通信行业动态点评 GTC 盛大召开，光铜互连多路线并进 2026 年 03 月 20 日 【行业动态】 【投资建议】 英伟达 GTC 大会展示了未来两年公司存算与网络产品的迭代方向，我 们认为各互联路径均有望受益，中短期内铜连接方案可凭成本与稳定 性等优势为终端客户所接受，从长期来看 CPO/NPO 等光连接方案也 将保持逐步渗透的趋势。建议持续关注数据中心光通信（中际旭创、 新易盛、光迅科技、华工科技、天孚通信、仕佳光子、光库科技、嘉元 科技、优迅股份等）、铜互连（兆龙互连、瑞可达等）、交换机（锐捷 网络、菲菱科思等）。 【风险提示】 客户集中度与订单履约风险、技术迭代风险、海外投资风险。 挖掘价值 投资成长 强于大市（维持） 东方财富证券研究所 证券分析师：马成龙 证书编号：S1160524120005 联系人：陈力涵 相对指数表现 -20% 4% 28% 52% 76% 100% 2025/3 2025/9 2026/3 通信 沪深300 相关研究 《英伟达投资 40 亿美金加码光通信产业， 博通发布单通道 400G DSP》 2026.03.17 《AI 算力投资与变革加速，智谱 GLM-5 适 配主 ...

Core Viewpoint - Nvidia has seen a remarkable increase in its stock price, rising 22,690% over the past decade, establishing itself as a leading company in the AI infrastructure sector [1] - The potential for Nvidia's stock to reach $500 hinges on continued significant investments in AI and the ability of its customers to generate adequate returns from these investments [3] Group 1: Market Performance and Projections - Nvidia shares would need to increase by 173% from a recent price of $183 to reach $500, which is considered a possibility if AI-related investments persist [3] - Nvidia's CEO announced expectations of $1 trillion in orders for its new architectures through 2027, doubling previous forecasts, indicating strong future demand [5] - Wall Street analysts project Nvidia's revenue and earnings per share to grow at compound annual rates of 36.5% and 39.4% respectively over the next three fiscal years, suggesting a positive outlook for the company [6] Group 2: Market Sentiment and Valuation - Nvidia is currently trading at a forward price-to-earnings ratio of 22.5, which may indicate that the market is underestimating the company's potential [6] - Despite concerns about potential slowdowns in AI spending, the overall sentiment remains bullish regarding Nvidia's growth prospects over the next five years [8][10] - Investors are encouraged to maintain a long-term perspective, as the belief in continued AI progress is crucial for Nvidia's stock price to reach $500 [9][10]

Core Insights - Nvidia's CEO Jensen Huang proposed a new compensation model for engineers that includes AI tokens in addition to their base salary, aimed at enhancing productivity through the deployment of AI agents [2][3] - Huang envisions a future workplace where engineers manage numerous AI agents capable of performing complex tasks autonomously, indicating a significant shift in workforce dynamics [3][4] Company Perspective - Nvidia plans to offer engineers a compensation structure where they could earn up to half of their base salary in AI tokens, which are intended to incentivize productivity [3] - The company is positioning itself at the forefront of AI integration in the workplace, with Huang stating that Nvidia will have hundreds of thousands of digital employees alongside its 42,000 biological employees [4] Industry Trends - Goldman Sachs estimates that AI could automate tasks accounting for 25% of all work hours in the U.S., potentially leading to a significant displacement of jobs [6] - The software industry is expected to see increased demand due to AI agents, as they will require more underlying software infrastructure [8][9] - A paradigm shift is occurring in software development, allowing engineers to communicate with computers in plain English rather than programming languages, significantly reducing project timelines [10] Labor Market Dynamics - There is a "talent paradox" in the job market, where a majority of executives anticipate headcount reductions due to AI while simultaneously facing talent shortages [11] - Entry-level jobs are particularly vulnerable to AI displacement, as AI takes over tasks traditionally used for training new workers [12] - Despite potential job losses, new job opportunities are expected to arise in sectors that did not exist previously, driven by technological advancements [16] Challenges in AI Integration - Integrating AI into existing corporate workflows may be challenging, with a high failure rate of AI projects reported since 2018 [17]

When Jeff Bezos said that one breakthrough technology would shape Amazon’s destiny, even Wall Street’s biggest analysts were caught off guard. Fast forward a year and Amazon’s new CEO Andy Jassy described generative AI as a “once-in-a-lifetime” technology that is already being used across Amazon to reinvent customer experiences. At the 8th Future Investment Initiative conference, Elon Musk predicted that by 2040 there would be at least 10 billion humanoid robots, with each priced between $20,000 and $25,000 ...

相较GB300混合冷却，Rubin实现100%全液冷+微通道冷板升级+无缆线无软管模块化连接。在维护难度 和泄漏风险下降的同时，单机柜液冷部件价值量提升，行业价值中枢有望同步抬升。微通道技术壁垒有 望为国产厂商带来量价齐升机遇。 英伟达供应链开放，国产液冷迎来黄金切入窗口 智通财经APP获悉，国联民生证券发布研报称，英伟达(NVDA.US)2026财年Q4及全年业绩大幅超越市 场预期，GB300液冷深度渗透+下一代Rubin平台100%强制液冷，单机柜功率密度大幅跃升，彻底将液 冷从"可选配置"升级为"AI算力基础设施标配"。英伟达下放冷板、CDU等核心部件采购权，在液冷技术 迭代需求压力下，开放趋势预期长期延续，具备微通道冷板、高密度CDU等核心技术储备的国产厂 商，正迎来入围英伟达合格供应商名单的历史性机遇。 国联民生证券主要观点如下： 英伟达业绩超预期，夯实液冷行业高景气 英伟达2026财年Q4及全年业绩大幅超越市场预期，数据中心营收占比高达91.5%。GB300液冷深度渗透 +下一代Rubin平台100%强制液冷，单机柜功率密度大幅跃升，彻底将液冷从"可选配置"升级为"AI算力 基础设施标配"，黄仁勋 ...

