FY26Q4&FY27Q1 指引业绩超市场预期，Agentic AI 拐点已至 ——英伟达（NVDA.O）FY26Q4&FY26 全年业绩点评 要点 事件：英伟达发布截至 2026 年 1 月 25 日的 FY26Q4 及全年业绩。截至美国 东部时间 2 月 25 日 20：00，公司盘后股价上涨 0.17%。1）FY26Q4 营收、 净利润超市场预期。FY26Q4 实现营收 681 亿美元，同比+73%，环比+20%， 超一致预期的 659 亿美元；FY26 全年实现营收 2159 亿美元，同比+65%。 FY26Q4/FY26 全年 Non-GAAP 毛利率分别为 75.2%/71.3%；FY26Q4 Non-GAAP 净利润 396 亿美元，同比+79%，环比+25%，超一致预期的 375 亿美元，对应 Non-GAAP EPS 1.62 美元，超彭博一致预期的 1.532 美元。2） FY27Q1 营收指引超一致预期：FY27Q1 公司预计营收 780 亿美元（±2%， 未计入来自中国的数据中心计算收入），中值高于彭博一致预期的 729 亿美 元；Non-GAAP 毛利率 75.0%（±50bp）。 ...

当季度营收创下纪录，达681亿美元，环比增长20%，同比增长73%。净利润429.6亿美元，同比增长超 90%，摊薄后每股收益1.76美元。2026财年全年营收为2159亿美元，同比增长65%。毛利率71.1%，净利 润1200.67亿美元，摊薄后每股收益4.90美元。 | | GAAP | | | | | | --- | --- | --- | --- | --- | --- | | (S 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 | 2.0 pts | | Operating expenses | $6.794 | $5,839 | $4,689 | 16 % | 45 % | | Operating income | $44.299 | $36.010 | $24.0 ...

Core Viewpoint - NVIDIA is set to unveil a groundbreaking new chip at the GTC 2026 conference, which is expected to further solidify its leadership in AI infrastructure [1][3]. Group 1: Upcoming Events and Innovations - The GTC 2026 conference will take place from March 16 to 19 in San Jose, California, focusing on the new era of AI infrastructure competition [1]. - NVIDIA's CEO Jensen Huang hinted at multiple new chips that will push current physical limits, with speculation that one of them may be based on the Rubin architecture [3]. Group 2: NVIDIA Rubin Platform - The NVIDIA Rubin platform, launched on January 5, consists of six new chips designed for building advanced AI supercomputers, setting a new benchmark for cost-effective deployment and operation of large-scale AI systems [5]. - The Rubin platform introduces five innovative technologies, including next-generation NVIDIA NVLink, Transformer engines, confidential computing, RAS engines, and NVIDIA Vera CPUs, which significantly reduce the GPU count needed for training MoE models [5]. Group 3: Industry Partnerships and Collaborations - Major tech leaders, including OpenAI and Microsoft, have endorsed the Rubin platform, emphasizing its role in enhancing AI capabilities and efficiency [6][7]. - Meta has announced a long-term partnership with NVIDIA, planning to deploy millions of NVIDIA chips in its data centers, marking a significant step in challenging the dominance of traditional x86 architectures [12]. Group 4: Market Trends and Financial Insights - Despite the increasing capital expenditures by major tech companies in AI infrastructure, NVIDIA's stock price has remained relatively stagnant, with a decline of 6.49% since its peak in October 2025 [10][11]. - A report from Huafu Securities indicates that the narrative of "selling shovels upstream" in the AI industry is still ongoing, with projected capital expenditures for major Chinese and American tech giants reaching $404.49 billion in 2026, an 18% year-over-year increase [12][13].

Investment Rating - The industry investment rating is "Outperform the Market" (maintained) [1][50]. Core Insights - The demand for AI infrastructure is expected to continue increasing due to the rapid growth in the size and token consumption of AI models. Major tech companies like NVIDIA are continuously empowering AI model and application development through product iterations, thus benefiting from the AI industry boom [3][48]. - The report highlights the launch of NVIDIA's Rubin platform, which features six new chips designed for AI supercomputers, significantly reducing inference costs and enhancing performance [4][30]. - AMD introduced the Helios rack, aimed at meeting the Yotta-level AI computing demands, showcasing a substantial increase in performance and efficiency [40][41]. Summary by Sections NVIDIA Developments - NVIDIA unveiled the complete Rubin platform at CES 2026, which includes six new chips that work in extreme synergy to lower inference costs significantly. The token cost for inference can be reduced to one-tenth of the previous Blackwell platform [2][4]. - The Rubin platform features innovations such as the NVIDIA Vera CPU and Rubin GPU, which enhance training time and reduce costs for large-scale AI models [4][16]. - NVIDIA also launched the Alpamayo series of open-source AI models for autonomous driving, which includes a model with 10 billion parameters designed for the assisted driving research community [30][31]. AMD Developments - AMD's Helios rack is designed to meet the increasing demand for AI computing, featuring a core computing tray powered by four new AMD hardware components, providing 2.9 exaflops of AI computing power [41][42]. - The MI455X GPU, part of the Helios platform, boasts a performance increase of 10 times compared to its predecessor, with 320 billion transistors and advanced cooling technology [43][44]. - AMD plans to release the next-generation MI500 product in 2027, which is expected to achieve a 1000-fold increase in AI performance over four years [46][47]. Investment Recommendations - The report recommends focusing on AI themes for investment opportunities, highlighting companies in AI algorithms and applications, AI computing, and autonomous driving sectors [3][48]. - Specific strong buy recommendations include companies like 恒生电子, 中科创达, and 盛视科技, while also suggesting to monitor other firms in the AI space [3][48].

