1）英伟达发布 4Q25 业绩，单季度每股盈利 1.62 美元，较分析师预期的 1.53 美元高出 5.81%，同比+82.02%；单季度 营收 681.27 亿美元，较分析师预期的 660.03 亿美元高出 3.22%，同比+73.21%。同时公司给出 1Q26 780 亿美元业绩 指引，超分析师预期的 727.8 亿美元，整体业绩&指引双 beat。但从股价看市场反应温和，我们认为主要系担心竞争 格局、预期交易充分的原因。展望 3 月，NVIDIA GTC 大会即将开幕。将首次公开下一代核心产品 Feynman 芯片，且将 搭载全球首款 1.6nm 制程工艺，同时我们预计有望展示 CPO 交换机、rubin ultra 等方案，或将给再催化光通信板块。 2）旭创发布业绩快报，2025 年全年收入 382.40 亿元，我们测算对应 4Q25 单季度收入 132.35 亿元，同比+102%/环比 +30%，增长稳健。2025 年全年营业利润 135.97 亿元，我们测算对应 4Q25 单季度营业利润 47.59 亿元，同比+190%/环 比+20%。整体处于此前业绩预告中枢。利润增速小于收入增速，我们认为主要 ...

光模块：本周光模块指数+4.84%，本月以来光模块指数+4.14%。NV 业绩&指引双 beat。但从股价看市场反应温和， 我们认为主要系担心竞争格局、预期交易充分的原因。展望 3 月，NVIDIA GTC 大会即将开幕,我们预计有望展示 CPO 交换机、rubin ultra 等方案，或将给再催化光通信板块。 IDC：本周 IDC 指数+2.41%，本月以来，IDC 指数+2.44%。OpenRouter 数据显示，2026 年 2 月 9 日至 15 日当周，中 国模型的调用量为 4.12 万亿 Token，首次超过美国模型（2.94 万亿 Token）。中国领先的芯片制造商正努力在两年 内将采用尖端工艺技术制造的芯片产量提高五倍，以满足国内人工智能行业的需求。 核心数据更新： 电信业务量收增速逐步提升。2025 年电信业务收入累计完成 1.75 万亿元，同比增长 0.7%。按照上年不变价计算的电 信业务总量同比增长 9.1%。12 月我国光模块出口金额当月同比增加 0.9%；1-12 月累计同比降低 16%。4Q25 微软/谷 歌/Meta/亚马逊资本支出分别为 299 亿/279 亿/214 亿/39 ...

Core Viewpoint - The U.S. government's approval for NVIDIA to export H200 AI chips to China comes with stringent conditions, leading to significant order cancellations from Chinese companies, which contradicts U.S. expectations of leveraging technology dependence to restrain China [1][4]. Group 1: U.S. Strategic Calculations - The release of H200 chips is a strategic move by the U.S. to alleviate pressure on American companies, as evidenced by Intel's record loss of $16.6 billion in Q3, the highest in 56 years, and significant stock declines of 25%-35% for major equipment firms [4]. - The U.S. aims to create a dependency in China's AI sector through limited technology exports, thereby delaying China's independent innovation efforts, similar to past strategies in high-end materials markets [4]. Group 2: China's Resilience and Growth - China's semiconductor exports are projected to exceed 1 trillion yuan in 2024, with a year-on-year growth of 21.4% from January to October, and a nearly 10 percentage point increase in self-sufficiency in chip production over the past decade [4][6]. - Domestic chip supply has improved, with Huawei's Ascend AI chips supporting over 40 large model applications, and SMIC's 28nm mature process capacity accounting for 25% of the global market [4]. Group 3: Impact of U.S. Restrictions - The "boomerang effect" of U.S. chip restrictions is evident, as China fills technological gaps through rapid innovation, with domestic chip equipment market share rising from 16% in 2020 to 28% in 2024 [6]. - Non-U.S. companies like ASML and Tokyo Electron are increasing their market presence in China, with ASML's DUV lithography machines capturing a 35% market share [6]. Group 4: Energy and Infrastructure Advantages - China's advantage in the AI race is bolstered by its energy capacity, with electricity generation reaching twice that of the U.S., potentially tripling by 2026, which supports high-energy AI data centers [7]. - The efficiency of infrastructure development in China allows for faster completion of data centers compared to the U.S., enabling China to leverage scale in computing power despite current chip performance disadvantages [7]. Group 5: Long-term Industry Outlook - China's advancements in the semiconductor industry are attributed to a comprehensive approach across the entire supply chain, achieving synergy in design, manufacturing, and packaging [9]. - Despite U.S. sanctions, Chinese chip companies have not only survived but have established a foothold in mature process markets and are extending into high-end sectors, indicating a robust competitive landscape [9].

