Core Viewpoint - Nvidia's introduction of NVLink Fusion aims to enhance flexibility and customization in AI infrastructure while maintaining its technological advantage in the market [8][17]. Group 1: Nvidia's Development and Products - Nvidia has evolved from focusing on GPUs to becoming a giant in AI infrastructure, with significant milestones such as the launch of CUDA in 2006 [1]. - The GB300 chip, set to launch in Q3, boasts a 1.5x improvement in inference performance, HBM memory, and a 2x increase in network bandwidth, while maintaining physical compatibility with previous generations [6]. - The Project DIGITS personal AI computer, DGX Spark, is now in full production, with availability expected by Christmas [6]. Group 2: NVLink Fusion Technology - NVLink Fusion extends Nvidia's NVLink technology to third-party CPUs and accelerators, allowing for a more open ecosystem while still requiring Nvidia chips in the system [8][10]. - The technology includes two components: a semi-custom CPU connection via NVLink C2C and the integration of NVLink 5 Chiplet into third-party accelerators [9][10]. - NVLink Fusion is designed as a "either/or" technology, allowing for either a semi-custom CPU or GPU but not both simultaneously, ensuring Nvidia's presence in the system [10]. Group 3: Market Implications and Partnerships - Current partners for NVLink Fusion include Alchip and AsteraLabs, with Fujitsu and Qualcomm developing new CPUs compatible with Nvidia GPUs [11]. - The limited openness of NVLink Fusion may accelerate diversification in AI computing infrastructure and provide pathways for third-party chips to enter the high-performance computing market [11][17]. - Nvidia's strategy reflects an understanding that a fully closed NVLink could limit market expansion, particularly among cloud service providers and sovereign AI projects [17]. Group 4: NVLink Advantages - NVLink 5 offers a dual bandwidth of 1.8 TB/s, significantly outperforming PCIe 5.0, which is crucial for scaling AI model training and inference [20]. - The NVLink Switch chip enables rack-level scalability, supporting up to 72 GPUs with a total bandwidth of 130 TB/s, a capability that competitors struggle to match [20]. - The integration of NVLink with Nvidia's SHARP protocol and Mission Control software optimizes AI workload throughput and latency, enhancing overall performance [20].

Core Viewpoint - Citi believes that 2026 will be the year when 800G optical modules dominate the market, with a projected sales volume of 37 million units, representing an 85% year-on-year growth driven primarily by demand from overseas cloud service providers [1][3]. Group 1: Market Demand and Supply Dynamics - The demand for 800G optical modules is expected to be robust, with over 90% of this demand coming from overseas cloud service providers (CSPs) [3]. - The deployment speed of 1.6T Ethernet switches may slow down, leading to a downward revision of demand forecasts for 1.6T modules [3]. - The market share distribution will be influenced by production capacity and delivery capabilities, with a preference for second-tier suppliers among overseas customers [3]. Group 2: Company Performance and Predictions - Citi's rating hierarchy for companies in the sector is as follows: Eoptolink > Innolight > T&S > TFC, with a particular focus on Eoptolink due to its strong performance in 800G/1.6T products and capacity [4]. - Eoptolink is expected to benefit significantly from the strong demand for 800G, with projections indicating that 75% of its sales will come from this segment by 2026 [15]. - Innolight is also anticipated to capture a substantial share of the 800G demand due to its superior supply capabilities and ongoing silicon photonics migration [17][18]. Group 3: Financial Forecasts and Market Trends - The 2025-2027 shipment forecasts for 400G, 800G, and 1.6T modules have been adjusted, reflecting an increase in 800G shipments and a decrease in 1.6T due to industry demand delays [17]. - The overall market valuation is expected to recover, with industry price-to-earnings ratios projected to rise from 8-10 times to 15-20 times by 2025, driven by cloud infrastructure upgrades and higher optical module integration rates [9]. - The anticipated strong demand for 800G and the potential delay in 1.6T migration may pose risks for CPO (Co-Packaged Optics) deployment, which could be pushed to 2027 [9]. Group 4: Competitive Landscape - The competitive landscape is shifting, with more second-tier suppliers entering the supply chain due to insufficient supply from first-tier suppliers [13]. - Companies like Suzhou Taicheng Light are facing challenges due to lower-than-expected InfiniBand penetration and weak GB200 rack numbers, which may impact their 1.6T demand [20][21]. - The ongoing silicon photonics migration is expected to provide cost advantages for companies like Innolight and Eoptolink, allowing them to maintain higher profit margins compared to competitors who rely on external design sources [18].

还有专注于半导体行业分析的SemiAnalysis的全部分析报告： 星球中每日还会更新Seeking Alpha、Substack、 stratechery的精选付费文章， 现在星球中领券后只需要 390元，即可每天都能看到上百篇外资顶尖投行科技行业的分析报告和每天的精选报告，无论是我们自 己做投资，还是对行业有更深入的研究，都是非常值得的。 想要看外资研报的同学，给大家推荐一个星球，在星球中每天都会上传几百篇外资顶尖投行的原文研 报：大摩、小摩、UBS、高盛、Jefferies、HSBC、花旗、BARCLAYS 等。 ...

