申请入围"中国IC独角兽" 半导体高质量发展创新成果征集 据知情人士透露，韩国政府计划限制 HBM （高带宽内存）设备的出口，特别是关键的 TC Bonder （热压键合机）设备。据悉，韩国政府已要求本土主要 的 TC Bonder 制造商，暂停向特定国家和地区的客户交付设备，同时收紧对相关技术出口的审批流程。 HBM TC Bonder 在深入探讨韩国断供事件之前，我们先来了解一下 TC Bonder 到底是什么，以及它在半导体行业中为何如此重要。 TC Bonder ，也就是热压键合机，别看它只是一台设备，却是半导体封装环节中的 " 关键先生 " 。尤其在 HBM 的制造过程中，它承担着将多个 DRAM 芯 片精准堆叠并实现电气连接的重任。打个比方，如果把 HBM 比作一座高楼大厦，那么 TC Bonder 就是那位技艺精湛的 " 建筑大师 " ，负责将每一块 " 芯 片砖块 " 严丝合缝地堆叠起来，确保整座 " 大厦 " 能够稳定运行。 目前，根据填充材料的不同，热压键合技术分为 TC-NCF 、 TC-NCP 、 TC-CUF 、 TC-MUF 等多种类型；而依据基板材料的差异，又有 Chip-to- ...

Core Viewpoint - The recent air combat between Pakistan and India highlights the technological superiority of China's J-10C fighter over France's Rafale, demonstrating the importance of integrated systems in modern warfare [2][34][37] Group 1: Air Combat Details - Pakistan's air force claimed to have shot down five Indian aircraft, including three Rafale jets, using the J-10C equipped with PL-15E missiles [2][4][37] - The J-10C's radar capabilities allow it to detect targets at a range of 260 kilometers, significantly surpassing the Rafale's detection range of 180 kilometers [10][13] - The engagement showcased the J-10C's ability to engage targets before the Rafale could respond, emphasizing the "first to detect, first to attack" principle [26][34] Group 2: Technological Advantages - The J-10C's KLJ-7A AESA radar, utilizing GaN technology, provides superior detection and tracking capabilities compared to the Rafale's RBE2-AA radar, which uses GaAs technology [10][13][15] - The PL-15E missile has a range of up to 200 kilometers, while the Rafale's MICA missile has a maximum range of only 80 kilometers, creating a significant tactical disadvantage for the Rafale [20][22][24] - The integration of electronic warfare capabilities in the J-10C allows it to disrupt enemy radar systems effectively, enhancing its combat effectiveness [16][28] Group 3: System Integration and Tactics - Pakistan's "Sky Net" operational framework, centered around the ZDK-03 early warning aircraft, enables real-time data sharing and coordination among its forces [28] - The combined tactics of using early warning systems, J-10C strikes, and ground-based electronic warfare resulted in rapid engagement and destruction of Indian aircraft [31][32] - The air combat demonstrated the limitations of India's multi-national aircraft fleet, which struggles with data integration and timely information sharing [28][31] Group 4: Market Implications - The successful performance of the J-10C in combat is expected to open new markets for Chinese military exports, with at least six countries likely to express interest in procurement within the next five years [37] - The stock price of Chengdu Aircraft Industry Group surged nearly 50% following the air combat, reflecting investor confidence in the company's capabilities and future prospects [37]

Core Viewpoint - Lenovo has developed its first self-researched 5nm chip, named SS1101, which is integrated into the new YOGAPad Pro 14.5 tablet, marking a significant step in its semiconductor capabilities and breaking the perception of being merely an assembler of components [2][4]. Group 1: Chip Development - The SS1101 chip features an ARM architecture with a 10-core CPU configuration (2+2+3+3), allowing for efficient multitasking and high performance [2]. - The chip operates at a main frequency of 3.29GHz, enhancing its processing speed for running multiple applications and gaming [2]. - The GPU component is based on Arm's "Mali-G720-Immortalis," although specific details about its capabilities and power consumption remain undisclosed until the product launch on May 20 [2][4]. Group 2: Strategic Importance - Lenovo's chip development is not a spontaneous decision; it has been a planned initiative since the establishment of Dingdao Zhixin in 2021, led by experienced professionals from major companies like Huawei and OPPO [4]. - The successful development of the 5nm chip represents a significant achievement for Lenovo, allowing it to optimize product performance, such as improving battery life and system responsiveness [4]. - This move enhances Lenovo's brand image and competitiveness in the global market, shifting the perception from being a mere assembler to a company with core technological capabilities [4].

