指引超预期，Token 经济学的最佳增长引擎 英伟达(NVDA.O) ——英伟达 FY26Q4 业绩点评 | [姓名table_Authors] | 电话 | 邮箱 | 登记编号 | | --- | --- | --- | --- | | 秦和平(分析师) | 0755-23976666 | qinheping@gtht.com | S0880523110003 | | 刁云鹏(研究助理) | 021-38674878 | diaoyunpeng@gtht.com | S0880125070016 | 本报告导读： 长期收入上调，毛利率维持坚挺。Agent 应用拐点已至，英伟达将凭借最优的 Token 成本，持续领跑 AI 基础设施。 投资要点： | 财务摘要(百万美元) | 2025A | 2026A | 2027E | 2028E | 2029E | | --- | --- | --- | --- | --- | --- | | 营业收入 | 130,497 | 215,938 | 380,098 | 523,847 | 637,418 | | 同比增长 | 114.2% | 65.5% | 76.0% ...

这意味着，不论是用于大模型训练的高端GPU，还是支撑大型集群的高速网络组件，均处于增长状态。 在算力部署中，除了大众熟知的Blackwell系列芯片外，NVLink计算架构、以太网、InfiniBand平台等英 伟达的互联产品，也在迅速发展。 英伟达这份财报披露的一些数据超过此前市场预期。而面向2027财年，英伟达的指引还在增长，预计 2027财年第一财季营收为780亿美元，再次超出分析师预期。不过，中国市场依然存在悬念，财报指 出，截至目前，英伟达还没有在H200许可项目下产生任何收入。 英伟达CFO透露，接下来全年，英伟达还会继续销售Blackwell，同时销售Rubin架构芯片。游戏业务方 面，希望内存供应今年底的情况有所改变，目前看，未来几个季度供应还会非常紧张。汽车及机器人业 务方面，robotaxi服务正在增长，一些robotaxi公司的运营规模将在未来十年扩大数百万辆。 "算力即收入" 英伟达以一份打破纪录的财报，试图回击外界对AI泡沫的质疑。2月25日美股盘后，英伟达公布最新财 报，营收和利润都双位数攀升，再创历史新高。在市场对AI泡沫和云厂商数据中心资本支出表现出巨 大的担忧之际，英伟达CE ...

Core Viewpoint - Nvidia's record-breaking financial report aims to counter skepticism regarding the AI bubble, showcasing significant revenue and profit growth amid concerns about capital expenditures in the AI sector [2][3]. Financial Performance - In Q4, Nvidia reported a record revenue of $68.127 billion, a 73% increase from $39.331 billion year-over-year; net profit reached $42.96 billion, up 94% from $22.091 billion [3]. - For the entire year, Nvidia's revenue was $215.938 billion, with a net profit of $120.067 billion, equating to daily earnings of approximately $32.8 million (RMB 220 million) [3]. - The data center business contributed $193.48 billion in revenue for the year, a 68% increase, and accounted for over 91% of total revenue in Q4, with Q4 revenue reaching $62.3 billion, up 75% year-over-year and 22% quarter-over-quarter [3]. Business Segments - Within the data center segment, the "compute business" (primarily GPU products) generated $51.3 billion, a 58% increase, while the "network business" contributed $11 billion, growing 263% [3]. - Nvidia plans to continue selling the Blackwell architecture and the Rubin architecture chips, while the gaming segment faces tight memory supply [4]. Market Outlook - Nvidia's guidance for Q1 of FY2027 anticipates revenue of $78 billion, exceeding analyst expectations [4]. - Concerns remain regarding the Chinese market, as Nvidia has not generated any revenue under the H200 licensing project to date [4]. Industry Context - Wall Street is worried about high capital expenditures from tech giants potentially leading to systemic credit risks, with an estimated $700 billion to be spent on AI expansion by companies like Google, Microsoft, Meta, and Amazon this year [5][6]. - Despite the positive outlook for Nvidia, analysts caution that a slowdown in tech investments could significantly impact the company [6]. Strategic Initiatives - Nvidia is working to solidify its position in the AI ecosystem, with CEO Jensen Huang indicating a near agreement with OpenAI for a potential $100 billion AI infrastructure project [7]. - The upcoming GTC 2026 conference is expected to unveil "world's first" new chips, with speculation around the Rubin and Feynman series [7][8]. Competitive Landscape - Nvidia faces competition from custom AI chips like Google's TPU and Amazon's Inferentia, which are challenging the dominance of general-purpose GPUs in data centers [8]. - TrendForce predicts that the shipment share of ASIC AI servers will rise to 27.8% by 2026, surpassing GPU AI servers in growth rate [8].

