"需求是否过于密集？" 记者|杨卉 编辑|程鹏 陈俊杰杜恒峰校对|陈柯名 当地时间2月25日，英伟达发布了2026财年第四季度及全年财务业绩。其中，数据中心的业绩增长尤为 亮眼。第四财季，英伟达数据中心业务实现营收623亿美元，创历史新高，全年营收1937亿美元（约合 人民币13251元），同比增长68%。 在稍后的财报电话会上，英伟达CEO黄仁勋和首席财务官Colette Kress反复强调，AI（人工智能）发展 已经迎来拐点，算力需求呈爆发式增长。在这一新时代背景下，算力即营收。 黄仁勋还在财报电话会上透露，目前各大云计算服务商、超大规模云计算企业、AI模型研发商和企业 客户部署的Blackwell（英伟达目前最高级的AI芯片）架构基础设施，算力已达90亿瓦，且处于满负荷 运行状态。另外，本周早些时候，英伟达已向客户交付首批Vera Rubin样品，量产计划将于2026年下半 年如期推进。 至于业内普遍关注的英伟达上一代芯片H200，根据该公司高管在财报电话会上的说法，面向中国客户 的H200产品出口尚未产生相关营收。另外，英伟达方面还谈到，中国本土竞争对手近期完成首次公开 募股，实力不断提升，或对全球A ...

每经记者｜杨卉 每经编辑｜陈俊杰 分业务来看，根据财报，英伟达大部分收入来自其数据中心业务，该业务因科技行业对人工智能基础设 施的巨大投资而得到提振。第四财季，英伟达数据中心业务实现营收623亿美元，创历史新高，环比增 长22%，同比增长75%；全年营收1937亿美元，同比增长68%。 但值得关注的是，去年以来，"数据中心需求过于集中"的声音就一直存在。 在财报电话会上，Colette Kress也表示，2026 年全球头部的五大云计算服务商和超大规模云计算企业贡 献了英伟达数据中心业务超50%的营收，这不免再次加重了华尔街对英伟达数据中心需求过于集中的担 忧。 对此，Colette Kress强调称，英伟达方面仍认为，传统数据中心负载向GPU加速计算的转型，以及AI对 现有超大规模云计算负载的赋能，将在长期为公司贡献约一半的增长机遇。如同电力和互联网，每个国 家都将打造并运营本土的AI基础设施。 当地时间2月25日，英伟达发布了2026财年第四季度及全年财务业绩。其中，数据中心的业绩增长尤为 亮眼。第四财季，英伟达数据中心业务实现营收623亿美元，创历史新高，全年营收1937亿美元，同比 增长68%。 在稍 ...

2026财年全年，英伟达营收达到2159.4亿美元（约合人民币14834.0亿元），同比增长65%；净利润则达 到1200.8亿美元（约合人民币8248.9亿元），同比增长65%。 据美国消费者新闻与商业频道（CNBC）、南华早报、英伟达网站消息，当地时间2月25日，英伟达公 布了截至2026年1月25日的2026财年第四季度财报以及2026财年全年业绩。 本财季，英伟达营收和净利润双双超越华尔街预期，营收达681.3亿美元（约合人民币4680.2亿元），同 比增长73%，环比增长20%；净利润（非GAAP）达395.5亿美元（约合人民币2716.9亿元），同比增长 79%，环比增长25%。 其中，数据中心营收达623亿美元，环比增长22%，同比增长75%。报道称，该公司超过91%的销售额来 自数据中心部门，该部门包含其市场领先的人工智能芯片。 克雷斯还称，"我们在中国的竞争对手，在近期IPO的推动下正在取得进展，从长期来看，有可能颠覆 全球AI行业的格局。" 据悉，英伟达对2027财年第一季度的展望为预计实现营收780亿美元，上下浮动2%；预计GAAP和非 GAAP毛利率分别为74.9%和75.0%，上下浮 ...

全球AI算力龙头又"杀疯了"，AI还远没到头。 英伟达最新财报"碾压式"超预期，Q4营收狂增73%，黄仁勋高呼"AI已到拐点"。 不过在财报公布后，公司股价盘后冲高回落，此前一度涨超4%。 截至发稿，夜盘交易微跌0.13%，最新总市值4.75万亿美元。 Q4狂揽681亿美元 美东时间25日，英伟达发布2026财年第四季度及全年财报。 第四季度，公司营收达创纪录的681亿美元，同比猛增73%，环比增20%。 GAAP和非GAAP毛利率分别为71.1%和71.3%；每股摊薄收益分别为4.90美元和4.77美元。 | GAAP | | | | | | | --- | --- | --- | --- | --- | --- | | ($ in millions, except earnings per share) | Q4 FY26 | Q3 FY26 | Q4 FY25 | 0/0 | Y/Y | | Revenue | $68,127 | $57,006 | $39,331 | 20 % | 73 % | | Gross margin | 75.0 % | 73.4 % | 73.0 % | 1.6 pts | ...

