据中证报报道，近日，北京大学电子学院科研团队创造性地制备了迄今尺寸最小、功耗最低的铁电晶体 管，有望为AI芯片算力和能效提升提供核心器件支撑。团队将铁电晶体管的物理栅长缩减到了1纳米极 限，这一技术打破了传统铁电晶体管的物理限制，使得能耗比国际最好水平整整降低了一个数量级。 中证报表示，该晶体管的"存算一体"特性，使其能够同时完成数据存储与计算，这与AI芯片的进化方向 高度契合。它有望彻底打破传统计算架构中"存储"与"计算"分离导致的效率瓶颈，为发展高算力、低能 耗的下一代人工智能芯片奠定关键技术基础。有着超低工作电压与极低能耗特性的纳米栅铁电晶体管， 不仅能为构建高能效数据中心提供核心器件方案，也为发展下一代高算力人工智能芯片奠定关键技术基 础。这项技术突破填补了国内在原子尺度铁电晶体管领域的空白，打破了国际巨头在AI芯片核心器件 领域的垄断。 公司方面，据中证报表示，A股相关概念股主要有国林科技、亿道信息等。 *免责声明：文章内容仅供参考，不构成投资建议 *风险提示：股市有风险，入市需谨慎 ...

然而，传统铁电晶体管存在能耗过高、逻辑电压不匹配等短板，限制了其大规模应用。为此，北京大学 邱晨光研究员－彭练矛院士团队，利用纳米栅极结构设计，巧妙解决了铁电材料"改变极化状态"需要高 电压高能耗的问题。 "我们持续精进工艺，将铁电晶体管的物理栅长缩减至极限1纳米，这一精度达到原子尺度，促成铁电层 内部形成高强度电场，仅需极少外部能量（0.6V电压）激发，即可轻松翻转铁电极化。"邱晨光表示， 这一技术打破了传统铁电晶体管的物理限制，使得能耗比国际最好水平整整降低了一个数量级。 "纳米栅的设计就好像是对电场进行了'杠杆放大'，能够以极低的电压代价，驱动铁电材料发生极化反 转，从而在物理机制上实现了能耗的跨越式降低。"邱晨光进一步解释，有着超低工作电压与极低能耗 特性的纳米栅铁电晶体管，不仅能为构建高能效数据中心提供核心器件方案，也为发展下一代高算力人 工智能芯片奠定关键技术基础。 科技日报北京2月23日电 （记者张佳星）"我们将铁电晶体管的物理栅长缩减到了1纳米极限。"23日， 北京大学电子学院研究员邱晨光告诉科技日报记者，团队创造性地制备了迄今尺寸最小、功耗最低的铁 电晶体管，有望为AI芯片算力和能效提升提供 ...

彭海琳介绍，长久以来，铁电材料因其可逆极化与非易失存储特性，被视为打通存算一体、突破冯·诺 依曼架构（在冯·诺依曼架构下，计算和存储是相互分离的）瓶颈的关键。然而，当芯片工艺逼近亚5纳 米（小于5纳米）节点，传统铁电薄膜面临均匀性差、界面缺陷多、厚度减薄后铁电性骤降等难题。研 究团队创新性依托其自主研发的高迁移率铋基二维半导体Bi2O2Se（硒氧化铋），首次实现了原子级平 整的二维铁电自然氧化物Bi2SeO5（硒酸铋）及异质结构晶圆级均匀制备。这种新型铁电氧化物不仅具 有高达24的介电常数和超过600℃的高温结构稳定性，更在单晶胞厚度（约1纳米）下仍保持优异铁电 性，彻底摆脱了传统铁电材料的尺寸限制。 在此基础上，研究团队还制备出高性能铁电晶体管阵列，能效领先其他存储技术1至2个数量级，并展现 出32个稳定多级存储态与超10年数据保持能力。尤为亮眼的是，在0.8伏超低电压和20纳秒高速写入条 件下，器件经受住1.5万亿次循环考验，可靠性远超云端AI计算的严苛标准。更进一步，团队利用该器 件构建出可动态重构的存内逻辑电路——在低于1伏的常规CMOS电压下，同一器件既能执行逻辑运 算，又能切换为非易失存储，真正实 ...

Core Viewpoint - The third phase of Yangtze Memory Technologies Co., Ltd. (YMTC) is set to be completed and put into production this year, which will attract around 200 upstream and downstream enterprises to the region [1] Group 1: Company Overview - Yangtze Memory was established in July 2016 and is a leading company in China's storage chip manufacturing sector, primarily providing 3D NAND flash wafers, embedded storage chips, and solid-state drives [1] - The second phase of YMTC was established in December 2021 with a registered capital of 60 billion yuan, while the third phase was established in September 2025 with a registered capital of 20.72 billion yuan [1] Group 2: Market Position and Projections - By 2025, YMTC is expected to hold approximately 7% to 8% of the global market share in storage capacity, with projections indicating it could exceed 10% by 2026, potentially surpassing Micron Technology to become the fourth-largest storage chip manufacturer globally [1] - Market research firm Omdia estimates that YMTC's capital expenditure will be more aggressive than its global peers, accounting for about 20% of total global NAND flash investment by 2025, with expectations for continued growth [3] Group 3: Industry Context - The global demand for storage chips is on the rise due to the ongoing AI computing power boom, marking a new growth cycle for the industry [4] - NAND flash prices are projected to increase by over 40% in the first quarter of this year, and the global storage chip market is expected to reach 1.584 trillion yuan by 2031, with a compound annual growth rate of 9.3% from 2025 to 2031 [4] Group 4: Strategic Initiatives - The company aims to enhance its product matrix and expand both domestic and international markets while focusing on technological innovation and the development of next-generation flash memory chips [3] - The recent establishment of a joint-stock company indicates that YMTC has completed its shareholding reform, and it has entered the list of China's top ten unicorns with a valuation of 160 billion yuan [4]

