Investment Rating - The report does not explicitly state an investment rating for the industry Core Insights - The report highlights a significant increase in demand for computing power driven by advancements in AI models, particularly in video generation and multi-modal applications, indicating a structural shift in the industry towards higher quality and more complex models [6][11][17] - The introduction of GLM-5 and its "Interleaved Thinking" mechanism represents a new paradigm where computing power is exchanged for intelligence, leading to increased computational demands for inference tasks [13][27] - The supply side is expected to transition from a state of scarcity to structural balance by 2026, with improvements in domestic chip performance and the approval of NVIDIA's H200 chips for the Chinese market [42][43] Summary by Sections Section 1: Supply and Demand Dynamics - Seedance 2.0 has shown a strong user demand with long wait times for video generation, indicating a critical need for computing resources [11][12] - The demand for computing power is expected to escalate as major internet companies continue to develop large-scale AI models, with a notable focus on multi-modal capabilities [17][18] Section 2: Rapid Release of Computing Demand - The competition among leading internet firms is intensifying, with significant investments in AI infrastructure and model development, leading to a sharp increase in inference computing demand [32][34] - The report predicts that by 2026, the AI application landscape will expand significantly, driving real-time inference computing consumption [6][33] Section 3: Supply Side Improvements - The approval of NVIDIA's H200 chips is expected to alleviate computing power shortages for major internet companies, enhancing model iteration speeds [42] - Domestic chip manufacturers are making strides in performance and ecosystem development, with several companies achieving significant advancements in their products [43][44] Section 4: Full-Chain Inflation in Domestic Computing Power - The report anticipates a "full-chain inflation" cycle in the computing power industry by 2026, with growth expected across various segments including AI data centers and cloud services [49] - Major tech companies are projected to increase their capital expenditures significantly, further driving demand for computing resources [51] Section 5: Related Companies - The report identifies several companies as relevant to the industry, including Dongyangguang, Cambrian, Haiguang Information, Wangsu Technology, and others [4][55]

"Hetzner本质上是一家云基础设施供应商，他们购买硬件进行管理，并出租其容量。对于这次涨价我并 不意外，"Omdia企业基础设施高级研究总监Vlad Galabov表示。他认为，其他云供应商也将陆续涨 价。"对于依赖低成本云托管的公司来说，今年的基础设施成本绝对会大幅跳升。" 在此之前，全球云计算市场已进入密集涨价周期。 从芯片、云到大模型，AI及算力产业链涨价潮正在逐步扩散。 日前，欧洲最大的数据中心运营商及云服务商之一Hetzner宣布，由于"IT多个领域成本大幅上涨"，公 司将于2026年4月1日起调高全线产品及服务报价，包括云服务、专用服务器、存储及负载均衡器等，多 款产品涨幅相当显著。此次调价将影响该公司位于欧洲、美国和新加坡数据中心的新订单及现有订阅用 户。 所有业务中，按欧元报价涨幅较为显著的是云服务报价，例如德国及芬兰的云服务价格根据配置等级的 不同，涨幅在30%到38%之间；美国地区的CCX专用vCPU云服务器价格普遍涨幅在30%水平。 这已是Hetzner本月第二次宣布提价。此前在2月2日，该公司曾宣布调高专用服务器的一次性设置费用/ 初装费，涨价原因主要在于"硬件组件采购价格异常高昂 ...

（编辑 丛可心） 证券日报网讯 2月25日，长光华芯在互动平台回答投资者提问时表示，目前在数通芯片领域，面向多种 技术路线和发展构建了完整的产品矩阵（如EML、DFB、VCSEL、PD、硅光芯片等）共同来满足日益 增长的算力需求。 ...

在行业分布上，中证人工智能主题指数(930713)覆盖AI全产业链，从底层算力、中层技术到上层应用， 完整覆盖人工智能三大核心环节。 指数前十大重仓股，按AI细分赛道分类来看，光模块/通信占比21.14%（中际旭创、新易盛），AI芯片/ 处理器占比14.51%（寒武纪、澜起科技、豪威集团），AI服务器/算力占比4.70%（中科曙光），算法/ 大模型/NLP占比4.36%（科大讯飞），计算机视觉占比4.11%（海康威视），AI应用/IP平台占比5.22% （金山办公、芯原股份）。 规模与流动性方面优势显著，截止2025年12月31日，天弘中证人工智能主题指数基金产品规模达33.94 亿元，A类：7.80亿，C类：26.14亿（数据来源产品2025年四季报）。 基金业绩方面，2025年历史收益亮眼：天弘中证人工智能主题指数基金A类：64.98%，C类：64.66% （数据来源产品2025年四季报）。 国内方面，2026年春节前后国产大模型迎来集体发布潮，智谱发布GLM-5大模型，参数达744B，编程 能力对标Claude Opus 4.5，且完成与多款国产算力平台的适配；MiniMax推出M2.5模型，在编程、工 ...

