架构瓶 颈。 应用潜力：支持硬件端直接训练AI，降低云端依赖。 0:00 ⚙️ 二、核心工艺与材料：打破垄断 光刻与刻蚀技术 1nm离子束刻蚀机：国光量超发布4英寸离子束刻蚀机，精度达0.02纳米，性能较国际主流2nm设备提升 百倍；中微半导体实现1纳米等离子 ...

为突破现有瓶颈，唐建石团队将芯片算力拆解为"晶体管集成密度 × 芯片面积 × 单个晶体管算力" 三个 核心要素，针对每个要素展开技术探索。 来源：半导体产业纵横 2025 年9月24日，清华大学集成电路学院副院长、长聘副教授唐建石在2025 IC WORLD 高峰论坛上， 发表题为《高算力芯片发展路径探索与存算一体芯片》的演讲。演讲围绕学院近年在高算力芯片与存算 一体芯片领域的思考、探索及实践展开，系统阐述了行业现状、技术突破与未来规划。 从他的演讲中，我们获取了以下关键信息： 唐建石指出，当前人工智能领域对算力的需求呈爆发式增长，国家计算力指数与数字经济、GDP 增长 紧密相关。其中，中国智能算力规模 2025 年已突破数十万亿亿次，且 AI 算力需求每不到六个月便实 现翻倍，这一增速远超摩尔定律驱动的硬件算力提升速度，构建更强力的芯片算力底座成为行业迫切需 求。 同时，计算芯片与存储芯片存在显著差异：存储芯片拥有统一的标准接口与定义，而计算芯片需依赖指 令集、工具链、操作系统构成的完整生态支撑。从行业格局看，美国长期主导计算芯片体系，我国则面 临双重硬件制约：一是摩尔定律逐步放缓，晶体管尺寸微缩难度加大， ...

Core Insights - The article discusses how venture capital firms can seize investment opportunities and create long-term value through empowerment in the rapidly changing technology landscape, particularly focusing on AI [2] Investment Focus on AI - Xunfei Venture Capital, established in 2016, specializes in investments in the technology sector, particularly artificial intelligence [4] - The investment strategy emphasizes core capabilities and future-oriented investments, focusing on the main AI track and various application layers such as AI + new hardware, AI + life sciences, and AI + energy revolution [4][7] Project Selection Methodology - The team employs a funnel model for project selection, which includes having excellent traffic entry points, clear algorithms to filter out unqualified projects, and providing genuine empowerment [5] - Each stage of the funnel model increases the success probability by 20%-30%, enhancing the likelihood of successful investments [5] Unique Challenges in Hard Technology - The article highlights that the hard technology era differs from the internet era, requiring distinct methodologies and valuation models [6] - Investments in hard technology projects often involve a lengthy process from laboratory to market, necessitating a tailored approach for each industry [6] Dual Empowerment and Ecosystem Building - Xunfei Venture Capital's second core investment logic is "dual empowerment and ecosystem building," aiming for mutual benefits between the firm and its portfolio companies [8] - The firm has achieved business synergy with over 70% of its invested projects, focusing on collaboration across technology, business, and branding to support project growth [8] Investment Philosophy - The investment philosophy combines industry, technology, and capital to aid in constructing an AI ecosystem [9] - The goal is to create a "carrier fleet" in the AI field, enabling collaborative exploration and advancement in the AI wave [9]

Core Insights - The article discusses how venture capital institutions can seize investment opportunities amidst rapid technological changes and a new investment landscape, with insights from Xu Jingming, co-founder of iFlytek and chairman of iFlytek Venture Capital [1] Investment Focus on AI - iFlytek Venture Capital, established in 2016, focuses on investments in the technology sector, particularly artificial intelligence (AI) [2] - The investment strategy emphasizes a "funnel model" for project selection, which includes having excellent traffic entry points, clear algorithms to filter out unqualified projects, and providing genuine empowerment [2][3] - The team aims to invest in a "1+3" strategy, focusing on the main AI track and three "AI+" areas: AI + new hardware, AI + life sciences, and AI + energy revolution [3] Industry Collaboration - iFlytek Venture Capital has achieved industry collaboration with over 70% of its invested projects, aiming for mutual empowerment through technology, business, and brand synergies [4][5] - The company emphasizes a dual empowerment approach, where both the venture capital firm and the invested companies benefit from collaboration [5] Ecosystem Building - The investment philosophy of iFlytek Venture Capital is centered on "industry + technology + capital," aiming to build a robust AI ecosystem [6] - The goal is to create a "carrier fleet" in the AI field, exploring and advancing together in the wave of artificial intelligence [6]

Core Viewpoint - The article discusses the challenges faced by traditional chip architectures due to the rise of generative AI models and the emergence of in-memory computing technology, which significantly enhances AI computing efficiency and is seen as a disruptive technology in the post-Moore era [1][3]. Group 1: In-Memory Computing Technology - In-memory computing technology has gained traction as it addresses the "storage wall" and "power wall" issues inherent in the von Neumann architecture, leading to a potential efficiency improvement of several times in AI computing [1][3]. - The in-memory computing chips developed by Zhichun Technology have already served over 30 clients in commercial applications, showcasing the technology's practical viability [5]. Group 2: Talent Acquisition and Development - Zhichun Technology has launched the "Genius Doctor Program" for 2026, aiming to attract top talent in semiconductor devices, circuit design, and AI algorithms, reflecting the industry's talent competition amid rapid technological advancements [1][7]. - The program offers a unique growth system that includes mentorship and rotation across core R&D positions, allowing participants to gain comprehensive experience in the technology development process [7][10]. Group 3: Industry Trends and Future Outlook - The semiconductor industry is expected to face a talent shortage of over 300,000 professionals by 2025, highlighting the urgency for companies to develop and attract skilled individuals [1]. - The current phase of in-memory computing technology is critical as it transitions from "production validation" to "scale application," indicating a pivotal moment for the industry [12].