申请入围"中国IC独角兽" 半导体高质量发展创新成果征集 2024 年，全球芯片分 销市场格局发生变化， 常年稳居第一的艾睿被文晔反超 （依次为 文晔、艾 睿、大联大、安富利 ） 。 随着今年一季度 TOP4 "双 W " （大联大 WPG 、文晔 WT ） 与"双 A " （艾 睿 Arrow 、安富利 Avnet ） 业绩出炉，排名又有新变化， 按营收高低依次为：大联大、文晔、艾 睿、安富利。 其中， 大联大 以2,488.3亿元 新台币 （约合82.09亿美元） 小幅领先文晔 2,474亿元 新台币 （约合 81.67亿美元） ， 暂居榜首 ；艾睿与安富利分别以68.14亿美元和53.15亿美元的营收分列第三、 第四位。 这四家芯片分销大厂的业绩增长情况如何？芯片分销市场整体好起来了吗？ 以下文章来源于芯世相 ，作者Silvia 芯世相 . 芯片电子元器件IC半导体分销教科书式必读公众号【芯世相】；缺芯片，销库存，保品质，市场行情早 知道，尽在芯片超人。加我：chrisjiang08 具体来看，大联大一季度营收达 新台币 2,488.3 亿元 （约合 82.09 亿美元） ， 季增 7.4% 、 年 增 ...

Core Viewpoint - The article discusses the recent departure of key personnel from Alibaba's Tongyi Laboratory, particularly focusing on the implications of these changes for Alibaba's AI strategy and the competitive landscape in the tech industry [2][4]. Group 1: Personnel Changes - Bo Liefeng, the head of the application vision team at Alibaba's Tongyi Laboratory, left the company on April 30, 2023, after more than two years of service [2][6]. - His departure follows that of another senior employee, Yan Zhijie, who was the head of the voice team, indicating a trend of high-level exits from the laboratory [4][6]. - Bo Liefeng is speculated to have joined a major internet company in the U.S., possibly ByteDance or Tencent, as the vice general manager of the multimodal model department [4][6]. Group 2: Implications for Alibaba - The exit of Bo Liefeng may pose challenges for Alibaba's large model strategy, potentially slowing down the advancement of related technologies and extending product iteration cycles [4][6]. - The integration and commercialization of multimodal technologies may also be disrupted, necessitating a reassessment of commercial promotion plans [4][6]. - The competitive landscape could shift if Bo Liefeng contributes to a rival company's AI initiatives, creating additional obstacles for Alibaba's expansion in the AI sector [4][6]. Group 3: Background of Bo Liefeng - Bo Liefeng, born in 1978, holds a Ph.D. from Xi'an University of Electronic Science and Technology and has extensive experience in machine learning, deep learning, computer vision, and natural language processing [9]. - Prior to joining Alibaba, he worked at Amazon as a chief scientist, where he was instrumental in developing the Amazon Go cashier-less shopping experience [9]. - He also served as the chief scientist at JD Digital Technology Group before transitioning to Alibaba in 2022 [9].

受相关消息影响，亚洲半导体芯片股集体上涨，日本东京威力科创上涨3%，Disco上涨6.5%， Screen上涨2.1%，Lasertec上涨3.6%，Advantest上涨3.9%。韩国海力士涨2.7%，三星电子涨 1.7%，Wonik IPS涨2%，DB Hitek涨1.8%。隔夜费城半导体指数上涨1.7%，英伟达上涨3.1%。 申请入围"中国IC独角兽" 半导体高质量发展创新成果征集 5月8日消息， 美国商务部发言人表示， 美国总统特朗普计划撤销和修改拜登时代限制人工智能芯片出 口的规定！ 据报道，特朗普政府计划取消拜登时代的人工智能芯片限制，作为修订半导体贸易限制的更广泛努力的 一部分，这些限制已引起主要科技公司和外国政府的强烈反对。 据了解，该政策将国家划分为三大类，以监管英伟达公司等企业的芯片出口。知情人士称，5月15日所 谓的人工智能扩散规则生效时，特朗普政府将不会执行该规则，不过这项废除决定尚未最终确定。 美国商务部发言人表示："拜登的人工智能规则过于复杂、官僚主义，会阻碍美国的创新。我们将用一 个简单得多的规则来取代它，这个规则将释放美国的创新能力，并确保美国在人工智能领域的主导地 位。"知情 ...

Core Viewpoint - The Trump administration plans to rescind Biden-era AI chip restrictions, aiming to revise semiconductor trade policies, which could significantly impact companies like Nvidia and the semiconductor industry as a whole [2][4]. Group 1: Policy Changes - The repeal seeks to reshape a policy from the Biden administration that established three tiers of countries for regulating chip exports from companies like Nvidia [2][4]. - The Trump administration will not enforce the AI diffusion rule starting May 15 and is working on a new rule to strengthen overseas chip controls [4]. Group 2: Market Implications - The changes are a clear positive for Nvidia and similar companies, as the adjustment in export restrictions could lead to significant revenue changes, particularly in the Middle East [4][5]. - Singapore is identified as a key player in this dynamic, with Nvidia's revenue from Southeast Asia heavily reliant on Singapore, which serves as a billing center for products that may be shipped elsewhere [5][6]. Group 3: Domestic Impact - For domestic clients, the policy changes could alleviate issues related to AI training resource consumption and supply instability, although large companies may find it challenging to procure chips through unconventional channels [6]. - The overall relaxation of restrictions is expected to increase chip availability in the market, benefiting AI training resource needs and providing time for the development of domestic AI chips [6]. Group 4: Long-term Trends - The sentiment towards domestic chips may experience slight fluctuations, but the demand for domestic AI chips is expected to grow steadily, regardless of external changes [6]. - Nvidia's role in the ongoing trade dynamics suggests it may not merely be a pawn but could be influencing the broader strategic landscape [5][7].