这篇文章，我们结合国盛证券的一份研报，来看下华为昇腾产业链上的公司们。主要讲下面4个产业链 的方向：整机、电源、散热和连接。本文中所有的图片都是来自国盛证券。 昇腾整机硬件伙伴要求拥有自有品牌 产品，能在昇腾产品基础上二次开发或加工生产，并销售与服务 至最终用户，伙伴类型 分领先级、优选级、认证级。 昇腾整机硬件伙伴： 一、整机 根据国盛证券，2024年新增算力规模约为2万Pflops，2028年中国智算中心市场投资规模有望达到 2886 亿元。2023年中国智算中心市场投资规模达879亿，同比增长 90%以上。未来，AI大模型应用场景不断 丰富，商用进程加快，智算中心市场增长动力 逐渐由训练切换至推理，市场进入平稳增长期，预计 2028 年中国智算中心市场投资规 模有望达到2886亿元。截至2024年8月，中国智算中心项目超过300 个，已公布算力 规模超50万PFlops。从已投用、在建、规划的智算中心项目来看，全国各省智算中心 总 计300 余个，约三分之一智算中心项目规划算力大于500PFlops，主要为政府或基础电 信运营商投建 项目，2024年当年投运项目数量超过50个，60%以上为地方政府、国资 ...

前几天第一时间转载了路透社的报道， 外媒：H20阉割版预计在7月推出，性能或大幅缩水 。 根据这个 报道，英伟达计划将在未来2个月推出H20的阉割版，网友调侃这是"丐中丐"版。Jefferies也第一时间对 这款"丐中丐"做解读分析。 报告的核心观点为： H20 因内置 HBM3 内存易受限，总带宽 4.0TB/s 。美国或设 GPU 内存带宽上限 1.7-1.8TB/s ，若如 此英伟达 H20 可能弃用 HBM 内存改用 GDDR6 ，降级版 H20 性能仍可能强于使用 GDDR6 的游戏 GPU ，如 RTX5090D 报告内容 Jefferies在之前的研究中，就提到过H20 容易受到限制，因为它内置了 HBM3 内存，其总内存带宽为 4.0TB/s（高于 H800）。去年 12 月底，拜登政府对向中国出售独立的 HBM3 及以上产品实施了限制。 科技媒体Tom's hardware上周也披露，英伟达在中国已停止接受游戏GPU RTX5090D 的订单， RTX5090D 内置 32GB 的 GDDR7 内存，总带宽为 1.79TB/s。因此，当英伟达在 4 月 16 日宣布对 H20 的限制措施（需 ...

黄仁勋还希望接触中国本土的顶尖人工智能人才。目前，英伟达正在招聘上海的职位，包括帮助 "指导 下一代深度学习硬件和软件研发" 的工程师，以及 "开发和优化具有全球竞争力的 ASIC 设计" 的岗位。 其中一位知情人士称，上海市政府已对该计划表示初步支持，而英伟达正在游说美国政府批准。这家硅 谷公司在上海约有 2000 名员工，主要从事销售和相关支持职能。 英伟达正在扩大其在中国的研究布局，以试图在其最大的海外市场之一保持领先地位。该公司担心，以 华为为首的中国本土竞争对手可能通过提供 rival AI 生态系统占据市场。 根据Financial Time的消息，英伟达计划在上海设立研究中心，以示对中国的新承诺。 英伟达（Nvidia）正寻求在上海建立一个研发中心，以帮助这家全球领先的人工智能处理器制造商在中 国保持竞争力。由于美国出口管制收紧，该公司在华销售额已大幅下滑。 据两位知情人士透露，英伟达首席执行官黄仁勋（Jensen Huang）上个月在上海会见上海市市长龚正 时，讨论了这一计划。目前，英伟达正在上海租赁新办公空间，以容纳现有员工并为潜在的扩张做准 备。 知情人士称，该研发中心将研究中国客户的特定 ...

