如果您希望可以时常见面，欢迎标星收藏哦~ 过去40年来，这项技术助力科学和工程领域取得了重大发现。如今，高性能计算正处于一个转折 点，政府、研究人员和科技行业今天的选择可能会影响创新、国家安全和全球领导力的未来。 高性能计算系统本质上是由数千甚至数百万个处理器同时协同工作的超强计算机。它们还使用先进 的内存和存储系统来快速移动和保存大量数据。 凭借如此强大的能力，高性能计算系统可以运行极其精细的模拟和计算。例如，它们可以模拟新药 如何与人体相互作用，或者飓风如何穿越海洋。它们还应用于汽车设计、能源生产和太空探索等领 域。 近年来，高性能计算因人工智能而变得更加重要。人工智能模型，尤其是用于语音识别和自动驾驶 汽车等领域的模型，需要海量的计算能力进行训练。高性能计算系统非常适合这项工作。因此，人 工智能和高性能计算如今正紧密合作，相互促进。 可以明见，高性能计算系统面临着比以往任何时候都更大的压力，对系统的速度、数据和能耗提出 了更高的要求。与此同时，我也看到高性能计算面临着一些严峻的技术问题。 技术挑战 高性能计算面临的一大挑战是处理器速度与内存系统跟上处理器输出速度之间的差距。想象一下， 你有一辆超快的汽车 ...

Core Viewpoint - Arm is transitioning from a component IP supplier to a platform-focused company, emphasizing its capabilities in AI and energy-efficient chip design to capture market opportunities in the growing AI sector [1][4][5]. Group 1: Company Strategy and Product Development - Arm's new product naming strategy reflects its shift towards complete computing platforms, which facilitate faster integration and scalability for AI chip development [4][5]. - The company aims to address the increasing complexity and power demands of AI workloads by providing integrated, validated systems that enhance performance per watt and enable scalable innovation [4][5]. - Arm has introduced specific platforms for key markets, including Arm Neoverse for infrastructure, Arm Niva for PCs, and Arm Lumex for mobile devices, along with maintaining the Mali brand for GPUs [4][5][9]. Group 2: Financial Performance - In the fourth quarter of fiscal year 2025, Arm reported a record revenue of $1.24 billion, a 34% year-over-year increase, driven by record licensing revenue of $634 million (up 53%) and royalty revenue of $607 million (up 18%) [7]. - The growth in royalty revenue is attributed to the increasing deployment of the Armv9 architecture and the adoption of Arm's computing subsystems in smartphones, cloud infrastructure, and edge AI [7]. Group 3: Market Opportunities - Arm views the automotive sector as a major growth area, having signed its first automotive computing subsystem agreement with a leading electric vehicle manufacturer [7][8]. - The company is also expanding its influence in data center computing, with cloud providers like AWS, Google Cloud, and Microsoft Azure increasingly utilizing Arm-based chips for AI workloads [8]. Group 4: Ecosystem and Developer Support - Arm is enhancing its hardware platforms with expanded software tools and ecosystem support, including free access to GitHub Copilot for developers to optimize code for Arm architecture [8]. - Over 22 million developers are currently working on Arm, with the Kleidi AI software layer installed over 8 billion times on devices, indicating a robust developer engagement [8]. Group 5: Implications for AI and Data Management - The strategic repositioning of Arm is expected to reshape how technology decision-makers handle AI, data, and security, providing clearer pathways for selecting optimized computing architectures for AI workloads [11][12]. - The introduction of predefined computing systems like Neoverse and Lumex can streamline the evaluation process for engineers managing large language models and AI pipelines, enhancing efficiency and reducing overhead [11][12].

