[Table_Date] 2026 年 02 月 08 日 [Table_Title] 关注商业航天，坚定算力基建扩张 证券研究报告|行业点评报告 [Table_Title2] 通信 [Table_Summary] 1、关注商业航天，持续进展催化市场关注度上升 当前时点，国内低轨卫星批量发射已进入实质阶段，与商业航天 高度重合，在相关政策、产业驱动，卫星网络作为 6G 网络建设基 础，当前时点建设确定性较强。伴随可回收火箭预研阶段开启首 个发射项目，后续多家商业火箭可回收发射计划有序进展，商业 航天进入加速落地阶段，卫星互联网发展势头迅猛，星座常态化 发射，实现通信容量增大，时延降低，同时有望看到手机宽带直 连卫星与低轨卫星网联智能驾驶系统的相关应用快速落地。卫星 通信、卫星物联网相关受益标的。 2、四部门发文加快培育数据流通服务机构 当前时点，面临近期的流动性担忧以及外围风险厌恶情绪传 导，我们预计市场仍将保持震荡，板块建议相对谨慎，中性配 置。 尽管短期市场对海外加大资本开支背景下 AI 相关收入的疑 惑，以及对 AI 软件板块受挫，但是我们认为目前 AI 发展阶段 仍处于 Scale up 和 Scal ...

当AI发展正从算力建设转向应用落地，一键布局"算力+AI应用"的创业板人工智能ETF（159363）及场 外联接（A类023407、C类023408），更直接受益于AI技术商业化爆发的增长红利。从赛道看，创业板 人工智能约六成仓位布局算力（光模块龙头+IDC龙头），约四成仓位布局AI应用，不止是"算力"核 心，也是真正的"AI应用"代表。 周五（2月6日），A股缩量盘整，创业板人工智能午后翻绿收跌，算力、AI应用双双走低。成份股大面 积飘绿，AI应用龙头昆仑万维领跌超7%，光模块龙头新易盛下跌5%，中际旭创下跌4%；IDC算力租赁 方向相对活跃，铜牛信息、奥飞数据盘中均拉升超5%。 热门ETF方面，双线布局"算力+AI应用"的创业板人工智能ETF（159363）场内收跌1.95%，日线三连 跌，单日缩量成交5.66亿元。单周来看，科技风格回调明显，创业板人工智能ETF（159363）周线十连 阳后首阴，短期调整或属正常技术回踩，中长期产业趋势未变，可关注逢跌布局机会。 IDC算力租赁方面，千问APP因春节活动高峰出现服务卡顿，暴露算力瓶颈，国产算力机会显现！消息 面上，据智通财经报道，阿里千问APP"春节30 ...

Market Overview - On February 6, A-shares experienced a volume contraction and consolidation, with the ChiNext AI sector turning negative in the afternoon, leading to declines in both computing power and AI applications [1][7] - Major components in the AI application sector saw significant losses, with Kunlun Wanwei dropping over 7%, Xinyi Sheng down 5%, and Zhongji Xuchuang falling 4% [1][7] - The ChiNext AI ETF (159363) closed down 1.95%, marking its first decline after ten consecutive weeks of gains, with a daily trading volume of 566 million yuan [1][7] Computing Power Sector - The IDC computing power leasing sector showed relative activity, with companies like Tongniu Information and Aofei Data rising over 5% during the trading session [1][7] - The Qianwen APP from Alibaba faced service interruptions during the Spring Festival promotion, highlighting a computing power bottleneck and revealing opportunities for domestic computing power [9] - National internet giants are engaged in a "red envelope war" to capture traffic for large models, indicating that computing resources are becoming a critical constraint in AI applications [9] AI Applications - The commercial model for AI applications is evolving towards large-scale implementation, with expectations for accelerated commercialization of AI agents by 2026 as the ecosystem matures [10] - The supply of AI agents is currently concentrated among leading model manufacturers, with high usage costs, suggesting that high-value scenarios on both consumer and business sides will be the first to achieve large-scale deployment [10] - The ChiNext AI ETF (159363) is positioned to benefit directly from the commercialization of AI technology, with approximately 60% of its portfolio allocated to computing power and 40% to AI applications [10]