Core Viewpoint - Nvidia's record-breaking financial report aims to counter skepticism regarding the AI bubble, showcasing significant revenue and profit growth amid concerns about capital expenditures in the AI sector [1][4]. Financial Performance - In Q4, Nvidia reported record revenue of $68.127 billion, a 73% increase from $39.331 billion year-over-year; net profit reached $42.960 billion, up 94% from $22.091 billion [3]. - For the entire year, Nvidia's revenue was $215.938 billion, with a net profit of $120.067 billion, equating to daily earnings of approximately $32.8 million (RMB 220 million) [3]. Business Segments - The data center segment generated $193.48 billion in revenue for the year, a 68% increase, and accounted for over 91% of total revenue in Q4, with $62.3 billion in revenue, up 75% year-over-year and 22% quarter-over-quarter [3]. - Within the data center segment, the "compute business" (primarily GPU products) contributed $51.3 billion, a 58% increase, while the "network business" generated $11 billion, a 263% increase [3]. Future Guidance - Nvidia's guidance for Q1 FY2027 anticipates revenue of $78 billion, exceeding analyst expectations [4]. - The CFO indicated ongoing sales of Blackwell and Rubin architecture chips, while the gaming segment faces tight memory supply [4]. Market Sentiment - Concerns persist among investors regarding potential threats from the AI bubble, with 23% of surveyed investors citing it as their primary concern, up from 9% in December [6]. - Despite positive performance, there are worries that capital expenditures by major tech firms may peak this year, impacting Nvidia [6][7]. Strategic Initiatives - Nvidia aims to solidify its position in the AI ecosystem, with plans to integrate various sectors onto its platform, including AI, robotics, and life sciences [8]. - The company is nearing an agreement with OpenAI for a potential $100 billion AI infrastructure project and has acquired technology from AI startup Groq for $20 billion [8]. Upcoming Developments - Nvidia's GTC 2026 conference is set for March 15, where new, unprecedented chips are expected to be unveiled [8][9]. - Speculation surrounds the new chips, likely from the Rubin series or the next-generation Feynman series, which are anticipated to be revolutionary [9].

在此起彼伏的AI泡沫担忧声中，AI巨头英伟达依然稳定地超预期增长。 2月26日，英伟达公布了2026财年第四季度财报（截至2026年1月25日，相当于自然年2025年第四季度），营收和利润都双位数攀升，再创历史 新高。 Q4当季营收681.27亿美元，同比增长73%，环比增长20%；净利润高达429.6亿美元，同比增长94%，环比增长35%；毛利率提高到了75%。 | | | Revenue by Market Platform | | | | | --- | --- | --- | --- | --- | --- | | ($ in millions) | Q4 FY26 | Q3 FY26 | Q4 FY25 | Q/Q | Y /Y | | Data Center | $62.314 | $51,215 | $35,580 | 22 % | 75 % | | Compute | 51.334 | 43.028 | 32.556 | 19 % | 58 % | | Networking | 10.980 | 8.187 | 3.024 | 34 % | 263 % | | Gaming | 3.727 | ...