如果您希望可以时常见面，欢迎标星收藏哦~ 来源：内容来自半导体行业观察综合 ，谢谢 。 昨夜晚间，小米雷军发布了一项消息，表示小米自主研发设计的、名为玄戒O1的手机SoC芯片即 将在五月下旬发布。 按照雷军在微博中所说，小米的造芯始于2014年9月。 翻开资料，小米当时成立了一家名为北京松果电子的全资子公司。据甲子光年在报道中描述，对于 当时的情景，雷军有过这样的描述：绝大部分人心里都会七上八下，因为不知道冲出去是死是活。 而我心里稍微平静一些，因为我已经做好了干十年的准备。 到了小米于2017年2月28日，小米也终于发布了手机芯片的首个成果——松果澎湃S1手机芯片。作 为一颗定位为中端手机芯片，澎湃S1采用台积电28nm工艺，4+4大小核心全A53架构，大核主频 2.2GHz，小核主频1.4GHz，GPU为ARM Mali T860 MP4，基带为可升级设计。该芯片也首发搭 载于小米5C上。 而后，小米的手机AP几经沉浮，也没有下文。不过，公司在此期间还是发布了电源管理芯片、自 研影像芯片等一系列大家眼里的"小芯片"，松果也在接下来几年被拆分，直到2021年成立上海玄 戒，小米造芯，重回主干道。 和之前一样 ...

如果您希望可以时常见面，欢迎标星收藏哦~ 来源： 内容 编译自 semiengineering ，谢谢 。 Chiplet 带来了半导体功能和生产力的巨大飞跃，就像 40 年前的软 IP 一样，但要实现这一点，还 有很多工作要做。这需要一个生态系统，而目前这个生态系统还非常初级。 如今，许多公司已达到光罩极限，被迫转向多芯片解决方案，但这并没有催生一个即插即用的芯片 市场。这些早期系统无需遵循标准即可运行，也不追求相同的优势。从设计角度来看，它们仍在构 建一个大型系统。 西门子EDA公司Tessent硅片测试解决方案DFT流程产品经理Vidya Neerkundar表示："chiplet背 后的理念是分而治之。你可以以更快的速度完成任务，并获得更高良率的所有好处。但是，当你分 而治之时，你还需要考虑其他事情。你解决了一个问题，然后又必须解决其他问题。你需要不断推 进问题，不断追赶。" 对这些新问题的全面理解仍在不断发展。"我们知道如何制造标准的 chiplet，"Marvell 技术副总 裁兼定制解决方案首席技术官 Mark Kuemerle 说道。"它叫做 HBM，也是目前唯一的 chiplet。 它是在 ...

Group 1 - Cognichip has announced its establishment and secured $33 million in funding to develop an AI model called "Artificial Chip Intelligence" (ACI) aimed at reducing the manual workload in processor development [1] - The ACI model is expected to accelerate chip development and reduce costs by up to 75% [1] - The CEO of Cognichip emphasizes the need for a fundamental change in the semiconductor industry due to a significant decline in venture capital investment and a shrinking pool of engineers [26][27] Group 2 - The semiconductor industry has been a transformative force since its inception, impacting various sectors including communication, automotive, and healthcare [3] - In 2023, global semiconductor device sales approached 1 trillion units, equating to over 100 chips per person on Earth [3] - The demand for computing power is expected to grow exponentially over the next 10-15 years, driven by the rise of artificial intelligence [4] Group 3 - The semiconductor industry faces a severe talent crisis, with Deloitte estimating a need for over 1 million new technical workers by 2030 [4][7] - The shortage of engineers is exacerbated by an aging workforce, with a significant portion of current employees nearing retirement [7] - The complexity of semiconductor design requires years of specialized experience, making it difficult to quickly fill the talent gap [7] Group 4 - The lengthy time-to-market (TTM) for semiconductor chips has increased, with development cycles now taking several years [9][17] - High R&D costs are a significant burden, with companies like Qualcomm projected to spend nearly $8.9 billion on R&D in 2024 [17] - The industry is experiencing a "2.0 innovator's dilemma," where the high costs and lengthy development times hinder innovation and market responsiveness [12][13] Group 5 - Cognichip aims to create a foundational AI model that can help accelerate chip development processes, potentially reducing production time by 50% [25] - The company is led by experienced professionals from top tech firms, focusing on integrating AI into semiconductor design to address industry challenges [28][29] - The goal is to simplify chip design, making it more accessible for smaller companies and fostering innovation in the semiconductor sector [29][41]