半导体高质量发展创新成果征集 5月5日有消息传出，美国特朗普政府可能最快于本周公布针对半导体加征关税的细节，市场预估税率区 间可能高达25%至100%，并且新规则中不排除以晶圆制造地（waferout）作为源产地来加征关税。这对 台积电、三星等产能集中在亚洲地区的晶圆制造大厂，以及英伟达、苹果、高通、联发科等依赖于亚洲 晶圆代工产能的芯片设计厂商，它们未来的发展似乎蒙上了一层阴影。 回顾美国在半导体领域的一系列动作，早在2024年9月13日，美国政府便宣布在原有对华301关税的基础 之上，对进口自中国的部分钢铁产品、铝产品、半导体等一系列"战略性产品"分阶段（2024年至2026 年）加征关税。其中，2025年半导体的关税税率从25%提高到50%，涉及太阳能电池板使用的多晶硅和 硅晶圆。进入2025年，4月1日，美国商务部下属部门工业与安全局（BIS）根据《1962年贸易扩展法》 第"232条款"赋予的权力，对进口半导体及其衍生产品、进口药品及药用成分发起国家安全调查，并征 求公众意见。此次调查主要聚焦外国补贴、供应链依赖风险、美国国内产能瓶颈等14项细节，旨在确定 进口半导体及其衍生产品对美国国家安全的影响 ...

申请入围"中国IC独角兽" 半导体高质量发展创新成果征集 康希通信在公告中表示，此次投资策略的变更不会对公司的业务开展、生产经营活动和财务状况 造成不利影响。这一调整反映了公司在面对市场环境和战略合作条件变化时的灵活应对能力。 康希通信在接受机构调研时表示，2025年公司Wi-Fi 7产品出货量有望与Wi-Fi 6持平，目前已参与 Wi-Fi 8协议的预研工作。根据行业预测，2025年全球Wi-Fi 7渗透率将维持在10%左右，该公司凭 借在Wi-Fi 7领域的提前布局及与高通、联发科等主芯片厂商的深度合作，持续优化产品功率效率 和低功耗特性，巩固市场竞争力。 关于人才激励机制，康希通信2024年已完成3,000万元股份回购用于股权激励及员工持股计划， 2025年具体实施方案正在筹备中。公司表示将持续通过长效激励政策强化核心团队稳定性，相关 细节将以公告形式披露。 公开资料显示，康希通信专注于射频前端芯片领域的研发与设计工作，致力于研发高性能、高线 性度、高可靠性的Wi-Fi射频前端芯片及模组。公司产品包括Wi-Fi FEM、IoT FEM、V2X FEM、 无人机等，集成了公司自主研发的PA、LNA及射频开 ...

Core Viewpoint - The rapid advancement of AI large models is driving the transition of AI from a supportive tool to a core productivity force, with computing power chips being crucial for training and inference of these models [2]. Group 1: Computing Power Indicators - **Process Technology**: Major overseas companies are utilizing advanced process technologies, with Nvidia's latest Blackwell series using TSMC's 4NP (4nm) technology, while AMD and Intel are at 5nm. Domestic manufacturers are transitioning from TSMC's 7nm to SMIC's 7nm [3][4]. - **Transistor Count and Density**: Nvidia's B200 chip, using Chiplet technology, has a transistor density of 130 million/mm², while Google's TPU Ironwood (TPU v7p) boasts a density of 308 million/mm², significantly higher than competitors [6][7]. - **Performance Metrics**: Nvidia's GB200 achieves FP16 computing power of 5000 TFLOPS, while Google's TPU Ironwood reaches 2307 TFLOPS, showcasing a significant performance gap [10][11]. Group 2: Memory Indicators - **Memory Bandwidth and Capacity**: Most overseas manufacturers are using HBM3e memory, with Nvidia's GB200 achieving a bandwidth of 16TB/s and a capacity of 384GB, significantly surpassing domestic chips that primarily use HBM2e [18][19]. - **Arithmetic Intensity**: Nvidia's H100 has a high arithmetic intensity close to 600 FLOPS/Byte, indicating efficient memory bandwidth usage, while domestic chips exhibit lower arithmetic intensity due to their lower performance levels [20][21]. Group 3: Interconnect Bandwidth - **Interconnect Capabilities**: Overseas companies have developed proprietary protocols with interconnect bandwidth generally exceeding 500GB/s, with Nvidia's NVLink5 reaching 1800GB/s. In contrast, domestic chips typically have bandwidth below 400GB/s, with Huawei's 910C achieving 700GB/s [26][27].