美光科技（MU）下跌 3%，因韩联社（Yonhap）报道称，三星计划最早于本月重启 HBM4（高带宽内存 4）的 大规模生产，以供应给英伟达（NVDA）的 Vera Rubin芯片。 我们早在去年 9 月就建议选择三星而非美光，因当时担忧美光在 HBM 领域的进展缓慢，而这些担忧如今似乎已在 一定程度上反映到市场中。 不过，自去年秋季以来，美光股价已累计上涨 200% 以上，而三星仅上涨 100% 美光科技（MU）下跌 3%，因韩联社（Yonhap）报道称，三星计划最早于本月重启 HBM4（高带宽内存 4）的 大规模生产，以供应给英伟达（NVDA）的 Vera Rubin芯片。 我们早在去年 9 月就建议选择三星而非美光，因当时担忧美光在 HBM 领域的进展缓慢，而这些担忧如今似乎已在 一定程度上反映到市场中。 今日与投资者交流后发现，多数人观点与我们一致，当前主要讨论焦点包括：亚马逊云服务（AWS）增速能有多 快、运营利润还有多少上调空间，以及资本支出周期将持续多久。 Spotify（SPOT）下跌 2%：上周末，TMTB 聊天群中就多空双方对该股的争论展开了热烈讨论，具体观点如 下： AnonCap（Ta ...

2月10日，行业分析机构SemiAnalysis发布的最新报告显示，下一代高带宽内存（ HBM4 ） 市场将由韩系厂商SK海力士与三星彻底主导，美国厂商美光（Micron）因技术路线选择失 误，已痛失英伟达下一代Rubin芯片的HBM4供应订单，其在该平台的供应份额恐将降至 0%。 报告明确指出，美光在NVIDIA Rubin平台的HBM4供应份额已被"清零"。与之形成鲜明对比 的是，韩系双雄将瓜分全部订单——市场预计 SK海力士 将拿下约70%的份额，剩余30%则 由三星收入囊中，这也意味着韩系厂商将全面垄断这一高端内存核心市场。 美光此次遭遇市场滑铁卢，核心症结在于技术路线的选择偏差。为降低生产成本并掌控供应 链主动权，美光采取了内部自主设计、制造HBM4基础裸片（ Base Die ）的"单打独斗"策 略，这与竞争对手的合作模式形成了鲜明差异。 不同于SK海力士选择与台积电强强联手、三星拥有自家逻辑代工能力，美光的独立研发生产 模式不仅引发了严重的散热问题，其产品的引脚速度（ Pin Speeds ）也未能达到英伟达的 客户标准。更为关键的是，美光不愿转向更先进的外部制程节点，最终导致自身在关键性能 指 ...

Economic Information - The "Qianwen Toolbox" mini-program by Alibaba has been restricted by WeChat due to numerous user complaints regarding potential fraud [2] - The peak sales season for liquor during the Spring Festival in 2026 is delayed by nearly a month compared to previous years, with overall sales expected to be difficult to exceed last year's levels due to more rational consumer behavior [2] - A price war in collagen products has emerged, with imported high-end products priced between 8,000 to 12,000 yuan, while new brands are priced below 1,000 yuan, leading to significant price differences for similar efficacy products [2] - Insurance companies are set to disclose their solvency reports for Q4 2025, with 57 life insurance companies reporting a net profit increase of over 150% year-on-year, marking a "strongest profit year" despite pressures from a low-interest-rate environment [2] Company Alerts - Ruihua Jewelry is under investigation by the China Securities Regulatory Commission for suspected information disclosure violations [4] - Jieqiang Equipment and its actual controller were fined 1.5 million yuan for committing bribery [4] - Several companies, including Congsheng Co. and Uder Precision, have announced plans to reduce their shareholdings by up to 5% and 3% respectively [4][6] Overseas Alerts - Research firm Semianalysis has downgraded Micron's market share in Nvidia's next-generation AI chip Vera Rubin's HBM4 to 0%, predicting that SK Hynix will capture 70% of the supply [4] - Federal Reserve Governor Christopher Waller indicated that the optimistic sentiment in the cryptocurrency market, which surged after Donald Trump's election, may be fading [4] Stock Movement Alerts - Companies such as Hengdian Film and Television and Zairun New Energy have experienced significant stock price fluctuations, triggering severe abnormal movement alerts due to high deviation percentages [8]