Core Viewpoint - The Chinese economy is undergoing a fundamental shift from reliance on the real estate sector to the burgeoning artificial intelligence (AI) industry, which is becoming a new pillar of economic growth [1][2][5]. Industry Development - The AI industry in China has surpassed a scale of 700 billion, maintaining a growth rate of over 20% [5]. - China is home to over 5,300 AI companies, accounting for approximately one-sixth of global AI firms [5]. - The AI sector is expected to exceed 1 trillion by 2025 due to continuous domestic investment and development [5]. Technological Impact - AI technology is penetrating various industries, leading to profound changes in sectors such as finance, healthcare, manufacturing, and agriculture [5][13]. - AI applications are enhancing production efficiency and altering employment structures and industry layouts [2][11]. Infrastructure and Support - The development of computing power infrastructure is crucial for the growth of the AI industry, with initiatives like "East Data West Computing" reducing computing costs in the western regions by 50% [7][9]. - Data centers in regions like Hainan and Guizhou are providing essential computing capabilities for the rapid development of the AI sector [9][19]. Economic Transition - The shift from a resource-driven economy, primarily reliant on real estate, to an innovation-driven model is evident, with AI leveraging data resources for sustainable growth [11][21]. - AI is facilitating a transformation in economic growth patterns, moving away from traditional resource constraints [21][23]. Future Outlook - The AI industry is expected to continue driving industrial upgrades and technological innovations, positioning itself as a key engine for high-quality and sustainable economic development in China [23].

Group 1 - The U.S. government has approved the export of H200 AI chips to China by Nvidia, but with conditions including a 25% revenue share to the U.S. government [1] - The U.S. Department of Commerce is revising its export licensing review policy for certain semiconductors to China, shifting from a presumption of denial to case-by-case reviews [1] - Restrictions include limiting the quantity of chips that Chinese customers can acquire to 50% of what U.S. customers purchase, and requiring third-party testing in the U.S. to confirm AI capabilities [1] Group 2 - Chinese companies are actively working on developing domestic AI chips to replace Nvidia's market share, with significant investments from major tech firms like Huawei, Alibaba, Tencent, Baidu, and ByteDance [2] - Huawei has announced a three-year product iteration roadmap for its Ascend AI chips, indicating a strategic focus on enhancing domestic chip capabilities [2] - The Chinese government advocates for cooperation between the U.S. and China to achieve mutual benefits in the context of AI chip exports [2]

Group 1 - The U.S. government has approved the export of H200 AI chips to China by Nvidia, but with conditions including a 25% revenue share to the U.S. government [1] - The U.S. Department of Commerce is revising its export licensing policy for certain semiconductors to China, shifting from a presumption of denial to case-by-case reviews [1] - The H200 chip features significantly improved performance, with six times the capabilities of its predecessor, the H20, although it is not Nvidia's highest-end product [1] Group 2 - Chinese companies are actively developing domestic AI chips to compete with Nvidia, with major players like Huawei, Alibaba, Tencent, Baidu, and ByteDance increasing their investment in chip R&D [2] - The Chinese government advocates for cooperation between China and the U.S. to achieve mutual benefits regarding the export of H200 chips [2]

Core Insights - NVIDIA's CEO Jensen Huang announced the new Rubin platform at the Consumer Electronics Show, highlighting significant improvements in inference cost and training efficiency, with full production set to begin and shipments planned for the second half of 2026 [1][3] - Huang emphasized the strong demand for the H200 AI chips in the Chinese market, which have been approved for export, while the company awaits further government approvals for shipments [1][6] Group 1: New Product Launch - The new Vera Rubin platform integrates six new chips, achieving a training performance that is 3.5 times better than the previous Blackwell platform and improving AI software performance by 5 times [3][4] - The Rubin platform can reduce inference token generation costs by up to 10 times compared to Blackwell [4] Group 2: Market Dynamics - Huang addressed concerns about increasing competition in the AI sector and the sustainability of AI spending growth, aiming to alleviate fears of a potential "AI bubble" [4] - Despite NVIDIA's dominance in training AI models, it faces growing competition from traditional rivals like AMD and major clients such as Google [4] Group 3: Export and Regulatory Challenges - The U.S. government, under former President Trump, allowed NVIDIA to export H200 chips to China, with conditions related to national security and a 25% revenue share for the U.S. [6] - The H200 chip reportedly has nearly six times the performance of its predecessor, the H100, enabling Chinese AI labs to build supercomputers comparable to top U.S. models, albeit at a higher cost [6] Group 4: Chinese Market Response - There are uncertainties regarding whether Chinese buyers will purchase the H200 chips, as some are reportedly reluctant to buy the "special version" designed for China [7][8] - Chinese companies are actively developing domestic AI chips to replace NVIDIA's market share, with significant investments from major tech firms like Huawei, Alibaba, Tencent, and ByteDance [8]

