2月26日，海外算力表现较强，通信ETF开盘后持续上行，收涨2.58%。创业板人工智能ETF同步上行，收涨1.73%。 资料来源：Wind 消息面上，2月26日凌晨英伟达发布FY26Q4财报。 从FY27Q1的指引看，同样表现亮眼。 -营收780亿美元±2%（不包含中国收入），超过市场预期的726亿美元。 -毛利率74.5%-75.5%，高位维持。 资料来源：nvdia 此外，英伟达GTC大会将在3月16日盛大开幕。众所周知，GTC大会是英伟达最为重要的发布会。从目前市场预期看，LPU、CPO、VR200 Superpod等产品 可能迎重磅更新或发布。 -单季总营收 681.3亿美元，qoq+20%，yoy+73%，超过市场预期的656.84亿美元。 -单季数据中心营收达623亿美元，qoq+22%，yoy+75%，高于市场预期的606.2亿美元。 -毛利率达75%，环比增长1.6%，同比增长2%。 -净利润从去年同期的 221 亿美元增至 430 亿美元（每股 1.76 美元）。 投资人应当充分了解基金定期定额投资和零存整取等储蓄方式的区别。定期定额投资是引导投资人进行长期投资、平均投资成本的一种简单易行的 ...

Market Performance - Overseas computing power recovery led to a rise of over 4% in the Communication ETF (515880) and the ChiNext AI ETF (159388) during trading today [1] Factors Driving the Increase - Recent US earnings reports indicate several key points: - Capital expenditures are accelerating, with Google projecting 2026 capital spending between $175 billion and $185 billion, nearly doubling year-on-year; Meta's full-year capital expenditure guidance is $115 billion to $135 billion, a 73% year-on-year increase; Amazon's guidance is $200 billion, a 53% increase year-on-year [2] - Microsoft did not provide full-year guidance but noted a seasonal quarter-on-quarter decline, likely due to factors like financing leases. The market previously expected a 42% growth in North American CSP capital expenditures for 2026, but recent earnings reports show capital expenditure growth significantly exceeding expectations [2] CPO Penetration and Market Sentiment - CPO penetration is exceeding expectations, but attention is needed on scale-up and scale-out dynamics. The consensus is that scale-up will dominate CPO, while scale-out remains led by pluggable optical modules. The domestic optical module leaders may secure some orders in the scale-up segment, indicating that CPO penetration in scale-up represents incremental growth rather than a replacement [3] - Market sentiment has rebounded from previous lows, with US markets recovering last Friday, influencing A-shares today. Both markets have faced issues related to funding and sentiment, with A-shares experiencing declining trading volumes since late January [3] Future Outlook - US earnings reports reaffirm the certainty of AI, with ongoing shortages in computing power. Google reported that Gemini 3.0 is the fastest model in its history, with over 750 million monthly active users for Gemini applications. Google also announced a partnership with Apple to develop the next-generation Apple foundational model. The management indicated that investments in AI infrastructure will gradually increase throughout the year, with a continued tight supply of computing power expected [4] - The focus remains on core segments like optical modules and servers, which are positioned at the heart of the global AI industry chain. With capital expenditures for 2026 significantly exceeding expectations, the outlook for optical modules and servers is strengthening, making Communication ETF (515880) and ChiNext AI ETF (159388) attractive for interested investors [4]

基本面无明显利空，今日回调主因cpo引发情绪面担忧，或可重视低位布局机会。 （1）科技行业技术的进展和迭代很正常，目前光模块龙头也有CPO技术积累，有望凭借产业链卡位和 技术优势首先受益。目前国内可插拔光模块大厂已建立了良好的客户粘性，不管是对于英伟达还是下游 的云厂。除可插拔光模块具备核心份额之外，CPO方面也有深厚技术积累，后续若CPO渗透，国内大厂 也有望承接多数订单，且CPO集成度较高，渗透会带来毛利率的改善。 （2）2026年光模块供不应求几成定局，2027年预期也在不断打入。由于芯片的出货及1.6T光模块的快 速渗透，2026年光模块市场空间或成长数倍，2027年也有望保持较高增速。对光模块龙头厂商而言，当 前正是利润爆发的黄金期。后续季报、GTC大会等催化良多，基本面十分强硬。 CPO的落地和渗透是一个漫长的过程，且国内大厂也具备相关技术能力，CPO影响有限。此外， 2026~2027年内1.6T光模块的景气度十分高企，催化良多。综合而言，今日算力下跌主因情绪影响，不 仅不会改变光模块行业的高成长趋势，反而提供更舒服的布局时点，或可继续关注通信ETF （515880）、创业板人工智能ETF（15 ...