Core Viewpoint - The company is making steady progress in the development of next-generation high-speed testing products, ensuring its continued leadership in the storage technology testing field [1] Group 1: ASIC Chips - The company is advancing the research and development of next-generation high-speed testing products for ASIC chips [1] Group 2: Computing Power Chip Test Machines - The company plans to accelerate the launch of high-end computing power chip testing prototypes, anticipating further progress this year in conjunction with strategic customers' new product timelines [1] Group 3: AI Layout with Smart Glasses - Smart glasses are viewed as a significant entry point in the AI era, with optimistic prospects; the company aims to leverage its entire product line to create a one-stop solution for edge AI, including storage testing, SoC testing, and optical inspection [1]

Core Insights - Reexen Technology, a leading AI chip company specializing in integrated storage and computing, has recently completed a B+ round of financing with participation from Shenzhen Changsheng Industrial Collaborative Private Equity Investment Fund [2] Company Overview - Reexen Technology has been rooted in Futian District, Shenzhen, focusing on the core technology research and industrialization of integrated storage and computing AI chips. The company has achieved significant breakthroughs in large capacity, high bandwidth, and high energy efficiency, particularly demonstrating strong technical advantages in large model inference [3] - The company successfully addresses the high demands for storage capacity and energy efficiency in cloud-edge terminal inference by deeply integrating multi-level storage and computing, providing efficient and low-power solutions for smart terminal devices and cloud AI inference [3] Financing Purpose - The financing will primarily be used to further optimize the multi-level architecture of integrated storage and computing AI chips, accelerate product research and iteration, and expand market coverage to promote commercialization in more application scenarios. Additionally, the company will increase investment in talent development and team building to maintain its leading position in technological innovation and market competition [4] Investment Rationale - The chairman of the Futian Guidance Fund stated that Reexen Technology has demonstrated outstanding technical strength and continuous innovation capability in the fields of storage-computing integration and edge AI chips. The company has performed exceptionally in technology route selection, product advancement, and industrialization practices, representing a strategic emerging industry in Futian District [5] Investment Ecosystem Perspective - The founder of Shandao Venture Capital Network noted that Reexen Technology's B+ round financing reflects the continuous optimization of Shenzhen's hard technology industry ecosystem. In recent years, the government has introduced multiple policies to support the development of the semiconductor and integrated circuit industries, with Futian District providing precise support for hard technology enterprises through guidance funds [6] - The successful financing not only showcases the company's technical advantages in integrated storage and computing AI chips but also sets an innovation benchmark for the industry. The active participation of venture capital institutions reflects a long-term positive outlook on the hard technology sector and highlights the precise judgment of project value by managers [6]

Core Insights - The storage industry is experiencing a significant upcycle driven by AI demand, extending from HBM to traditional storage sectors, with Micron's gross margin reaching a historical high of 66-68% for the next quarter, indicating a stronger cycle than previous ones [1][3]. Group 1: AI Demand and Storage Market Dynamics - The price increase of storage products reflects the supply-demand relationship in the market, primarily driven by AI server demand [3]. - The current AI storage cycle is characterized by a shift in focus from training to inference, leading to differentiated demands for "low latency, high capacity, and high bandwidth" storage [3][14]. - The three major manufacturers (Samsung, SK Hynix, Micron) are prioritizing capital expenditures towards HBM and DRAM, resulting in structural supply-demand imbalances and significant price increases [3][6]. Group 2: Role of Different Storage Types in AI Servers - HBM serves as the "performance ceiling" for AI servers, being a high-bandwidth, high-power product that directly impacts the model scale and response speed [11]. - DRAM (DDR5) acts as a data exchange hub, connecting HBM and NAND, and is crucial for handling concurrent tasks in AI servers [12]. - NAND (SSD) functions as a fast persistent layer for frequently accessed data, while HDD serves as a low-cost container for large volumes of cold data [12][14]. Group 3: Addressing the "Memory Wall" Challenge - The "memory wall" bottleneck arises from the disparity between computing speed and data transfer speed, leading to high GPU idle rates [5][16]. - Solutions to this issue include upgrading HBM to 16-Hi stacks to enhance bandwidth and implementing 3D stacked SRAM to reduce latency [18][19]. - The integration of computing capabilities within storage (compute-in-memory) is anticipated to be a long-term solution to the "memory wall" problem [21]. Group 4: HBM Market Supply and Demand - HBM demand is closely tied to AI chip shipments, with expectations for HBM supply to increase by over 60% by 2026 due to significant capital investments from the three major manufacturers [6][24]. - The combined monthly HBM production capacity of the three manufacturers is projected to rise from approximately 390,000 wafers to 510,000 wafers by the end of 2026, translating to an estimated supply of 41.9 billion GB of HBM [29][34]. - The HBM market is expected to be in a "tight balance" state in 2026, with demand estimated at around 42 billion GB, indicating a competitive landscape among manufacturers [39][40].