Core Concept - The article discusses the concept of "Compute In Memory" (CIM), which integrates storage and computation to enhance data processing efficiency and reduce energy consumption [1][20]. Group 1: Background and Need for CIM - Traditional computing architecture, known as the von Neumann architecture, separates storage and computation, leading to inefficiencies as data transfer speeds cannot keep up with processing speeds [2][10]. - The explosion of data in the internet era and the rise of AI have highlighted the limitations of this architecture, resulting in the emergence of the "memory wall" and "power wall" challenges [11][12]. - The "memory wall" refers to the inadequate data transfer speeds between storage and processors, while the "power wall" indicates high energy consumption during data transfer [13][16]. Group 2: Development of CIM - Research on CIM dates back to 1969, but significant advancements have only occurred in the 21st century due to improvements in chip and semiconductor technologies [23][26]. - Notable developments include the use of memristors for logic functions and the construction of CIM architectures for deep learning, which can achieve significant reductions in power consumption and increases in speed [27][28]. - The recent surge in AI demands has accelerated the development of CIM technologies, with numerous startups entering the field alongside established chip manufacturers [30][31]. Group 3: Technical Classification of CIM - CIM is categorized into three types based on the proximity of storage and computation: Processing Near Memory (PNM), Processing In Memory (PIM), and Computing In Memory (CIM) [34][35]. - PNM involves integrating storage and computation units to enhance data transfer efficiency, while PIM integrates computation capabilities directly into memory chips [36][40]. - CIM represents the true integration of storage and computation, eliminating the distinction between the two and allowing for efficient data processing directly within storage units [43][46]. Group 4: Applications of CIM - CIM is particularly suited for AI-related computations, including natural language processing and intelligent decision-making, where efficiency and energy consumption are critical [61][62]. - It also has potential applications in AIoT products and high-performance cloud computing scenarios, where traditional architectures struggle to meet diverse computational needs [63][66]. Group 5: Market Potential and Challenges - The global CIM technology market is projected to reach $30.63 billion by 2029, with a compound annual growth rate (CAGR) of 154.7% [79]. - Despite its potential, CIM faces technical challenges related to semiconductor processes and the establishment of a supportive ecosystem for design and testing tools [70][72]. - Market challenges include competition with traditional architectures and the need for cost-effective solutions that meet user demands [74][76].

Group 1 - The 2025 Yangpu Global Promotion Conference showcased cutting-edge technology products, attracting significant attention from attendees [1] - Featured products included a vertical take-off and landing fixed-wing drone with a wingspan of 6.3 meters, capable of carrying 30 kilograms, with a flight range of 700 kilometers and a 6-hour endurance [1] - Other innovations included a waist exoskeleton robot designed to prevent worker injuries, an AI product combining hearing aids and AR glasses for the hearing impaired, and various advanced robotics and chips from local enterprises [1] Group 2 - Yangpu District is supportive of technology companies, providing substantial assistance and promotion for enterprises [2] - The conference introduced 14 riverside plots covering approximately 34 hectares, along with two premium plots in the Wujiaochang sub-center, totaling around 9 hectares, inviting global investors [2] - The district launched InnoMatch, a global innovation element supply chain platform, in collaboration with the National Technology Transfer Eastern Center [2]

Core Viewpoint - The "5th China Integrated Circuit Design Innovation Conference and IC Application Ecological Exhibition" (ICDIA) aims to promote breakthroughs in chip technology, showcase China's IC innovation achievements, and foster a self-controlled industrial ecosystem for large-scale applications in AI, new energy vehicles, IoT, and digital economy [1] Group 1: Event Overview - The conference will be held from July 11-12, 2025, at the Suzhou Jinji Lake International Conference Center [1] - The theme is "Independent Innovation • Application Landing • Ecological Co-construction," focusing on AI computing power, photonic integrated circuits, heterogeneous computing, RISC-V ecosystem, 5G/6G semiconductors, AIoT, and smart vehicles [1] - The event will feature a "1+1+4+1" model, including one summit forum, one AI developer conference, four sub-forums, and one IC application ecological exhibition [1] Group 2: Agenda Highlights - The agenda includes a welcome dinner and the 2025 China Strong Chip IC Awards on July 11, followed by various thematic sessions on advanced design, automotive chips, AIoT, and industry-academia cooperation on July 12 [2] - The summit forum will focus on cutting-edge technology breakthroughs, advanced design tools, application scenarios, and industrialization [3] Group 3: AI Developer Conference - The AI Developer Conference will invite influential experts to discuss AI high-performance chips, AI models, and applications, covering deep learning, big data processing, and innovative trends in AI [6][8] - Topics will include AI-driven chip design, emerging computing paradigms, high-end EDA tools, and low-power, high-reliability designs [6] Group 4: Exhibition Areas - The exhibition will showcase China's IC innovation achievements, AI technologies, and intelligent ecological application scenarios across four main areas: advanced design, design innovation alliance, applications and intelligent ecology, and AI and robotics [9][11][12][13][14] - The advanced design area will feature EDA tools, RISC-V ecosystem, and various types of chips including 5G/6G communication chips and AI acceleration chips [11] - The AI and robotics area will include humanoid robots, industrial robots, and AI interaction experiences [14] Group 5: Strong Chip Evaluation - The "Strong Chip Evaluation" aims to identify and promote leading Chinese chip products, providing references for system integrators and end-users, thereby fostering a self-sustaining industrial ecosystem [17]