该职位涉及"设计和改进新的硬件架构以突破计算效率的界限"，以及在硬件设计过程中使用人工智能。 候选人需要熟悉加速器友好的硬件设计语言，如 Chisel、VHDL 和 Verilog。 他们最好具有"在新型 AI 硬件架构上模拟训练工作负载"的经验。 公众号记得加星标⭐️，第一时间看推送不会错过。 来源：内容来自 datacenterdynamics，谢谢 。 招聘信息指出"新颖的人工智能硬件架构"。 埃隆·马斯克的生成式人工智能公司 xAI 正在招聘定制硅片开发人员。 在一份此前未公布的招聘信息中，该公司表示正在寻找能够"共同设计下一代人工智能系统，涵盖从硅片到软件编译器再 到模型"的员工。 参参考考链链接接 https://www.datacenterdynamics.com/en/news/elon-musks-xai-plans-to-develop-its-own-custom-silicon/ *免责声明：本文由作者原创。文章内容系作者个人观点，半导体行业观察转载仅为了传达一种不同的观点，不代表半导体行业观察对该观点赞同或支持， 如果有任何异议，欢迎联系半导体行业观察。 该员工将加入 xAI 团队， ...

公众号记得加星标⭐️，第一时间看推送不会错过。 在大多数人看来，因为全球正在疯抢人工智能主权，这就引致大家争夺GPU，进而推动HBM需求。 知名分析机构Yole在今年三月发布的一个报告中称，自2022 年底 ChatGPT 横空出世以来，生成式人工智能蓬勃发展，推 动 2023 年 HBM 比特出货量同比增长 187%，增幅空前，2024 年更是飙升 193%。 "预计这一增长势头将持续下去。HBM 的增速远超整体 DRAM 市场。全球 HBM 收入预计将从 2024 年的 170 亿美元增 长至 2030 年的 980 亿美元，复合年增长率达 33%。"Yole在报告中接着说。三大存储巨头最新的营收数据显示，HBM也 的确走在Yole预言的这条轨道上。 作为新晋的DRAM龙头，受高带宽存储器 (HBM) 需求激增的推动，SK 海力士预计第二季度营业利润将达到近 9 万亿韩 元（66 亿美元），预计HBM销售额今年将占SK海力士DRAM总收入的50%以上，高于2024年第四季度的40%以上。另一 家HBM供应商美光也在HBM的推动下创下业绩新高。 报告还援引分析师的警告称，到 2026 年，三星的 HBM 出 ...

Core Viewpoint - The article emphasizes the critical role of the physical layer (PHY) in data communication, particularly in the context of emerging technologies such as artificial intelligence and high-performance computing, highlighting its importance in meeting the increasing demands for bandwidth, low latency, and energy efficiency [3][6][11]. Summary by Sections Importance of PHY - The physical layer has evolved from supporting traditional industries to becoming foundational for AI factories and large-scale data centers, acting as a key driver for data transmission and communication [3]. - As data centers handle massive amounts of data, the significance of PHY increases, especially for AI and HPC workloads that require unprecedented system performance [3][6]. Standards and Applications - Understanding the physical layer is crucial for maintaining competitiveness in various applications, with different standards developed to address specific issues [4]. - Standards like HDMI and DisplayPort illustrate the need for compatibility and efficiency in system design, balancing cost and functionality [5]. Design Challenges - Designing PHY for speeds exceeding 100G presents numerous challenges, including process technology dependence, signal integrity, system design constraints, and packaging integration [8]. - The transition from NRZ to PAM signaling represents a significant shift in technology, necessitating advanced design techniques to meet increasing bandwidth demands [7][8]. Chip-to-Chip Communication - The development of chip-to-chip communication standards, such as UCIe, aims to achieve high bandwidth with low power consumption, which is essential for modern 3D systems [9]. - The integration of multiple dies in a system-on-chip (SoC) architecture requires careful consideration of physical layer protocols to optimize performance [9][10]. Collaboration Between Disciplines - Effective collaboration between analog and digital engineers is necessary to bridge the gap between different domains, ensuring that physical effects are adequately addressed in system design [10][11]. - A comprehensive understanding of how physical effects impact system performance is vital for optimizing designs and achieving desired outcomes [11]. Future Outlook - As the industry progresses towards higher standards like 448G, the challenges will intensify, particularly with the emergence of chip decomposition and optical I/O [11]. - The PHY layer is increasingly viewed as a strategic enabler, necessitating continuous innovation and commitment to pushing technological boundaries [11].