智通财经APP获悉，国联民生证券发布研报称，英伟达(NVDA.US)2026财年Q4及全年业绩大幅超越市 场预期，GB300液冷深度渗透+下一代Rubin平台100%强制液冷，单机柜功率密度大幅跃升，彻底将液 冷从"可选配置"升级为"AI算力基础设施标配"。英伟达下放冷板、CDU等核心部件采购权，在液冷技术 迭代需求压力下，开放趋势预期长期延续，具备微通道冷板、高密度CDU等核心技术储备的国产厂 商，正迎来入围英伟达合格供应商名单的历史性机遇。 国联民生证券主要观点如下： 英伟达业绩超预期，夯实液冷行业高景气 英伟达供应链开放，国产液冷迎来黄金切入窗口 从H100"指定独供"到GB300"白名单+ODM自主采购"，英伟达下放冷板、CDU等核心部件采购权。下一 代Vera Rubin平台单机柜功率密度大幅提升，其架构设计实现"100%"全液冷散热，支持45℃热水冷却、 无需传统冷水机，对微通道冷板、高密度CDU等技术提出更高要求。在液冷技术迭代需求压力下，开 放趋势预期长期延续，具备微通道冷板、高密度CDU等核心技术储备的国产厂商，正迎来入围英伟达 合格供应商名单的历史性机遇。 ASIC放量崛起，成为液冷市场第 ...

【环球网科技综合报道】据路透社3月20日消息，英伟达宣布与亚马逊云科技（AWS）日前达成大规模芯片采购协议： 2027年底前，将向AWS供应100万颗GPU及多款AI相关芯片，供货从今年正式启动。 英伟达高管伊恩・巴克透露，此次合作不仅包含100万颗GPU，还涵盖Spectrum网络芯片、最新推出的Groq芯片等多款产 品。AWS将采用七款英伟达芯片组合，重点优化AI推理性能，提升AI响应与执行效率。 值得关注的是，英伟达网络产品将首次进入AWS数据中心。AWS长期使用自研网络设备，此次合作意味着英伟达在高端 云数据中心基础设施领域取得关键突破。（旺旺） 该周期与英伟达CEO黄仁勋提出的Blackwell、Rubin系列芯片总计1万亿美元市场机遇时间线一致，凸显双方对AI云计算 长期增长的信心。 ...

3月16日，美国加州圣何塞的冰球场又将座无虚席。英伟达CEO黄仁勋将穿着他那标志性的皮夹克走上舞台，开启一年一度的GTC大会。 但今年，气氛有些微妙。 过去十年，英伟达可以称为AI芯片市场唯一的"王"。 《华尔街日报》统计的数据显示，从2025年2月到10月，英伟达卖出了1478亿美元的芯片和相关硬件，比上年同期的910亿美元增长了62%。去 年7月，英伟达成为全球首家市值突破4万亿美元的公司，后来一度摸到5万亿的门槛儿。 但这个芯片帝国正被一群对手围猎。这场围猎的参与者可以大致分成三股势力： 第一是博通领衔的定制芯片（ASIC）阵营。博通可以说是几乎所有大客户"叛逃"的技术后台。谷歌的TPU、Meta的MTIA、OpenAI即将推出的 自研芯片Titan，背后都有博通的身影。 博通上季度AI收入84亿美元，同比暴增106%。根据Counterpoint Research预计，博通明年将控制定制AI芯片市场60%的份额。当英伟达的大客 户们纷纷转向定制芯片，博通就成了这场围猎中最关键的"军火商"。 第二股是超大规模云服务商的自研芯片浪潮。谷歌的第七代TPU Ironwood峰值性能4.6 petaFLOPS， ...

国芯网[原:中国半导体论坛] 振兴国产半导体产业！ 不拘中国、 放眼世界 ！ 关注 世界半导体论坛 ↓ ↓ ↓ ***************END*************** 半导体公众号推荐 半导体论坛百万微信群 加群步骤： 第一步：扫描下方二维码，关注国芯网微信公众号。 3月20日消息，据报道，在近几年的生成式人工智能热潮中，黄仁勋担任CEO的英伟达是一大受益者，他们的算力芯片被各大厂商大量购 买，他们的业绩也大幅提升，营收已连续11个财季创下新高，在截至今年1月25日的2026财年第四财季已经达到了681.27亿美元。 而从市场研究机构的报告来看，在生成式人工智能热潮中获益颇丰的英伟达，已连续两年是全球营收最高的半导体供应商，去年的营收遥遥 领先。 市场研究机构的报告显示，在2024年和2025年，英伟达的营收分别为973.95亿美元、1503.01亿美元，均是营收最高的半导体供应商。 在2024年和2025年，营收第二高的半导体供应商是三星电子，他们的半导体业务部门在这两年的营收分别为750.91亿美元、857.59亿美 元，去年虽然同比有增加，但与英伟达的差距有扩大。 就营收来看，当前全球半导体 ...

Nvidia Deepens Grip on Cloud AI With Major AWS Chip Dealhttps://t.co/KuqjbPeMV4 ...