Chan表示："不过，我们认为，随着 2025 年下半年和 2026 年先进工艺节点支出的回升，该公司 可以克服这一不利因素。" 如果您希望可以时常见面，欢迎标星收藏哦~ 来源：内容综合自路透社等 ，谢谢 。 受出口限制不确定性的影响，芯片设备制造商应用材料公司最大业务部门的销售额低于预期，该公 司周四公布的第二季度营收未达华尔街预期。 这家总部位于加利福尼亚州圣克拉拉的公司股价在盘后交易中下跌了 5% 以上。 根据伦敦证券交易所汇编的数据，半导体系统部门的收入为 52.6 亿美元，低于分析师估计的 53.2 亿美元，而半导体系统部门是该公司总收入的最大贡献者。 应用材料公司指出，ICAPS 市场（涵盖物联网、通信、汽车、电力和传感器）的投资有所放缓， 但表示这已被先进芯片的巨额投资部分抵消。 Summit Insights Group 高级研究分析师 Kinngai Chan 表示："显然，美国政府对中国某些半导 体设备的出口限制正在影响公司的业绩和前景。" 美国政府去年 12 月曾表示，将对向中国（应用材料公司最大的海外市场）出口生产先进节点芯片 所需的半导体制造设备实施新的管制。 该公司来自中国的收入占 ...

Core Viewpoint - Nvidia has postponed the commercialization of its next-generation low-power DRAM module, SOCAMM, shifting its focus from the initially targeted Blackwell series to the Rubin series, in response to challenges in ensuring stable production and performance amid the rapid advancements in AI accelerator technology [1][4][5]. Summary by Sections SOCAMM Development and Supply Chain - SOCAMM is designed to integrate four LPDDR low-power DRAMs to enhance power efficiency and is removable for easier upgrades and maintenance [1]. - Major memory partners, including Samsung, SK Hynix, and Micron, have adjusted their SOCAMM supply plans following Nvidia's notification about the change in application timeline [1][4]. Changes in GB300 Design - The GB300, initially set to incorporate SOCAMM, has undergone a design change from the Cordelia motherboard to the Bianca design, which does not support SOCAMM and instead uses existing LPDDR memory [2][3]. - The Cordelia design was intended to support two Grace CPUs and four Blackwell GPUs, while the Bianca design only supports one Grace CPU and two Blackwell GPUs [3]. Reliability and Production Challenges - The switch to the Bianca design was influenced by reliability issues with the Cordelia design, which was prone to data loss, and similar reliability concerns with SOCAMM, including heat dissipation problems [3][4]. - Nvidia is facing supply chain difficulties, impacting its ability to manage production yields, prompting a reliance on existing technologies to mitigate these issues [3][5]. Market Impact and Future Prospects - The postponement of SOCAMM is seen as a strategy to avoid excessive technological advancement while ensuring product reliability and timely market entry [4]. - Micron's SOCAMM modules, which are now in mass production, are expected to provide significant performance and energy efficiency benefits for AI training and inference applications [6][8].