（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 Insights - The global AI infrastructure spending reached $82 billion, with servers accounting for 98% and storage devices only 1.4% of the total spending, indicating a significant "storage gap" in the industry [2] - The emergence of AI NAS (Network Attached Storage) is seen as a solution to the "data hunger" problem, with the AI NAS market projected to reach $730 million by 2031, growing at a compound annual growth rate (CAGR) of 49% [2][11] - Traditional NAS systems are evolving into AI NAS, transforming from mere data storage to powerful computing nodes capable of handling AI workloads [2][12] AI NAS Market Dynamics - The global AI NAS market is expected to grow significantly, with a forecasted market size of $730 million by 2031 and a CAGR of 49% [2][11] - AI NAS is positioned to address the inefficiencies of traditional storage systems, which struggle with the increasing volume of unstructured data generated by AI applications [3][12] - The demand for AI NAS is driven by the need for local data processing and compliance with data privacy regulations, making it attractive for both individual users and small to medium-sized enterprises (SMEs) [6][7] Technological Advancements - AI NAS integrates storage and computing capabilities, allowing for efficient data processing and retrieval, which traditional NAS systems cannot achieve [3][5] - The architecture of AI NAS includes AI modules that enhance data synchronization, retrieval, and processing, significantly improving operational efficiency [3][4] - The performance of AI NAS is exemplified by devices that support high-speed data transfer rates, enabling real-time editing and processing of large video files [4][5] Application Scenarios - AI NAS is gaining traction in various sectors, including healthcare, where it can process medical imaging data rapidly while ensuring compliance with data security regulations [7][12] - The technology is also appealing to SMEs, which face challenges related to high computing costs and a lack of IT talent, as AI NAS offers a cost-effective solution [12] - The growing awareness of data sovereignty and privacy concerns is driving the adoption of AI NAS among enterprises that prefer local data storage solutions [6][8] Cost Efficiency - AI NAS presents a cost advantage over cloud services, with significant savings for SMEs in terms of operational costs and response times for AI applications [8][10] - The decreasing costs of storage technology, such as NVMe SSDs, are further enhancing the economic viability of AI NAS solutions [8][10] - Companies deploying AI NAS have reported substantial improvements in IT efficiency and reductions in labor costs, making it an attractive investment [8][10]

Company and Industry Summary Company Overview - **Company Name**: 精智达 (Jingzhida) - **Industry**: Semiconductor Testing Equipment Key Points Business Progress and Technical Breakthroughs - In 2025, Jingzhida achieved significant business breakthroughs, particularly in the AIC chip, which supports a testing rate of 9Gbps, primarily used in FT high-speed testing machines and HBM CPU processors [3] - The company’s high-speed FT testing machine received customer validation and secured its first order, with some metrics surpassing international counterparts [3][4] - The HBM high-speed CP testing machine is nearing the final validation phase, with expectations for significant developments in 2026 [3][5] Revenue and Financial Outlook - Revenue orders for 2025 have exceeded initial expectations, reaching a historical high, with an increasing proportion of high-margin products contributing to overall gross margin growth [2][5] - For 2026, the company anticipates further revenue growth, driven by AI technology and a focus on five key product lines: storage chip testing, computing chip testing, probe cards, AR/VR/XR testing, and advanced optical testing [2][5] Challenges and Strategic Responses - Current challenges include overall capacity saturation and tight supply-demand conditions, reflecting high market recognition of the company’s products [6] - The company plans to reference international manufacturers' expansion experiences to significantly increase capacity to meet customer demands, prioritizing existing strategic clients [6][10] - Jingzhida aims to maintain its leading position in DRAM and high-speed FT sectors while responding to new equipment demands from key clients [7] New Product Development and Future Plans - The company is accelerating the launch of high-end professional chip testing machine prototypes, leveraging its SOC and ASIC R&D advantages to meet integrated processing and testing needs [8] - Collaborations with international clients like Meta for optical detection solutions are underway, alongside partnerships with domestic clients for comprehensive solutions in storage, SOC, and optical detection [8] - Future plans include significant advancements in 2026, with new product launches and mass deliveries, particularly in emerging sectors like smart glasses [8][9] Production Capacity Expansion - The company’s expansion plans are informed by historical experiences, with international peers having expanded capacity by nearly three times over the past two to three years [8] - Jingzhida's expansion strategy is customer-driven, ensuring confidence in post-expansion delivery capabilities [8][10] - The company is reallocating production lines to meet key customer delivery needs amidst high semiconductor orders and market conditions [10]