公众号记得加星标⭐️，第一时间看推送不会错过。 来源：内容来自 straitsresearch，谢谢 。 印度正采取大胆举措，力争成为全球半导体行业的主力军。这一转变源于印度对经济增长、技术独立和加强国家安全的渴 望。半导体，这种微型芯片为智能手机、汽车、医疗设备和数据中心等各种设备提供动力，是现代生活不可或缺的一部 分。随着印度电子和科技产业的发展，对这些关键部件的需求也日益增长。 印度为何需要强大的半导体行业 印度政府正在采取有力措施，力图改变该国的半导体格局。这项举措的核心是"印度半导体计划"（ISM），这是一项耗资 100亿美元的计划，旨在为芯片制造和设计打造强大的生态系统。该计划还得到了"生产挂钩激励计划"（PLI）和"设计挂 钩激励计划"（DLI）等其他计划的支持，旨在鼓励企业在印度投资。 ISM 正在努力： 印度的努力已初见成效。多家大型公司已宣布对半导体制造和封装进行大规模投资。例如： 塔塔集团与台湾 PSMC 合作，在古吉拉特邦建造一座价值 91,000 亿印度卢比的芯片工厂。 美光科技正在投资 27.5 亿美元建设一个组装、测试、标记和封装 (ATMP) 工厂。 阿达尼集团、HCL-富士康 ...

公众号记得加星标⭐️，第一时间看推送不会错过。 来源：内容来自 semiconductor-digest，谢谢 。 明尼苏达大学双城分校的研究为一种可以使计算机内存更快、更节能的材料提供了新的见解。 在这项新研究中，该团队展示了一种更有效的方法，利用一种名为Ni₄W（镍和钨的组合）的材料来控制微型电子设备中的 磁化。研究团队发现，这种低对称性材料能够产生强大的自旋轨道扭矩（SOT），这是下一代存储器和逻辑技术中操控磁 性的关键机制。 "Ni₄W 降低了写入数据的功耗，有可能显著降低电子设备的能耗，"该论文的资深作者、明尼苏达大学双城分校电气与计 算机工程系 (ECE) 杰出麦克奈特教授兼罗伯特 F. 哈特曼主席王建平 (Jian-Ping Wang) 表示。 这项技术可以帮助减少智能手机和数据中心等设备的电力消耗，使未来的电子产品更加智能、更加可持续。 "与传统材料不同，Ni₄W 可以在多个方向上产生自旋电流，无需外部磁场即可实现磁态的'无场'切换。我们观察到 Ni₄W 无论是单独使用还是与钨层叠时，都具有较高的自旋场效率和多方向性，这表明其在低功耗、高速自旋电子器件中具有巨 大的应用潜力。"王教授团队五年级 ...

Core Viewpoint - The article discusses the advancements made by the startup CDimension in the development of two-dimensional (2D) semiconductors, specifically focusing on their ability to grow molybdenum disulfide (MoS2) on silicon at low temperatures, which could revolutionize chip manufacturing and reduce power consumption significantly [3][5]. Group 1: CDimension's Technology - CDimension claims to have solved key challenges in the industrialization of 2D semiconductors, including wafer-level uniformity, device performance, and compatibility with silicon manufacturing processes [3][4]. - The proprietary process developed by CDimension allows for the growth of single-layer MoS2 at approximately 200°C, avoiding damage to the underlying silicon circuits, which is a significant improvement over traditional methods that require temperatures up to 1000°C [4]. - The startup is currently shipping silicon wafers with grown 2D materials for customer evaluation and integration into devices, showcasing the potential for 2D materials to be used in scalable logic devices [4][5]. Group 2: Industry Implications - Major chip manufacturers like Intel, Samsung, and TSMC are exploring the replacement of silicon nanosheets with MoS2 and other 2D semiconductors, indicating a shift in the semiconductor industry towards these advanced materials [4]. - The low-temperature synthesis demonstrated by CDimension's team can produce MoS2 transistors with multiple stacked channels, potentially meeting or exceeding the performance requirements of future 10A (1 nanometer) nodes [4]. - The motivation for adopting 2D semiconductors includes a significant reduction in power consumption, with devices made from CDimension's materials consuming only one-thousandth of the power of traditional silicon devices [5].

Core Viewpoint - Synopsys has successfully completed its acquisition of Ansys for $35 billion, receiving approval from Chinese regulators, which clears the final hurdle for the transaction that was previously approved by U.S. and European regulators with specific conditions [1][8]. Group 1: Strategic Importance of the Acquisition - The merger is seen as a transformative milestone for Synopsys, enhancing its capabilities in chip design and system-level simulation, which is crucial for developing complex intelligent systems [1][3]. - The acquisition opens new growth opportunities in sectors such as aerospace, automotive, and industrial equipment, potentially expanding Synopsys's market and business portfolio [2][7]. Group 2: Technological Integration - The combined company will provide a unified platform for developing complex multi-domain products, integrating EDA tools with advanced simulation capabilities [4][5]. - The integration of Ansys's simulation data with Synopsys's AI-driven EDA tools will enable smarter, automated collaborative design processes, optimizing power, performance, thermal characteristics, and reliability [5][6]. Group 3: Market Position and Competition - The merger reduces the number of independent players in critical technology areas, which may intensify competition and regulatory scrutiny, particularly from U.S., Chinese, and EU authorities [8][9]. - The combined entity is expected to strengthen Synopsys's technological leadership and position within the semiconductor ecosystem, potentially leading to further mergers or ecosystem changes in response to increased competition [2][7]. Group 4: Regulatory Considerations - The acquisition has raised antitrust concerns, but it has been approved by regulatory bodies with conditions to ensure interoperability with competitors' solutions [8][9]. - The merged company may face increased regulatory pressure due to its influence on key systems and design workflows in sensitive sectors like aerospace and defense [9].