针对上述传闻，董事会作出明确澄清：近年来AI大模型发展加速，相应对于算力及相关服务的需求也快速增长，而集团的主营业务恰恰受益于这一趋势。 公告援引公司于2025年11月12日披露的业绩数据，截至2025年9月30日止9个月，集团83.9%的收入为算力相关收入，以算力使用量为主要计价基础；仅 16.1%的收入来自软件服务业务。2025年1月至9月，集团收入同比增长36.8%，并于2025年第三季度首次实现单季度盈利。 第四范式在公告中强调，集团核心主业以算力为核心，通过AI技术服务各行业企业客户，推动客户实现智能化升级与全面数字化转型。AI大模型的快速迭 代显著提升了市场对算力需求的增长，也为集团业务的进一步发展拓宽了广阔空间。 经作出一切合理查询后，董事会确认：集团业务运营正常，且截至公告日期，并不知悉任何应披露而未披露的内幕消息。 2月24日，北京第四范式智能技术股份有限公司（股份代号：6682）发布自愿性澄清公告，针对近期公司股价下跌及市场上关于AI大模型迭代可能对软件行 业造成冲击的传闻作出正式回应。 公告显示，董事会注意到近期公司股价出现较大幅度下跌。经了解，市场出现相关传闻称，AI大模型的快速迭代可 ...

Core Insights - The market is increasingly recognizing the imminent threat of AI disruption, with the perceived risk in the MSCI Europe index rising from 4% to 24% in just over a month, including the banking sector [1][9] - Morgan Stanley has shifted its stance from neutral to cautious regarding cyclical stocks versus defensive stocks, highlighting opportunities in the European credit market for downside protection [1][15] AI Capability Advancements - The latest AI model, GPT-5.2, has achieved human expert-level performance in 71% of professional tasks, marking a significant leap in AI capabilities [5][8] - The speed of AI advancements is accelerating, with predictions that upcoming models in 2026 will far exceed current capabilities due to increased computational power [8] Market Disruption Dynamics - Initial concerns about AI's impact on the software industry have rapidly expanded to broader economic disruption risks, reminiscent of market reactions during the early COVID-19 pandemic [9][10] - Approximately 10% of the MSCI Europe index (excluding banks) is now viewed as facing substantial AI disruption risks, with this figure rising to 24% when including banks [9][10] Valuation Trends - The valuation of "disruption stocks" has decreased from a peak P/E ratio of 24x to 16.4x, with further downward potential indicated by comparisons to "uncontested disruption stocks" [10] Resilience Assessment Framework - Morgan Stanley proposes a framework to evaluate sectors and stocks based on five dimensions of risk exposure, identifying utilities, semiconductors, defense, and tobacco as the most resilient sectors [11] - Sectors such as software, commercial services, and banking are identified as facing the highest disruption risk [11] Non-AI Replicable Assets - The report emphasizes the rising value of assets that cannot be replicated by AI, including physical assets, regulatory barriers, and unique human experiences [4][12][14] Credit Market Insights - Despite AI disruption concerns affecting some credit markets, European investment-grade spreads remain low, presenting opportunities for investors to hedge against potential downturns [15] Computing Power Demand - There is a significant and growing demand for computing power, with projections indicating that the growth rate of demand will outpace current supply forecasts [16][21] - The intensity of computing requirements for AI queries is increasing rapidly, with predictions that companies may need to double their computing power every six months [19][21]

Investment Rating - The report maintains a rating of "Buy" for the electronic industry, with a focus on specific stocks such as 香农芯创, 东山精密, 兆易创新, 北方华创, 中微公司, and 胜宏科技 [6][62]. Core Insights - The release of AI models by Alibaba and ByteDance is expected to significantly increase the demand for computing power, particularly in video generation applications, which are more resource-intensive than text generation [2][12]. - The report highlights that the domestic AI applications are transitioning from "technology development" to "scaled deployment," indicating a robust demand for AI computing resources [2][12]. - Kioxia reported record revenue of 543.6 billion yen (approximately 3.547 billion USD) for FY25Q3, driven by increased average selling prices (ASP) and bit shipments, with expectations for continued growth in the NAND market due to AI-driven demand [3][46][61]. Summary by Sections Section 1: AI Model Releases and Computing Power Demand - Alibaba's Qwen-Image-2.0 model integrates image generation and editing, supporting 1K token input and 2K resolution, enhancing performance significantly [1][13]. - ByteDance's Seedance 2.0 model, released on February 12, 2026, demonstrates a 3-5 times increase in computing power consumption compared to standard video generation models, addressing complex motion and interaction scenarios [2][19][20]. Section 2: Kioxia's Financial Performance - Kioxia's FY25Q3 revenue reached a historical high of 543.6 billion yen, with a quarter-over-quarter growth of 21.3%, attributed to increased ASP and bit shipments [3][46]. - The company anticipates FY25 revenue between 2.18 trillion yen and 2.27 trillion yen, with a focus on data center and enterprise SSD shipments driving profitability [3][61]. Section 3: Related Stocks and Market Opportunities - The report identifies several key stocks in the semiconductor and AI sectors, including companies involved in computing power chips, storage modules, and semiconductor equipment, indicating a broad range of investment opportunities [62][63].