12月8日算力板块表现强势，延续前期上涨势头，创业板人工智能ETF（159388）收涨5.51%，通信ETF（515880）收涨5.49%。 资料来源：wind 事件方面，谷歌全栈AI生态持续发酵。谷歌加速了产业链进展，谷歌TPU的出货量预期持续上调。谷歌的成长，不仅在硬件侧带来更多的 计算芯片出货增量，而且在模型、应用等方面极大推动了产业成长。随着大模型持续不断迭代跃进，巨头的AI军备竞赛持续，必然会拉动 整体算力需求的进一步增长（无论NVGPU还是谷歌TPU）。A股光模块、服务器等厂商"卖铲人"逻辑较强，不管在芯片端的竞争中时谁获 胜，板块都将受益。 投资人应当充分了解基金定期定额投资和零存整取等储蓄方式的区别。定期定额投资是引导投资人进行长期投资、平均投资成本的一种简 单易行的投资方式。但是定期定额投资并不能规避基金投资所固有的风险，不能保证投资人获得收益，也不是替代储蓄的等效理财方式。 无论是股票ETF/LOF基金，都是属于较高预期风险和预期收益的证券投资基金品种，其预期收益及预期风险水平高于混合型基金、债券型 基金和货币市场基金。 基金资产投资于科创板和创业板股票，会面临因投资标的、市场制度以及交易 ...

Core Viewpoint - The AI infrastructure capital expenditure is expected to drive significant growth in the sector, with major cloud providers projected to increase their capital spending substantially in the coming years [1][2]. Group 1: Market Performance - The computing sector continued its upward trend on December 1, with the Communication ETF (515880) rising by 2.84% and the ChiNext AI ETF (159388) increasing by 2.45% [1]. - The anticipated capital expenditure for the four major U.S. cloud providers (Google, Microsoft, Meta, Amazon) is projected to reach $367 billion in 2025, representing a year-on-year growth of 59%, with a further increase to $495 billion in 2026, reflecting a 35% growth [1]. Group 2: Demand Drivers - AI infrastructure capital expenditure remains a strong driving force for demand, with the ongoing advancements in AI models and hardware, such as Gemini3 and TPU7, contributing to increased computational needs [1]. - The competition among major players in the AI space is expected to further boost overall demand for computing power, benefiting various sectors including optical modules and servers in the A-share market [1]. Group 3: Investment Strategy - The market outlook for artificial intelligence remains positive, with suggestions for investors to consider low-position opportunities in the Communication ETF (515880) and ChiNext AI ETF (159388) through methods like dollar-cost averaging [2].

Market Overview - The computing sector experienced a significant rise, with the Communication ETF (515880) increasing by 7% and the ChiNext AI ETF (159388) rising over 6% [1][2]. Factors Driving the Market - Increased expectations for interest rate cuts have emerged, as several Federal Reserve officials, including allies of Powell, expressed support for a potential 25 basis point cut in December, raising the likelihood from approximately 40% to 80% [2]. - Google's recent release of the Gemini 3 model, which has outperformed its predecessors in various benchmarks, has contributed to positive market sentiment [2][4]. AI Model Performance - The Gemini 3 Pro model has shown superior performance across multiple benchmarks, including achieving 95% in mathematics and 91.9% in scientific knowledge assessments, indicating its advanced capabilities compared to competitors [3]. Hardware Developments - Google's TPU v7 (Ironwood) has exceeded expectations in terms of memory capacity and energy efficiency, posing a competitive challenge to NVIDIA's offerings. Google anticipates a doubling of computing power every six months to meet rising AI service demands [5]. Market Outlook - There is a gradual improvement in risk appetite as pessimistic expectations are being absorbed by the market, with strong support levels in the A-share market despite rapid style shifts [6]. - While the competitive landscape for computing chips may evolve, NVIDIA is expected to maintain revenue growth due to its existing market share and new demands from sovereign AI projects [6]. - The AI industry is progressing rapidly, with significant investments expected to sustain growth, although short-term volatility should be monitored [7].