Core Insights - Nvidia reported Q3 FY2026 earnings with revenue of $57.006 billion, a 62% year-over-year increase, surpassing market expectations of $55.212 billion, and a net profit of $31.91 billion, up 65% [1] - The data center business, a key segment, achieved record revenue of $51.2 billion, with a 25% quarter-over-quarter increase and a 66% year-over-year increase [1] - Nvidia's stock price rose over 4% in after-hours trading following the earnings report [1] Revenue Breakdown - Data center computing revenue reached a record $43 billion, a 56% year-over-year increase and a 27% quarter-over-quarter increase [1] - Networking revenue also hit a record $8.2 billion, driven by the launch of NVLink computing architecture and increased demand, showing a 162% year-over-year increase and a 13% quarter-over-quarter increase [1] - Gaming revenue increased by 30% year-over-year but saw a 1% quarter-over-quarter decline due to inventory normalization [2] - Professional visualization revenue grew by 56% year-over-year and 26% quarter-over-quarter, driven by the new DGX Spark™ release [2] - Automotive revenue rose by 32% year-over-year and 1% quarter-over-quarter, supported by increased adoption of autonomous driving platforms [2] Market Position and Challenges - CEO Jensen Huang criticized analysts suggesting an AI "bubble," asserting Nvidia's readiness to succeed at every stage of AI development [2] - Despite strong demand, Nvidia faces challenges, including the loss of the Chinese market and competition from Qualcomm's new AI chips [3] - CFO Colette Kress indicated that Blackwell sales continue to grow, with projected revenues from Blackwell and Rubin expected to reach $500 billion by the end of 2026 [3] - Supply chain capacity remains a significant constraint on Nvidia's revenue growth, as noted by JPMorgan [3] Market Sentiment - Huang's comments did not fully alleviate market concerns, as Nvidia's stock price gains narrowed after the earnings call [4] - Analysts express ongoing worries about the sustainability of growth in AI infrastructure spending [4]

Core Viewpoint - The company anticipates that AI-related revenue will surpass the total revenue from software and non-AI businesses within two years, with a long-term goal of reaching $120 billion in AI revenue by fiscal year 2030, directly linking this target to the CEO's compensation [1][3][5]. AI Revenue Growth - The CEO emphasized that meeting the AI computing needs of specific clients is the company's top priority, predicting that AI revenue will become the absolute core pillar of the business [2][3]. - The "Tan PSU Award" executive incentive plan ties the CEO's compensation to achieving specific AI revenue milestones, highlighting management's confidence in the AI business [3][5]. Market Segmentation - The AI accelerator market is expected to see a division, with large cloud service providers favoring customized ASIC chips for their specific workloads, while enterprise clients will continue to rely on commercial GPUs [6]. - The company identifies its XPU (customized processor) business opportunities primarily from existing and potential clients among the seven major cloud service providers [6]. Networking in AI - The CEO highlighted the growth potential of AI networking, asserting that Ethernet will play an increasingly important role in AI networks due to its proven technology and the rising demand for scalable networks [7]. - The company expects large-scale deployment of Ethernet in these expanded networks within the next 18-24 months [7].

Group 1 - The core sentiment in the market is driven by the dovish stance of the Federal Reserve, leading to significant gains in various stock indices, including a 2% rise in the Hang Seng Index, reaching a four-year high [1] - The technology sector, particularly Chinese concept stocks, has shown strong performance, with a 3.8% increase last Friday, also marking a four-year high [1] - The real estate sector has seen a rebound, with Vanke's stock hitting the limit up, indicating a potential recovery in the housing market, supported by debt management strategies [4] Group 2 - The rare earth sector is experiencing a surge due to news of import controls on rare earth minerals, highlighting China's dominance not only in mining but also in refining capabilities [3] - The chip industry, particularly stocks related to Nvidia, continues to perform well, driven by new product launches and positive market sentiment [3] - The white wine sector is showing strength, with several stocks hitting limit up, indicating a recovery in this previously underperforming segment [10] Group 3 - The overall market sentiment is mixed, with a significant increase in trading volume but a relatively low number of stocks hitting limit up, suggesting caution among investors [7] - The current market rally is characterized by sector rotation rather than broad-based gains, with technology stocks leading the charge while other sectors lag behind [8] - The automotive sector is facing challenges, with competitive pressures leading to underperformance, particularly in the complete vehicle segment [5]