Core Viewpoint - The recent $200 billion agreements between the U.S. and the UAE may not be as beneficial as anticipated, reflecting a disparity between idealistic expectations and practical realities [2][3]. Group 1: Agreements and Impacts - The U.S. has signed two significant agreements with the UAE and Saudi Arabia, which are expected to reshape the power dynamics in AI, with implications for economic, geopolitical, and national security [5][6]. - The agreements are projected to unlock a trillion-dollar capital influx, benefiting both the Gulf region and the U.S. by enhancing AI infrastructure and alleviating power bottlenecks [6][18]. - The UAE's G42 is set to lead a 5GW data center project, with the first phase of 1GW already underway, indicating a strong commitment to AI infrastructure development [10][12]. Group 2: Geopolitical and Economic Considerations - The agreements deepen the technological ties between the UAE, Saudi Arabia, and the U.S., potentially increasing the region's dependency on American hardware and software [6][10]. - The Gulf region is expected to emerge as a new AI hub, with predictions indicating that by 2030, the Middle East's operational data center capacity will exceed 6GW [6][40]. - The capital from the Gulf is anticipated to flow into U.S. AI infrastructure, with significant investments from companies like Datavolt and HUMAIN, totaling hundreds of billions [16][20]. Group 3: Risks and Challenges - Concerns exist regarding the reliability of past commitments from Gulf states, with previous projects often failing to materialize as planned due to political and economic fluctuations [3][27]. - There are significant security risks associated with the transfer of GPUs to the UAE, including potential unauthorized use and the risk of technology transfer to China [27][28]. - The agreements necessitate stringent security measures to ensure that GPU resources are not misappropriated, with proposals for physical inspections and robust KYC protocols [28][29][30]. Group 4: Infrastructure and Capacity - The Middle East's data center market is currently dominated by G42, which is expected to expand rapidly due to the new agreements, with U.S. hyperscale companies increasing their investments [13][15]. - The region's energy resources, including solar, natural gas, and nuclear, will support the development of AI infrastructure, although challenges related to cooling costs and skilled labor shortages remain [14][42][44]. - The U.S. is facing a data center capacity shortage, with predictions of over 1GW of power shortfall by 2026, creating opportunities for Middle Eastern investments to fill this gap [38][40].

关于GB300，尽管市场信息显示其将沿用GB200的Bianca主板而非新型Codelia主板，但根据 Jefferies 与供应链的沟通，这一计划尚未最终确定。目前仍存在多种并行解决方案。预计GB300的时间表保 持不变，即2025年第四季度小批量生产，2026年第一季度进入量产阶段。 根据Jefferies的一份研报，GB200在2025年第一季度的总出货量仅为约1,500台，远低于最初预测的 3,800台。对于2025年第二季度，Jefferies已将预估出货量从之前的7,200台下调至6,000台。同时，将 2025年GB200的出货量预测调整为：NVL72机型24,000台，NVL36机型12,000台。 GPU+FPGA 我们这有H200/B200/RTX5090/FPGA不错的资源，有兴趣的朋友可以加微信或者进小程 序商城选购： 知识星球 Jefferies预计，良率将从4月开始提升，关键组件短缺问题也将自4月起逐步缓解。2025年第二季度 GB200的出货量将显著高于第一季度。广达（Quanta）公布的4月销售额为新台币1,540亿元，环比下 降20%，但同比增长58%，占富邦证券（Fubon ...

今年受美国的各种不确定政策，导致很多科技企业都开始回调。今天看到一篇The Information的文章， 认为台积电目前虽然还面临些地缘政治风险，但长期看还是具有投资价值，这篇文章我们来看下The Information给的观点。 原文的链接如下：https://www.theinformation.com/articles/tsmc-critical-ai-shares-sale?rc=lr2ufv 人工智能的发展高度仰仗英伟达的芯片技术，而英伟达的芯片制造又离不开台积电的支持。从长期投资 视角来看，台积电当前展现出比英伟达更具吸引力的布局价值。 这两家科技巨头在产业链与资本市场形成紧密联动。受人工智能浪潮推动，双方股价曾同步飙升，但今 年均遭遇回调——部分源于特朗普政府关税政策的不确定性。随着本周特朗普宣布暂停贸易争端，市场 情绪回暖带动芯片股回升，台积电单周涨幅近9%，但两家公司年内累计跌幅仍超过标普500指数基准， 反映出市场对半导体板块的集体估值修正。 由于业务的高度关联性，投资者可能低估了台积电的结构性优势。当前估值水平、关税政策影响有限以 及行业竞争格局演变，共同构成了其利好因素。桑伯格投资管 ...

Core Insights - Siemens EDA is a global leader in Electronic Design Automation software, enabling faster and more cost-effective development of innovative electronic products [1] - The company emphasizes the importance of diversity and equality in its workforce, with over 377,000 employees across more than 200 countries [4] - Siemens Software offers flexible working arrangements and a comprehensive benefits package, including competitive salaries and private healthcare [5] Responsibilities - Conduct FPGA-based ASIC prototype bringup by porting ASIC RTL code to proFPGA platforms, collaborating with hardware and software teams for system integration [6] - Optimize FPGA resource utilization and timing performance to resolve technical bottlenecks, integrating high-speed interface modules [6] - Explore FPGA applications in AI acceleration, 5G communication, and autonomous driving while innovating prototyping methodologies [6] Qualifications - A Bachelor's or Master's degree in electrical engineering, Computer Science, or related fields is required, along with 3+ years of FPGA experience [6] - Proficiency in Verilog/VHDL and familiarity with complex logic design, as well as experience with tools like Vivado or Quartus [6] - Strong analytical problem-solving abilities and customer-oriented skills are essential, with limited business travel expected [6]