Macroeconomic Dynamics - Industrial enterprise profits in November decreased by 13.1% year-on-year, indicating a weakening of overall domestic demand despite strong export performance [1] - The Federal Reserve's December meeting minutes are expected to reveal significant internal divisions, with Trump potentially designating a new Fed chair in the first week of January [1] - The 14th National People's Congress will convene for its fourth session on March 5, 2026, in Beijing [1] Satellite Internet - The commercial rocket industry is subject to new detailed listing rules under the Sci-Tech Innovation Board, requiring successful orbital launches of reusable rockets [1] - The market interest has expanded to materials related to satellite technology, including packaging, solid rocket engine materials, perfluoropolyether lubricants, coatings, and lightweight materials [1][2] Artificial Intelligence - Overseas data centers are seeking jet engine turbines and fossil fuel generators as alternative power sources due to difficulties in grid access and rising costs [5] - ZJ is rumored to procure 400 e Ascend AI chips next year, with the first batch expected to be delivered at a scale of 10 billion [6] - Huawei announced plans to launch Ascend AI chips in South Korea next year, offering a "full-stack" solution [7] - The timeline for Artificial General Intelligence (AGI) has been pushed to the 2030s, but AI adoption is expected to accelerate, with significant revenue breakthroughs anticipated in 2026 [9] New Energy - CATL plans to apply sodium batteries on a large scale in 2026 across various sectors, including battery swapping, passenger vehicles, commercial vehicles, and energy storage, indicating a "sodium-lithium dual-star" trend [9] - The battery production schedule in January is expected to remain active despite being traditionally a slow season, with positive impacts on processing fees due to positive electrode maintenance [10] - Domestic demand for lithium batteries is projected to decline significantly in early 2026 compared to Q4 [11] - A new narrative in photovoltaics links it to commercial aerospace, particularly in "space computing" energy demand [12] Robotics - The Ministry of Industry and Information Technology has established a standardization technical committee for humanoid robots and embodied intelligence [9] Autonomous Driving - The first fully vehicle-compliant unmanned logistics vehicle is set to launch in early next year [9] Market Observation - On Friday, trading volume reached 2.16 trillion, with an increase of 235.7 billion, resulting in an eight-day consecutive rise in indices. The sectors leading the gains included non-ferrous metals, electric power equipment, and steel [9] - The technology sector experienced fluctuations, potentially awaiting adjustments in satellite market conditions and index corrections before resuming upward momentum [9] - Market sentiment remains high, but caution is advised regarding the rhythm of continuous increases and changes in trading volume [9]

