2025年的二季度已经结束，GB200也已经开始大规模的出货，各个机构也调高了GB200的出货预期。用一句话总结今年的AI服务器市场就是：NVIDIA 领跑，ASIC崛起。 根据Bernstein的数据，2024-2026年，全球服务器市场预计以3%的年复合增长率（CAGR）稳步增长，到2026年规模将逼近4000亿美元。而AI服务器无疑 是这波增长的核心动力，其在整体服务器市场的占比将从2024年的低个位数跃升至2026年的高个位数。 具体来看，2024年全球服务器出货量同比增长4%，未来两年增速略有放缓，但AI服务器的出货量却保持两位数增长。特别是高端GPU服务器（每台配 备8个以上GPU，比如NVIDIA的GB200和GB300机架），2025年预计增长超50%，2026年也有低20%的增幅。2025年全球将部署约450万个NVIDIA GPU 芯片，这数字直接反映了AI服务器需求的火爆。 更有意思的是，高端AI服务器的平均售价（ASP）因NVIDIA下一代Rubin芯片的引入而水涨船高。这不仅推高了市场规模，也让整个供应链嗅到了巨大 商机。 NVIDIA服务器 GB200在2025年第二季度开始量产， ...

Group 1 - Industrial Fulian's preliminary forecast for Q2 2025 indicates a net profit of 6.727 billion to 6.927 billion yuan, representing a year-on-year increase of 47.72% to 52.11% [2] - The company reported that AI server revenue grew over 60% compared to the same period last year, highlighting its leading position in core product shares among major clients [2] - Despite the traditional off-season for consumer electronics in Q2, the strong growth in AI contributed to a significant overall profit increase [2] Group 2 - The AI industry chain, including sectors like server ODM, PCB, and IC manufacturing, is experiencing high growth rates, with some companies seeing accelerated performance in recent months [3] - Demand for ASICs is also rising, with Broadcom holding a 55% to 60% market share, and its financial reports indicate that ASIC business is a core growth driver [3] - Google has launched its seventh-generation TPU, which can achieve computing power 24 times that of the fastest supercomputer globally, indicating strong demand for AI accelerators [3] Group 3 - The A-share market is witnessing rapid growth in sectors like PCB and optical modules, alongside a new wave of investment expansion [5] - Companies like Huadian Co. are planning significant investments, such as 3.6 billion yuan for new projects, to enhance production capacity [5] - The AI wave is expected to continue driving growth in the industry chain, with recommendations for strategic investments in related sectors [5]

Google（谷歌） 今年受惠主权云项目以及东南亚新数据中心落成，显著提升Server需求。此外， Google本是自研芯片布局比例较高的业者，其针对AI推理用的TPU v6e已于上半年逐步放量成为主 流。 AWS（亚马逊云科技） 的自研芯片目前以Trainium v2为主力平台，据TrendForce集邦咨询了解， AWS已启动不同版本的Trainium v3开发，预计于2026年陆续量产。受惠于Trainium平台扩充与AI运 算自研策略加速，预估2025年AWS自研ASIC出货量将达双倍成长，为美系CSP最强。 和前四大CSP相比， Oracle（甲骨文） 更着重采购AI Server与IMDB（ In-Memory Database ） Server。该公司今年将更积极布局AI Server基础设施，除整合自家核心业务云端数据库及AI应用外， 针对美国等主权云项目，其对NVIDIA GB Rack NVL72需求也明显提升。 此外，近期因国际形势变化，多数Server Enterprise OEM重新检视2025年下半年市场规划， 目前 TrendForce集邦咨询预估全年整体Server（ 含通用型 ...

Core Viewpoint - The article discusses the increasing competition in the AI chip market, particularly how major tech companies like Google and Meta are accelerating their development of custom chips to reduce reliance on Nvidia's GPUs, with predictions that ASIC shipments will surpass Nvidia's AI GPU shipments as early as next year [1][2]. Group 1: Market Dynamics - Nvidia has historically dominated the AI chip market, holding over 80% market share in AI servers, while ASIC-based servers currently account for only 8% to 11% [2][4]. - Google is expected to ship between 1.5 million to 2 million of its self-developed AI chips (TPUs) this year, while Amazon's AWS is projected to ship 1.4 million to 1.5 million ASICs, bringing their combined shipments close to half of Nvidia's estimated annual GPU shipments of 5 million to 6 million [2][4]. Group 2: Cost Efficiency - The key advantage driving tech giants to develop their own chips is the reduction in Total Cost of Ownership (TCO), with ASICs potentially saving 30% to 50% in TCO compared to GPUs [3]. - Google claims its TPUs can deliver three times the performance of Nvidia GPUs per unit of energy consumed, highlighting the efficiency of custom chips [3]. Group 3: Competitive Landscape - Meta is focusing on launching its new high-performance ASIC chip "MTIA T-V1" in Q4 of this year, aiming to outperform Nvidia's next-generation AI GPU "Rubin" [5]. - Despite ambitious plans, Meta faces production challenges due to limited advanced packaging capacity from TSMC, which can only provide 300,000 to 400,000 units, creating a bottleneck [5]. Group 4: Nvidia's Response - In response to the competitive threat, Nvidia has opened its proprietary "NVIDIA NVLink" communication protocol to facilitate integration with other companies' CPUs or ASICs, aiming to retain its major clients [6]. - Nvidia's established software ecosystem, CUDA, remains a significant barrier for competitors, as it allows AI developers to efficiently build and deploy applications, maintaining Nvidia's competitive edge [6].