Investment Rating - The report maintains a "Recommended" rating for the industry, indicating an expected relative performance exceeding 15% compared to the benchmark index over the next 12 months [8]. Core Insights - The launch of Seedance 2.0 by ByteDance marks a significant advancement in AI video generation, enabling users to create videos with simple text prompts, which is expected to enhance user engagement and frequency of use over the long term [3][5]. - The model's capabilities include multi-modal input acceptance, narrative coherence, and audio-visual synchronization, addressing key industry challenges such as character consistency and audio-visual mismatch [4][5]. - The demand for computational power is projected to increase exponentially due to the anticipated high concurrency from both consumer (C-end) content creation and business (B-end) API calls [6]. Summary by Sections Seedance 2.0 Launch and Features - Seedance 2.0 is integrated into various platforms, allowing users to generate short videos easily, which is expected to drive higher usage rates [3]. - The model supports multiple input types and can generate videos with synchronized audio, enhancing the creative process for users [4]. Computational Demand Projections - The report estimates that by October 2025, the monthly usage of AI-generated content on the platform will reach 63,900 times, with a total usage time of 1,668.2 hours [6]. - The complexity of video generation is significantly higher than that of text and images, leading to a substantial increase in computational requirements [6]. - Projections indicate that the number of required H100 GPUs for video generation could range from 2.4 million to 4.284 million under different scenarios, with a market space for H100 GPUs estimated between $4.8 billion and $8.568 billion [14][12]. Long-term Market Opportunities - The report highlights the potential for domestic computational power providers to benefit from the growth of AI video generation, with specific companies recommended for investment, including chip designers and AI server manufacturers [13]. - The user base for short videos in China is expected to exceed 1 billion by the end of 2024, with increasing daily usage times projected [7][11].

Group 1 - In January, the A-share market saw increased trading activity and risk appetite, with all three major indices achieving gains, notably the Shanghai Composite Index rising by 3.76% and the CSI 500 Index increasing by 12.12% [1] - The rise in cyclical industries such as precious metals, non-ferrous metals, chemicals, and oil and gas was driven by multiple factors including complex international dynamics, a weakening US dollar, and domestic industrial policy support, although there was significant volatility at the end of the month [1] - The technology sector, particularly AI applications, is expected to enter a critical validation window as large model technologies mature, with strong demand for computing power and investments in AI hardware like storage chips and semiconductor equipment [1] Group 2 - As of January 2026, the total number of bank wealth management products in the market reached 46,845, an increase of 553 from December 2025, with fixed-income products making up the majority at 43,215 [2] - In January, 2,969 new wealth management products were launched, a decrease of 305 from the previous month, with wealth management subsidiaries of banks accounting for 76.56% of the new issuances [2] - The overall risk level of wealth management products in January increased significantly, with 51 products rated as medium risk and 4 additional products rated as medium-high risk compared to the previous month [12] Group 3 - The asset allocation of wealth management products in January showed a high proportion of bonds and public funds, with a focus on cash, bank deposits, and short-term bond ETFs for liquidity and low risk [17] - Mixed wealth management products had diverse asset allocations, covering debt, convertible bonds, equity assets, and gold, with many products cautiously investing in gold ETFs to capture commodity investment opportunities [18] - The top five institutions accounted for 43.81% of the products listed, indicating a competitive advantage for state-owned and joint-stock bank wealth management subsidiaries [14]

Core Insights - The domestic AI models, particularly ByteDance's Seedance 2.0 and Zhipu's GLM-5, are gaining significant traction, pushing the industry towards the global top tier in AI capabilities [1][8] - The recent surge in stock prices for companies like Zhipu and MiniMax indicates a strong market response to advancements in AI applications [1][2] - Zhipu's decision to raise prices for its GLM Coding Plan reflects robust market demand and the need for enhanced computational resources to maintain service quality [3][5] Industry Trends - The AI application index has seen notable increases, with companies like Deep Technology and Fengyuzhu rising by 10.02% and 9.98% respectively, indicating a shift in industry focus towards computational power [1] - The semiconductor equipment index also rose by 2.02%, with specific companies like Fuchuang Precision and Jingyi Equipment experiencing gains of 11.6% and 5.92% respectively, highlighting the growing demand for AI-related infrastructure [1] Product Developments - Zhipu's GLM-5 has made significant advancements, with its parameters doubling from 355 billion to 744 billion, and pre-training data increasing from 23 terabytes to 28.5 terabytes, enhancing its capabilities [4] - The introduction of multi-modal models like Seedance 2.0 is leading to a dramatic increase in token consumption, with applications in video generation and coding consuming significantly more resources than traditional text-based models [4][8] Competitive Landscape - The competition in the AI coding space is intensifying, with Zhipu's GLM-5 directly targeting Anthropic's Claude Opus 4.5, although it still lags behind in certain advanced capabilities [8] - Chinese companies, including Zhipu and others, are rapidly closing the gap with global leaders in AI, with experts noting that the development pace of Chinese AI models may only be months behind that of the U.S. [9]