Group 1 - The A-share market experienced a strong rally, with all three major indices rising over 1%, driven by the rapid rise of the Generative Engine Optimization (GEO) concept [1][3] - AI application and marketing-related stocks surged, with companies like BlueFocus and Handan Information hitting the daily limit, and over 200 stocks in total reaching their upper limits [1] - The GEO concept aims to enhance brand information visibility in AI search results by optimizing content's semantic structure and credibility, indicating a shift in traditional marketing paradigms [3] Group 2 - The Ministry of Industry and Information Technology emphasized that intelligence is the core driving force behind the integration of information technology and industrialization, promoting the "AI + manufacturing" initiative [3] - Recent advancements in AI infrastructure were highlighted at CES 2026, with NVIDIA announcing the mass production of its new Rubin platform and AMD unveiling its Helios platform and MI400 series accelerators [3] - The competition in large models is shifting from parameter comparison to actual task execution capabilities, marking a transition to the "agent" phase in AI development [4]

Group 1 - The market experienced a significant rally on January 12, with all three major indices rising over 1%, and the total trading volume in the Shanghai and Shenzhen markets reaching 3.6 trillion yuan [1] - The Software ETF fund (561010) hit the daily limit, increasing by 9.97%, with a turnover rate of 30.33% and a transaction volume of 49.55 million yuan, indicating active market trading [1] - Key stocks in the software sector, including Torus, New Point Software, and Keda Guokuan, among others, also reached the daily limit [1] Group 2 - According to Dongfang Securities, the upcoming release of the DeepSeek V4 model in mid-February is expected to surpass mainstream models in programming tasks and has made significant improvements in handling long code prompts [2] - Major internet companies are accelerating their AI application strategies, with Alibaba launching the "Qianwen" app for personal AI assistance and Ant Group releasing a multimodal assistant named "Lingguang" [2] - The Software ETF fund closely tracks the CSI All-Share Software Development Index, providing investors with analytical tools to assess the performance of various industry companies [2]

Core Viewpoint - The AI concept stocks have experienced a significant surge, with A-shares' major indices collectively rising, and the AI sector seeing over 150 stocks with gains exceeding 10% [1][2]. Market Performance - On January 12, A-shares saw the Shanghai Composite Index rise by 1.09%, marking a 17-day winning streak and reaching a 10-year high. The Shenzhen Component and ChiNext Index increased by 1.75% and 1.82%, respectively. The total trading volume exceeded 3.6 trillion yuan, setting a historical record [1]. - The AI marketing index surged over 10%, with specific indices like Zhiyu AI and multimodal AI rising over 8%, and AIGC and AI agent indices increasing over 7% [2]. Government Initiatives - The Minister of Industry and Information Technology, Li Lecheng, announced plans to implement the "Artificial Intelligence + Manufacturing" initiative, focusing on developing general and industry-specific large models and fostering key industry intelligent entities [1][2]. Industry Trends - The AGI—Next summit highlighted a shift in large model competition from mere chat capabilities to intelligent agents capable of performing complex tasks in real environments. The year 2026 is anticipated to be pivotal for the practical application of these models in business processes [2]. - According to Dongfang Securities, AI is expected to transition from technological innovation to industrial application and global rule-making by 2026, with retail sectors benefiting significantly from AI-driven logic [3]. AI Infrastructure Development - The CES 2026 showcased advancements in AI computing power, with companies like NVIDIA and AMD unveiling new platforms and products aimed at enhancing AI capabilities [5][6]. - Continuous improvements in global AI infrastructure are expected to drive the growth of the AI industry, with domestic models in China also advancing towards practical applications across various sectors [6]. Investment Opportunities - Analysts recommend focusing on AI applications with revenue-generating potential, including AI marketing, short-form content generation, AI e-commerce, and gamified AI experiences [7]. - The year 2026 is projected to mark the widespread adoption of edge AI applications in areas such as smart driving, embodied robotics, and assisted healthcare [7].