Core Viewpoint - Nvidia's transition to the new Blackwell AI chips has faced significant deployment challenges, particularly for major clients like OpenAI and Meta, but the company has managed to maintain its market position and address many technical issues [2][3][4]. Group 1: Deployment Challenges - Nvidia's CEO Jensen Huang indicated that the complexity of the new Blackwell AI chips would make the transition from the previous generation challenging for clients, requiring adjustments across various system components [2]. - Major clients, including OpenAI and Meta, struggled with the deployment and operation of Blackwell servers, which contrasted sharply with the quicker deployment of previous Nvidia AI chips [2][3]. - Despite these challenges, Nvidia's business has not been severely impacted, maintaining a market capitalization of $4.24 trillion and resolving many technical issues hindering client deployment [2][3]. Group 2: Client Reactions and Adjustments - Clients like OpenAI and Meta have expressed private dissatisfaction regarding the inability to build chip clusters at the expected scale, which limits their capacity to train larger AI models [3][4]. - To address client dissatisfaction, Nvidia provided refunds and discounts related to issues with the Grace Blackwell chips [3][4]. - Nvidia has collaborated closely with leading cloud service providers to improve the deployment process, indicating a commitment to joint engineering development [4]. Group 3: Product Improvements - Nvidia has learned from the deployment challenges and has optimized the existing Grace Blackwell systems while also improving the upcoming Vera Rubin chip servers [5]. - An upgraded version of the Grace Blackwell chip, named GB300, has been introduced to enhance stability and performance, addressing issues encountered with the first generation [5]. - Some clients have adjusted their orders to the upgraded products, indicating a shift in demand towards improved chip versions [5]. Group 4: Financial Implications - Delays in chip deployment have led to financial losses for cloud service partners of OpenAI, who invested heavily in Grace Blackwell chips expecting quick returns [9][10]. - Some cloud service providers negotiated discount agreements with Nvidia to alleviate financial pressure due to delayed chip usage [9]. - Oracle reported significant losses in its AI cloud business due to the slow deployment of Blackwell chips, highlighting the financial risks associated with new technology launches [10].

公众号记得加星标⭐️，第一时间看推送不会错过。 英伟达首席执行官黄仁勋昨晚宴请供应链伙伴高层，「兆元宴」再度登场。黄仁勋表示，台积电今年 必须要非常努力工作，因为英伟达需要很多晶圆，他预期「未来十年台积电的产能可能会成长超过百 分之百，是非常显著的规模扩张，而光是为英伟达的需求就得翻倍」。 黄仁勋再度选在台北砖窑古早味怀旧餐厅举行「兆元宴」，和供应链伙伴餐叙，出席的包括台积电董 事长魏哲家、联发科首席执行官蔡力行、广达董事长林百里、英业达董事长叶力诚、纬创董事长林宪 铭、鸿海董事长刘扬伟、宏碁董事长陈俊圣、矽品董事长蔡祺文、和硕董事长童子贤、共同首席执行 官郑光志及邓国彦、华硕董事长施崇棠、共同首席执行官胡书宾及许先越、纬颖董事长洪丽寗、台达 电董事长郑平、仁宝董事长陈瑞聪以及云达总经理杨麒令等人。餐叙结束后黄仁勋亲自送魏哲家离 开，显见对台积电的重视程度。 黄仁勋说，台积电今年要非常努力工作，因为英伟达需要很多晶圆和CoWoS，他也说，台积电做得 非常好。英伟达已经全面投产Blackwell、Vera Rubin芯片，而Vera Rubin包含六款不同的芯片，每 款都是世界上最先进的芯片。黄仁勋强调，今年英 ...

Core Viewpoint - The strategic partnership between Eli Lilly, a leading pharmaceutical company, and Nvidia, a top technology giant, marks a significant shift in the pharmaceutical industry, focusing on AI-driven drug development and manufacturing processes [1][14]. Group 1: Partnership Details - Eli Lilly and Nvidia will invest $1 billion over five years to establish a joint innovation lab in the San Francisco Bay Area [1]. - The lab will not only serve as a computing center but will also aim to completely restructure the drug development process using AI [2]. - The partnership will utilize Nvidia's latest AI chip architecture, Vera Rubin, which is designed for high-precision scientific calculations essential for drug development [2][3]. Group 2: Technological Integration - The collaboration will integrate hardware and software, with Nvidia's BioNeMo platform and Eli Lilly's TuneLab platform combining to enhance drug discovery [3][4]. - BioNeMo will function as a generative AI platform for biology, capable of generating new protein structures, while Eli Lilly will contribute its extensive historical experimental data [3][4]. - The partnership aims to address the data and model gap in AI healthcare, leveraging federated learning technology [4]. Group 3: Manufacturing Innovations - The collaboration extends to manufacturing, with plans to create a "digital twin" of Eli Lilly's production line using Nvidia's Omniverse platform [5]. - This digital twin will simulate production processes to optimize supply chain efficiency, potentially leading to significant revenue increases for high-demand products [5]. Group 4: Industry Context and Implications - Eli Lilly's decision to partner with Nvidia reflects a strategic move to overcome the challenges of traditional drug development, which is often time-consuming and costly [6][7]. - The partnership signifies a shift from a "Discovery" to a "Design" paradigm in drug development, allowing for targeted molecular design rather than random screening [7][8]. - The collaboration is expected to accelerate industry changes, prompting other major pharmaceutical companies to seek similar technological partnerships [16][18]. Group 5: Future Outlook - The partnership is seen as a potential turning point in AI-driven pharmaceutical development, creating a new model of collaboration between top pharmaceutical and technology companies [15][16]. - The competition in the pharmaceutical industry is likely to intensify as companies race to secure technological alliances, with AI becoming a critical component of drug development [19][20].