Core Insights - The data center rack industry is a crucial infrastructure component that enhances the management, security, and maintenance of IT equipment, facilitating the planning, design, and construction of data centers [2][5] - The market for data centers in China is projected to grow significantly, reaching a scale of 2,773 billion yuan by the end of 2024, with the number of operational data center racks exceeding 8.8 million [6][8] - By the end of 2025, the market size is expected to reach 3,180 billion yuan, with standard rack numbers surpassing 12 million [6][8] Industry Definition and Role - Data center racks are essential components that support and install various computer parts, providing a secure, reliable, and efficient management structure for IT equipment [2][5] Industry Development Status - The data center rack industry is experiencing rapid growth, driven by the development of the digital economy and the increasing demand for information network data [6][8] - The compound annual growth rate (CAGR) for the number of operational data center racks from 2019 to 2024 is over 20% [6] Industry Chain - The upstream of the data center rack industry includes raw materials such as metals (steel, aluminum, copper), plastics, and rubber, while the midstream involves the production and system integration of data center racks [6] - The downstream applications include enterprise data centers, edge data centers, AI computing centers, and high-performance computing/AI clusters [6] Industry Policy Environment - The data center is recognized as a vital part of the new infrastructure in China, with government policies supporting its investment and construction [10] Industry Competitive Landscape - The high-end market for data center racks is dominated by international giants, while domestic players like Huawei, Inspur, and ZTE are capturing significant market share in the IDC market and large internet companies [11][13] - The rise of domestic enterprises is accelerating the trend of local substitution, leading to increased market concentration [11] Industry Development Trends - The future of data center racks is expected to focus on "intelligent, liquid-cooled, and high-density" physical infrastructure units, integrating power supply, efficient heat dissipation, comprehensive sensing, and intelligent management capabilities [15]

Core Viewpoint - The AI sector experienced a significant downturn, with the ChiNext AI ETF (159243) dropping by 5.27%, while major component stocks like Wangsu Science & Technology and Guanghui New Network fell over 9%. Despite this, the ETF saw a net inflow of approximately 26 million CNY, indicating a "buy the dip" mentality among investors [1][2]. Group 1: Market Dynamics - The recent sharp adjustment in the AI sector is attributed to a combination of external shocks, internal rotations, and changes in market ecology. A notable trigger was the collective plunge of AI tech stocks in the US market, raising concerns that emerging AI tools could disrupt traditional software business models [3][4]. - The market is witnessing extreme "high-low switching" and "risk aversion" characteristics, with funds moving away from high-valuation tech growth sectors to defensive assets amid macroeconomic volatility and geopolitical uncertainties [3][4]. Group 2: Investment Sentiment - The investment landscape for artificial intelligence has shifted to a new phase, where the focus is on the commercialization and actual performance of companies rather than speculative narratives. Stocks that require significant capital expenditure or are still in the investment phase are facing increased short-term profit pressures, leading to a drastic correction in valuations [3][4]. - The current extreme adjustment in the AI sector serves as a "stress test" influenced by short-term emotions, funding behaviors, and the evolution of the industry stage. This adjustment has not undermined the long-term value of the sector but has highlighted the certainty of industry trends and the support from policies [4].

算力租赁概念盘中持续调整，宏景科技、网宿科技、开普云跌超10%，此前大位科技跌停，卓易信息、 利通电子、奥飞数据、世纪恒通、优刻得等跌幅靠前。 ...

Group 1 - The core viewpoint of the news highlights a significant sell-off in AI software stocks in the US market, which has impacted the A-share market, particularly affecting AI-related ETFs and stocks [1] - The A-share market is experiencing increased volatility with a notable divergence among sectors, but the long-term upward trend remains intact, supported by ongoing policies for the domestic AI industry [2] - The recent sell-off in US tech stocks, especially software companies, has raised concerns about the potential threat of AI technology replacing core business functions [1] Group 2 - The AI sector is currently in a rapid development phase, with expectations for increased capital expenditure in the global computing power sector as domestic support policies continue to be implemented [2] - The Spring market rally is expected to be followed by a consolidation phase, where the focus will be on clarifying industry trends and digesting valuations [2] - The long-term outlook remains positive for sectors such as overseas computing power chains, AI applications, energy storage, robotics, and commercial aerospace [2] Group 3 - The Huaxia AI ETF (159381) tracks an index with nearly 50% weight in CPO, covering domestic software and AI application companies, and has a low fee rate of 0.20% [3] - The Huaxia Communication ETF (515050) has a scale exceeding 8 billion yuan, covering key areas such as optical modules and computing power infrastructure, with a combined weight of over 76% in CPO and CPB concept stocks [3] - The Huaxia Cloud Computing ETF (516630) focuses on domestic AI software and hardware computing power, with a combined weight of 83.7% in computer software, cloud services, and computer equipment, also featuring a low fee rate of 0.20% [3]