Core Insights - The article discusses the launch of IBM's Power11 processors, which are designed to support large-scale computing environments with high memory and I/O capabilities [4][10] - It highlights the evolution of IBM's Power processors, particularly the transition from Power10 to Power11, and the challenges faced during the manufacturing process [5][7] - The article emphasizes the competitive landscape in the server market, particularly against AMD and Intel, and how IBM aims to maintain its position through advanced technology [6][19] Summary by Sections Power11 Launch - Power11 processors were launched and are set to be available on July 25, featuring enhancements over the previous Power10 generation [4] - The processors are designed to handle extensive memory and I/O requirements, making them suitable for enterprise-level applications [3] Evolution of Power Processors - Power11 is a refined version of the Power10, which faced manufacturing delays due to issues with GlobalFoundries' process technology [5][6] - The transition to 7nm technology was necessitated by delays in the 10nm process, leading to a redesign of Power10 [7][8] Competitive Landscape - The article notes that AMD has gained market share due to Intel's manufacturing challenges, which parallels IBM's own struggles with Power10 [5][6] - IBM's strategy includes leveraging its long-standing expertise in chip design and manufacturing to compete effectively in the high-performance computing market [8][19] Technical Specifications - Power10 and Power11 both feature 16 cores and 18 billion transistors, with Power11 offering improved clock speeds and core utilization [10][11] - The Power11 architecture supports advanced memory configurations, including DDR5, which enhances performance and efficiency [12][20] System Architecture - The Power E1180 server, based on Power11, can support up to 256 cores and 64TB of memory, showcasing a balance between I/O and computational power [18][19] - The architecture allows for flexible memory configurations, enabling users to optimize performance based on their specific workloads [21][22]

Core Viewpoint - TSMC reported record profits in Q2, driven by strong demand for AI chips, with a projected revenue growth of nearly 30% for the year despite potential tariff impacts [2][3][11]. Financial Performance - TSMC's Q2 revenue reached NT$933.79 billion, a year-on-year increase of 38.6%, with net income of NT$398.27 billion, up 60.7% [2][4]. - Earnings per share (EPS) for Q2 was NT$15.36, reflecting a 60.7% increase compared to the previous year [5][11]. - The gross margin stood at 58.6%, with an operating margin of 49.6%, indicating strong operational efficiency [5]. Market Outlook - TSMC's chairman, C.C. Wei, indicated that AI chip demand remains robust, with non-AI applications also expected to recover moderately [3][11]. - The company has revised its full-year revenue growth forecast from 24-26% to nearly 30%, outperforming market expectations [12][22]. - Despite potential tariff uncertainties, TSMC's overall outlook remains positive, with no significant changes in customer behavior observed [3][4]. Capital Expenditure and Dividends - TSMC plans to maintain its capital expenditure forecast at US$38-42 billion for the year, ensuring sustainable dividend policies [4][13]. - The company anticipates a total cash dividend of at least NT$18 per share this year, increasing to at least NT$20 next year [4]. Advanced Technology and Production Capacity - TSMC is set to begin mass production of its 2nm process technology in the second half of the year, which is expected to provide significant revenue contributions [15][17]. - The company is expanding its manufacturing capabilities in Arizona, with plans for six wafer fabs and two advanced packaging facilities to support AI and HPC applications [18][19]. Industry Context - Analysts view TSMC's strong performance as a positive signal for the semiconductor market, particularly amid tariff and policy challenges [21][22]. - Concerns remain regarding potential impacts from tariffs and the performance of major clients like Apple, which could affect TSMC's outlook for the latter half of the year [22][23].

公众号记得加星标⭐️，第一时间看推送不会错过。 2025 年 7 月 16 日-18日，第五届 RISC-V 中国峰会在上海张江科学会堂盛大召开。在这场聚 焦 RISC-V 生态创新与未来发展的行业盛会上，隼瞻科技重磅发布两款革新性新品 ——DSA处 理器敏捷开发平台ArchitStudio 与 "智翼" 系列端侧AI融合解决方案。两款产品分别从DSA敏 捷开发与端侧应用落地两个维度，破解行业痛点，推动DSA处理器设计模式与智能终端算力边 界的双重革新。 DSA处理器敏捷开发平台—— ArchitStudio 随着人工智能、具身机器人、低空经济、HPC等领域的发展，DSA处理器的应用场景不断增加。在追 求极致效率与创新的垂直应用领域，DSA处理器凭借其卓越的性能功耗比，正成为推动技术革新的核 心引擎。然而，传统DSA处理器的设计之路却布满荆棘： 技术门槛高不可攀、项目复杂度指数级增 长、设计指标难以早期评估、实现端工作浩如烟海、人力与时间投入巨大…… 这些痛点严重阻碍了 创新想法的快速落地。 破局时刻已至！ 隼瞻科技全新发布的划时代解决方案——基于RISC-V架构的DSA处理器敏捷开发平 台 ArchitSt ...