Core Viewpoint - The A-share computing power sector underwent a significant value reshaping in 2025, transitioning from policy-driven expectations at the beginning of the year to performance verification by year-end, with the index showing an impressive annual increase of 80% [1][15]. Performance Summary - As of December 25, 2025, the Wind Computing Power Theme Index (399363.SZ) closed at 7108 points, reflecting a substantial rise [1][15]. - A total of 22 stocks in the A-share computing power sector doubled in value, with four stocks exceeding a 400% increase: Shijia Photon (457.28%), Xinyi Sheng (441.57%), Yongding Co. (425.89%), and Zhongji Xuchuang (410.21%) [2][16]. Market Dynamics - The capital market's perception of the computing power industry shifted from "concept speculation" to "value anchoring," indicating a necessary evolution of the Chinese computing power industry chain amid intensified global competition and domestic upgrade demands [3][18]. - The computing power sector exhibited a "ice-fire duality" structural differentiation, where the market focus transitioned from short-term policy benefits to long-term performance realization and technological barriers [4][18]. Key Sector Highlights - AI chips and semiconductor equipment emerged as the core areas of capital pursuit, with leading companies like Haiguang Information reporting a 54.65% year-on-year revenue increase to 9.49 billion yuan and a net profit of 1.96 billion yuan, up 28.56% [6][19]. - Another AI chip representative, Cambricon, reported a staggering 2386.38% year-on-year revenue growth to 4.607 billion yuan, marking a significant turnaround to profitability [7][19]. - The global AI server leader, Industrial Fulian, achieved record revenue of 603.931 billion yuan, a 38.4% increase, with net profit rising by 48.52% [7][19]. Technological Advancements - The data center sector is prioritizing quality over quantity, with national regulations mandating that new large data centers must achieve a PUE (Power Usage Effectiveness) of ≤1.2. Domestic liquid cooling technology has achieved a PUE of 1.03-1.08, becoming essential for energy efficiency [8][20]. - Companies like Yingwei Ke have reported a 40.19% revenue increase to 4.026 billion yuan, driven by their liquid cooling solutions that meet stringent energy efficiency requirements [8][20]. Domestic Substitution Progress - The domestic substitution process in the A-share computing power sector has transitioned from "policy-driven" to "market-driven," with a comprehensive ecosystem emerging that includes chips, servers, and foundational software [9][21]. - Leading companies like Huawei have established a vertically integrated capability from Ascend AI chips to the Euler operating system, playing a crucial role in the construction of AI computing centers [9][21]. Future Outlook - The computing power sector is expected to continue its value reshaping into 2026, with a focus on high-performance AI chips and the widespread adoption of liquid cooling technology [12][25]. - The competition will shift from individual products or technologies to ecosystem collaboration, with companies that can provide comprehensive solutions from hardware to software gaining a competitive edge [12][25].

Core Insights - The A-share computing power sector experienced significant value reshaping in 2025, transitioning from policy-driven expectations to performance verification, with the National Computing Infrastructure Index showing an 80.74% increase year-to-date by December 25, 2025 [1] - The market's perception of the computing power industry shifted from "concept speculation" to "value anchoring," reflecting intensified global competition and domestic industrial upgrade demands [1] - The domestic substitution process in the computing power sector evolved from policy support to market selection, driven by real demand and the establishment of a comprehensive domestic computing ecosystem [5][6] Group 1: Performance Highlights - AI chip and semiconductor equipment sectors were the main focus of capital investment, with companies like Haiguang Information reporting a 54.65% year-on-year revenue increase to 9.49 billion yuan and a 28.56% rise in net profit [2] - Cambricon Technologies reported a staggering 2386.38% year-on-year revenue growth to 4.607 billion yuan, indicating the rapid market adoption of domestic AI chips [3] - Industrial Fulian, a global leader in AI servers, achieved a 38.4% revenue increase to 603.931 billion yuan, with net profit rising by 48.52% [3] Group 2: Technological Advancements - Liquid cooling technology emerged as a critical growth point, with companies like Yingwei achieving a revenue increase of 40.19% to 4.026 billion yuan, meeting stringent energy efficiency standards [4] - The integration of hardware upgrades and software ecosystem improvements is driving the performance of domestic AI chips, with frameworks like MindSpore facilitating easier application development [9] - The shift towards liquid cooling technology is not only a response to energy efficiency requirements but also a foundational element for future high-density computing infrastructure [8] Group 3: Market Dynamics - The market is increasingly favoring companies with clear technological pathways and real profit growth, leading to a structural differentiation in the computing power sector [2] - The demand for self-controlled and secure computing solutions is accelerating domestic substitution, with companies like Haiguang Information and Cambricon creating valuable market space [7] - The competition is evolving from product-based to ecosystem-based, with companies that can provide comprehensive solutions across hardware, software, and services gaining a competitive edge [10] Group 4: Future Outlook - The value reshaping in the A-share computing power sector is expected to continue into 2026, with a focus on high-performance AI chips and broader applications of liquid cooling technology [10] - The growth structure of AI computing demand will further differentiate, favoring solutions that demonstrate high energy efficiency and specific application capabilities [10] - Companies that can prove their technology's commercial viability and integrate deeply into the domestic computing ecosystem will continue to attract capital investment [10] Group 5: Industry Transition - The Chinese computing power industry is transitioning from scale expansion to high-quality development, marking a critical period of proactive leadership in the digital economy [11]