END 如果您希望可以时常见面，欢迎标星收藏哦~ 来源：内容来自人民日报 。 2025年5月14日，吉林省吉林市中级人民法院一审公开宣判紫光集团原董事长赵伟国贪污、为亲友非法牟 利、背信损害上市公司利益案，对被告人赵伟国以贪污罪判处死刑，缓期二年执行，剥夺政治权利终身， 并处没收个人全部财产；以为亲友非法牟利罪判处有期徒刑五年，并处罚金人民币一千万元；以背信损害 上市公司利益罪判处有期徒刑三年，并处罚金人民币二百万元；决定执行死刑，缓期二年执行，剥夺政治 权利终身，并处没收个人全部财产。对追缴在案的赵伟国违法所得返还被害单位。 图源：人民日报 经审理查明：2018年至2021年，被告人赵伟国利用担任紫光集团董事长的职务便利，与特定关系人李禄 媛共谋，由李禄媛实际控制的公司低价购买原本应当由紫光集团购买的房产，获取房产溢价利益，非法占 有国有资产价值人民币4.7亿余元。2014年至2021年，赵伟国利用担任紫光集团董事长等职务便利，将本 单位的盈利业务交由李禄媛等特定关系人经营，或者以明显高于市场的价格向李禄媛经营管理的公司购买 代建管理服务，造成国家直接经济损失人民币8.9亿余元。2019年，赵伟国还指使其实 ...

该公司在一份新闻稿中宣布了其2260万英镑的A轮融资，这是英国太空科技公司A轮融资金额最高 的一次。该公司表示，这笔资金将加速其第二代卫星ForgeStar-2的研发，并促进其首颗制造卫星 ForgeStar-1的发射。 如果您希望可以时常见面，欢迎标星收藏哦~ 来源 ：本文 编译自 tomshardware ，谢谢。 英国初创公司 Space Forge 已获得 3000 万美元的 A 轮融资，以支持其创建和发射制造卫星的努 力，该卫星可用于制造用于半导体和量子计算的独特材料。 Space Forge 的愿景是，通过利用太空的独特条件（包括微重力、真空和极端温差），该公司"正 在释放制造地球上无法生产的材料的能力"。 Space Forge 表示，这些进展"将在半导体、量子计算、清洁能源和国防技术领域具有广泛的应 用"。此外，该公司表示，这些任务"将展示一个可扩展、可重复使用的平台，用于在太空中制造高 性能材料，为安全、清洁能源和基础设施带来突破。" Space Forge 在其他地方表示，这些材料"有望在高价值领域带来显著的性能提升"，同时还可能将 包括数据中心在内的基础设施的二氧化碳排放量减少高达 75 ...

Core Viewpoint - The article highlights the advancements and strategic focus of Naxin Micro in the semiconductor industry, particularly in the signal chain products for automotive and energy sectors, emphasizing their approach to system-level solutions rather than just component specifications [2][4][11]. Group 1: Product Innovations and Market Positioning - Naxin Micro showcased its new magnetic sensor series and real-time control MCU NSSine™ series at the Munich Shanghai Electronics Show, demonstrating its comprehensive semiconductor solutions across various sectors including automotive electronics and renewable energy [1]. - The company has made significant breakthroughs in technology innovation, market expansion, and industry chain collaboration, particularly in the automotive chip market, which is increasingly penetrating intelligent and safety domains [2][4]. - Naxin Micro's strategy focuses on specific customer needs and application scenarios, aiming to excel in targeted markets rather than attempting to cover all areas indiscriminately [4][5]. Group 2: Competitive Advantages and Customer Engagement - The company emphasizes the importance of understanding customer systems and addressing their specific pain points, which allows them to optimize chip design and enhance reliability in real-world applications [5][6]. - Naxin Micro is shifting from a parameter-driven market to an application-driven market, focusing on high-performance signal chain products that meet the stability and adaptability requirements of automotive and energy sectors [6][11]. - The company has developed a dedicated sales and technical support team to assist clients in utilizing their products effectively, addressing concerns about using domestic chips [9][10]. Group 3: Industry Trends and Future Directions - Amidst global semiconductor supply chain disruptions, Naxin Micro's signal chain products are gaining traction as domestic alternatives, with many clients incorporating them into their official BOMs [9]. - The company is committed to serving the main market segments of renewable energy, industrial applications, and automotive, where high stability and performance are critical [10][11]. - Naxin Micro's approach reflects a broader industry shift from merely providing components to offering comprehensive solutions that enhance system performance and reliability [11].