Group 1 - The internet services and software development sectors are experiencing significant gains, with the CSI Software Service Index rising by 6.43% as of 11:05 AM on January 12, 2026 [1] - Key stocks such as Keda Xingtong, Hehe Information, and Weining Health have seen substantial increases, with respective rises of 16.05%, 14.97%, and 14.89% [1] - The CES 2026 exhibition highlights the continuous improvement in global AI computing power, with NVIDIA launching the mass-produced NVIDIA Rubin platform and AMD previewing the new Helios platform and MI500 series GPUs [1] Group 2 - CITIC Securities notes that Chinese independent third-party model vendors like MiniMax have significant commercial potential compared to leading US players, with MiniMax reporting $53.44 million in revenue for the first three quarters of 2025, indicating a potential for exponential growth [2] - The top ten weighted stocks in the CSI Software Service Index as of December 31, 2025, include iFLYTEK, Kingsoft Office, and Tonghuashun, collectively accounting for 60.89% of the index [2] - The Software ETF (159852) tracks the CSI Software Service Index, providing an accessible investment tool for the computer software industry, with additional opportunities available through the Software ETF linked fund (012620) for AI software investments [2]

Market Strategy - The market continues to rise with increased volume, and the Shanghai Composite Index has broken through previous highs [5][10] - The overall A-share market saw a rise of 1.59%, with the Shanghai Composite Index up by 1.50% and the ChiNext Index up by 0.75% [7] - The performance of the hard technology sector is leading, while the innovation growth sector is lagging [8] Industry Dynamics - Global sodium-ion battery shipments are expected to reach 9 GWh in 2025, a year-on-year increase of 150% [26] - The main applications for sodium-ion batteries include energy storage, light electric vehicles, and AI data centers, with significant market share growth [27] - NVIDIA has launched the Rubin platform, designed for AI supercomputing, featuring six new chips that enhance training speed and reduce costs [28][29] Company Updates - Haizhi Science (002653.SZ) has received acceptance for its innovative drug HSK39297 for treating paroxysmal nocturnal hemoglobinuria [32] - Meiya Optoelectronics (002690) won a lawsuit with compensation exceeding 198 million yuan [34] - Yisheng (002458.SZ) reported a 40% year-on-year increase in white feather chick sales in December 2025, while breeding pig sales declined by 43% [36] - Shennong Group (605296.SH) sold 3.0742 million pigs in 2025, generating revenue of 4.729 billion yuan [38] - Taotao Vehicle (301345.SZ) expects a net profit of 800-850 million yuan for 2025, a year-on-year increase of 85.5%-97.1% [41] - Ugreen Technology (301606.SZ) anticipates a net profit of 653-733 million yuan for 2025, reflecting a growth of 41.26%-58.56% [43]

Core Insights - NVIDIA's CEO Jensen Huang highlighted the transformative impact of next-generation accelerated computing and AI across industries during his keynote at CES 2026 [2] - The demand for AI training and inference computing is surging, with the Rubin architecture entering full-scale production and expected to launch in the second half of 2026, offering up to a 10x reduction in token costs compared to the previous Blackwell generation [2][4][5] NVIDIA Rubin Platform - The NVIDIA Rubin platform features six new chips designed for extreme collaboration, significantly reducing training times and inference token costs [4] - The six chips include NVIDIA Vera CPU, NVIDIA Rubin GPU, NVIDIA NVLink 6 Switch, NVIDIA ConnectX-9 SuperNIC, NVIDIA BlueField-4 DPU, and NVIDIA Spectrum-6 Ethernet Switch [4] - Innovations in the Rubin platform include the latest NVIDIA NVLink interconnect technology, a Transformer engine, confidential computing, and RAS engine [4] AI Model Advancements - The Rubin platform accelerates intelligent agent AI, advanced reasoning, and large-scale mixture of experts (MoE) model inference, reducing the number of GPUs needed for training MoE models by four times compared to previous generations [5] - The platform introduces a new generation of AI-native storage architecture designed for gigascale inference context, enhancing response capabilities and throughput [5] Market Deployment and Partnerships - NVIDIA Rubin products will be available through partners like AWS, Google Cloud, Microsoft, and others in the second half of 2026 [5] - CoreWeave will collaborate with NVIDIA to leverage Rubin's advancements in inference and MoE models, while major server manufacturers like Cisco, Dell, HPE, Lenovo, and Supermicro are expected to launch Rubin-based servers [6] Physical AI and Open Source Models - Huang announced the arrival of "physical AI's ChatGPT moment," with machines beginning to understand and act upon real-world data [12][13] - NVIDIA introduced the open-source physical AI foundational model, Cosmos, which has been pre-trained on vast datasets to understand the workings of the world [13] - The Alpamayo series of open-source AI models aims to accelerate the development of safe, reasoning-based autonomous vehicles, garnering interest from industry leaders [14] Robotics and Ecosystem Development - Global robotics leaders are developing products based on NVIDIA's Isaac platform and GR00T foundational model, covering various applications from industrial to consumer robotics [15] - NVIDIA emphasizes the importance of building an open-source AI ecosystem, with models like DeepSeek R1 demonstrating rapid industry adoption and collaboration [15] Industry Implications - The introduction of the Vera Rubin platform is expected to drive demand for high-speed optical modules and CPO technology, with companies in the supply chain already preparing for this shift [9][10] - The increased power requirements of the Rubin GPU, estimated at around 1800 watts, will elevate the demands on power supply and cooling systems [10]