Core Viewpoint - The article discusses the competition between InfiniBand and Ethernet in AI networking, highlighting the advantages of Ethernet in terms of cost, scalability, and compatibility with existing infrastructure [6][8][22]. Group 1: AI Networking Overview - AI networks are primarily based on InfiniBand due to NVIDIA's dominance in the AI server market, but Ethernet is gaining traction due to its cost-effectiveness and established deployment in large-scale data centers [8][20]. - The establishment of the "Ultra Ethernet Consortium" (UEC) aims to enhance Ethernet's capabilities for high-performance computing and AI, directly competing with InfiniBand [8][9]. Group 2: Deployment Considerations - Teams face four key questions when deploying AI networks: whether to use existing TCP/IP networks or build dedicated high-performance networks, whether to choose InfiniBand or Ethernet-based RoCE, how to manage and maintain the network, and whether it can support multi-tenant isolation [9][10]. - The increasing size of AI models, often reaching hundreds of billions of parameters, necessitates distributed training, which relies heavily on network performance for communication efficiency [10][20]. Group 3: Technical Comparison - InfiniBand offers advantages in bandwidth and latency, with capabilities such as high-speed data transfer and low end-to-end communication delays, making it suitable for high-performance computing [20][21]. - Ethernet, particularly RoCE v2, provides flexibility and cost advantages, allowing for the integration of traditional Ethernet services while supporting high-performance RDMA [18][22]. Group 4: Future Trends - In AI inference scenarios, Ethernet is expected to demonstrate greater applicability and advantages due to its compatibility with existing infrastructure and cost-effectiveness, leading to more high-performance clusters being deployed on Ethernet [22][23].

Core Insights - The article discusses the rise of AI Networking, particularly focusing on the "Scale Up" segment, highlighting its technological trends, vendor dynamics, and future outlook [1] Group 1: Market Dynamics - The accelerator market is divided into "commercial market" led by NVIDIA and "custom market" represented by Google TPU and Amazon Tranium, with the custom accelerator market expected to gradually match the GPU market in size [3] - Scale Up networking is transitioning from a niche market to mainstream, with revenue projected to exceed $1 billion by Q2 2025 [3] - The total addressable market (TAM) for AI Network Scale Up is estimated at $60-70 billion, with potential upward revisions to $100 billion [12] Group 2: Technological Evolution - AI networking has evolved from "single network" to "dual network," currently existing in a phase of "multiple network topologies," with Ethernet expected to dominate in the long term [4] - The competition between Ethernet and NVLink is intensifying, with NVLink currently leading due to its maturity, but Ethernet is expected to gain market share over the decade [5] - Scale Up is defined as a "cache coherent GPU to GPU network," providing significantly higher bandwidth compared to Scale Out, with expectations of market size surpassing Scale Out by 2035 [8] Group 3: Performance and Cost Analysis - Scale Up technology shows a significant performance advantage, with latency for Scale Up products like Broadcom's Tomahawk Ultra at approximately 250ns, compared to 600-700ns for Scale Out [9] - Cost-wise, Scale Up Ethernet products are projected to be 2-2.5 times more expensive than Scale Out products, indicating a higher investment requirement for Scale Up solutions [9] Group 4: Vendor Strategies - Different vendors are adopting varied strategies in the Scale Up domain, with NVIDIA focusing on NVLink, AMD betting on UA Link, and major cloud providers like Google and Amazon transitioning towards Ethernet solutions [13] - The hardware landscape is shifting towards embedded designs in racks, with a potential increase in the importance of software for network management and congestion control as Scale Up matures [13]