Al Semi EQUITY: TECHNOLOGY Meta's ambition on ASIC Al server - 2 Opportunities and challenges: complicated ASIC design with major volume mismatch in upstream/downstream Meta's aggressive ASIC roadmap to generate new business opportunities for semi supply chain despite significant volume mismatch in upstream and downstream We had highlighted that CSPs (cloud service provider) are getting increasingly aggressive in deploying their own AI ASIC solutions, such as Meta (META US, Not rated) from 2026F in our Asia ...

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 野村 。 据报道，追随谷歌的脚步，Meta 正在全面进入 ASIC（专用集成电路）竞赛。根据野村证券分析师 Anne Lee 及其团队的最新研究报告，Meta 在 AI 服务器领域的雄心正在迅速升级。他们专有的 ASIC 服务器项目 MTIA 预计将在 2026 年实现重大突破，这可能会挑战 Nvidia 长期以来的市场主 导地位。 报 告 援 引 最 新 的 供 应 链 信 息 称 ， Meta 计 划 在 2025 年 底 至 2026 年 期 间 推 出 数 百 万 颗 高 性 能 AI ASIC 芯片（100 万至 150 万颗）。与此同时，Google 和 AWS 等云服务巨头也在加速部署其自研 ASIC。这表明 AI 服务器市场的竞争格局正在悄然转变。 ASIC 芯片热潮：明年出货量可能超过英伟达 报告显示，英伟达目前持有 AI 服务器市场价值的 80% 以上，而 ASIC AI 服务器仅占 8-11%。 然而，从出货量的角度来看，情况正在发生变化。到 2025 年，谷歌的 TPU 出货量预计将达到 15- 200 万台，亚马逊 AW ...

| 研究所： | 证券分析师： | 李永磊 | S0350521080004 | | | --- | --- | --- | --- | --- | | liyl03@ghzq.com.cn | 证券分析师： | 董伯骏 | S0350521080009 | | | dongbj@ghzq.com.cn | 联系人 | ： | 王鹏 | S0350125040009 | | wangp05@ghzq.com.cn | | | | | [Table_Title] "20"家族亮相巴黎航展，朱雀三号完成一子级 动力系统试车 2025 年 06 月 22 日 行业研究 评级：推荐(维持) ——新材料产业周报 最近一年走势 | 行业相对表现 | | | 2025/06/20 | | --- | --- | --- | --- | | 表现 | 1M | 3M | 12M | | 基础化工 | -2.4% | -3.2% | 8.8% | | 沪深 300 | -1.3% | -3.2% | 9.8% | 相关报告 《高温合金行业动态研究：全球燃气轮机和航空发 动机需求大增，关注铬盐和高温合金行业（推荐） *基础化工* ...

1) AWS Trainium: it remains a global debate whether Marvell will win part of the Trainium3 business (link). At least our checks suggest that Alchip taped out the Trainium3 design back in February, and the wafers were already out in May. Charlie Chan thinks Alchip has a higher chance to win the 2nm Trainium4, which should be decided this summer. Astera Labs and Alchip just formed a partnership in connectivity chip design (I/O chiplets). We think that may help Alchip compete for next-generation XPU ASIC proje ...

Asia AI Server Global Markets Research EQUITY: TECHNOLOGY Meta's ambition on ASIC AI server Meta's MTIA AI servers could mark a milestone in 2026F AI ASIC is booming; a potential cross over in units from 2H26-2027 In the world of AI servers, nVidia (NVDA US, Not rated) has been dominating 80%+ market "value" so far, vs. ASIC AI servers that have roughly 8-11% in value share (based on Bloomberg consensus estimates). However, if we just compare the number of units for AI ASICs vs. nVidia's AI GPUs (which can ...

J P M O R G A N North America Equity Research 15 June 2025 Semiconductors AVGO and MRVL Continue to Dominate ASICs - Both Secure Next-Gen 2nm AI ASIC Programs; Overall Advanced Custom Chip (ASIC) Market Opportunity of $30B in CY25 and 30%+ Growth Outlook We are increasing our high-end custom ASIC chip market outlook to $30B (from our prior outlook of ~$25B ) growing at a 30% CAGR driven by the continued aggressive adoption of AI (training and inference), which is driving accelerating demand for custom chip ...