Group 1 - The core viewpoint of the news is that Aofei Data's stock performance and financial metrics indicate a mixed outlook, with significant trading activity and growth in revenue and profit, but also notable changes in shareholder composition and financing status [1][2][3]. Group 2 - On February 2, Aofei Data's stock fell by 2.31%, with a trading volume of 1.757 billion yuan. The financing buy-in amount was 169 million yuan, while the financing repayment was 181 million yuan, resulting in a net financing outflow of 12.0755 million yuan. The total financing and securities balance reached 1.228 billion yuan [1]. - As of February 2, the financing balance of Aofei Data was 1.225 billion yuan, accounting for 5.34% of the circulating market value, which is above the 60th percentile level over the past year, indicating a high level of financing [1]. - In terms of securities lending, Aofei Data had a repayment of 12,200 shares and a sell amount of 51,000 shares, totaling 118,700 yuan. The remaining securities lending volume was 133,100 shares, with a balance of 3.0986 million yuan, which is below the 30th percentile level over the past year, indicating a low level [1]. - As of September 30, Aofei Data had 113,600 shareholders, an increase of 9.06% from the previous period, while the average circulating shares per person decreased by 8.31% to 8,667 shares [2]. - For the period from January to September 2025, Aofei Data achieved operating revenue of 1.824 billion yuan, a year-on-year increase of 15.33%, and a net profit attributable to the parent company of 145 million yuan, a year-on-year increase of 37.29% [2]. - Since its A-share listing, Aofei Data has distributed a total of 113 million yuan in dividends, with 52.9344 million yuan distributed over the past three years [2]. - As of September 30, among the top ten circulating shareholders, Hong Kong Central Clearing Limited was the second-largest shareholder with 12.9293 million shares, a decrease of 777,700 shares from the previous period. New entrants included Huaxia CSI 1000 ETF and GF CSI 1000 ETF [3].

智通财经APP获悉，东吴证券发布研报称，Counterpoint Research预计，2026年Q1存储价格还将上涨 40%至50%，Q2继续上涨约20%。供需端视角上，供给侧硬件成本提升，需求侧AI Agent爆款不断推陈 出新，全球大厂纷纷下场，云作为AI的基础设施，海外大厂谷歌(GOOGL.US)和亚马逊(AMZN.US)已经 开始涨价，为多年来首次。随着春节临近，国内AI应用需求逐步提升，国内云厂也会提价，AI云产业 链有望引来投资机会。 东吴证券主要观点如下： 海外大厂云已经开始涨价 光纤涨价不断。在中国市场，G.652.D光纤价格相较于2025年11月大幅上涨约80%。2026年1月，该类型 单模光纤价格更是创下近七年新高，平均价格突破40元/芯公里，部分厂商报价甚至高达50元/芯公里， 仅一个月内涨幅就超过75%。 相关标的 算力租赁厂商：宏景科技(301396.SZ)、协创数据(300857.SZ)等。云厂商：网宿科技(300017.SZ)、优刻 得-W(688158.SH)、青云科技(688316.SH)、阿里巴巴-W(09988)等。CPU：澜起科技(688008.SH)、海光 信息(6 ...

Investment Rating - The report maintains an "Overweight" investment rating for the computer industry [1] Core Insights - Cloud service providers like Google Cloud and Amazon Web Services have announced price increases, marking a significant shift in the market [5][10] - The demand for AI Agents is accelerating, with notable growth in applications and user engagement [11][13] - Supply-side pressures are leading to rising hardware costs, including storage and CPU prices [16] - The report suggests that the upcoming Chinese New Year will boost domestic AI application demand, potentially leading to further price increases in cloud services [17] Summary by Sections Price Increases by Major Cloud Providers - Google Cloud announced price hikes effective May 1, 2026, with data transfer costs in North America increasing from $0.04/GB to $0.08/GB (100% increase), Europe from $0.05/GB to $0.08/GB (60% increase), and Asia from $0.06/GB to $0.085/GB (42% increase) [10] - Amazon Web Services raised prices for its EC2 machine learning capacity blocks by approximately 15% [10] Demand Side: Acceleration of AI Agent Adoption - The rapid growth of AI Agents is exemplified by the success of Moltclaw, which has seen exponential user growth [11] - Google has integrated its Gemini 3 into Chrome, transforming it into a comprehensive AGI platform for 3.8 billion users [13] Supply Side: Hardware Price Increases - Storage prices are expected to rise by 40% to 50% in Q1 2026 and by about 20% in Q2 2026 [16] - Intel and AMD plan to increase server CPU prices by 10-15% due to high demand and limited supply [16] - Prices for G.652.D optical fibers in China have surged by approximately 80% compared to November 2025, with some prices exceeding 50 yuan per core kilometer [16] Investment Recommendations - The report identifies potential investment opportunities in the AI cloud industry, particularly as hardware costs rise and demand for AI applications increases [17] - Recommended companies include cloud service providers like Wangsu Science and Technology, Yike Technology, and Alibaba, as well as CPU manufacturers like Lianqi Technology and Longxin Zhongke [18][19][20]