Core Viewpoint - The communication industry is experiencing an upward trend driven by the powerful engine derived from the AI wave, with significant performance differentiation across various segments [1]. Communication Industry Summary - In Q3 2025, the total revenue reached 632.1 billion yuan, representing a year-over-year increase of 3.7% [2]. - The net profit for the same quarter was 60.305 billion yuan, showing a year-over-year growth of 12.3% [2]. - Profitability metrics for Q3 2025 indicated a gross margin of 27.97% and a net margin of 9.54%, with a gross margin decrease of 0.40 percentage points and a net margin increase of 0.73 percentage points year-over-year [2]. Optical Module Industry Summary - The Wind Optical Module Concept Index reported a revenue of 29.590 billion yuan in Q3 2025, reflecting a year-over-year increase of 56.0% [5]. - The net profit for the optical module sector in Q3 2025 was 7.336 billion yuan, which is a remarkable year-over-year increase of 118.8% [6]. - Profitability for the optical module sector showed a gross margin of 39.07% and a net margin of 24.79%, with increases of 5.59 percentage points and 7.12 percentage points year-over-year, respectively [7]. Inventory Insights - As of the end of Q3 2025, the inventory balance stood at 33.212 billion yuan, marking a year-over-year increase of 47.6% and a quarter-over-quarter increase of 7.44% [8]. Investment Opportunities - Investors are encouraged to consider the communication ETF (515880) and the ChiNext AI ETF (159388) as potential investment vehicles in light of the strong performance in the optical module sector [1].

Core Insights - The rapid demand for high-quality and convenient AI tools is reflected in the popularity of applications like Nano Banana, which significantly boosts the demand for GPU and accelerated hardware computing power [1] - The implementation of the "Artificial Intelligence +" action plan emphasizes strengthening foundational support capabilities and continuously activating the demand for AI infrastructure, solidifying the technological base for industry development [1] - With accelerated global AI investments and domestic replacements, the industry is transitioning from performance recovery to a valuation recovery phase in a bull market, indicating a clear long-term growth logic [1] Industry Overview - The ChiNext AI ETF (159388) has seen a net inflow of 22 million units, indicating strong capital interest in AI assets [1] - The ChiNext AI Index (970070), which the ETF tracks, has a 20% fluctuation and includes listed companies involved in software development, intelligent hardware manufacturing, and big data processing, reflecting the overall performance of AI-related companies in the ChiNext market [2] - Major companies like Alibaba are making significant investments in AI and cloud infrastructure, creating a closed-loop logic of "AI investment - cloud growth - commercialization," marking a critical development window for domestic computing power [1]

Core Insights - The article highlights a significant inflow of 7 million units into the ChiNext AI ETF (159388), indicating strong investor interest in AI assets [1] - The ChiNext AI ETF tracks the ChiNext AI Index (970070), which has experienced a 20% fluctuation, reflecting the performance of listed companies involved in software development, smart hardware manufacturing, and big data processing related to AI technologies [1] - The index focuses on the innovation capabilities and growth potential of China's emerging technology industry, encompassing representative enterprises in the AI sector [1]

Group 1 - The core viewpoint emphasizes the sustained optimism in the technology sectors such as communications and semiconductors in the Chinese market, driven by favorable policies and increasing domestic capabilities in AI infrastructure [1][2] - The policy document outlines ambitious targets for AI application penetration rates, aiming for over 70% by 2027 and 90% by 2030, which is expected to catalyze domestic computing power [1] - TrendForce predicts a significant shift in the Chinese AI server market, with the proportion of foreign-sourced chips decreasing from 63% in 2024 to 42% by 2025, while local chip suppliers are expected to capture 40% of the market, indicating a strong trend towards domestic substitution [1] Group 2 - The current A-share market benefits from ongoing liquidity and strong performance, with clear advantages in the AI industry chain, including established business models and significant market size [2] - The AI industry chain is experiencing positive marginal changes, with accelerated domestic computing power development, making it a primary focus for asset revaluation in the A-share market [2] Group 3 - In terms of product allocation, it is recommended to balance investments between North American and domestic computing power, utilizing strategies such as grid trading and regular investment [3] - The North American computing power segment is represented by a communication ETF (515880), where over 75% is allocated to "optical modules, servers, copper connections, and optical fibers," indicating a solid fundamental basis for computing hardware [3] - For domestic computing power, the recommended ETF is the Guotai Sci-Tech Chip ETF (589100) for those considering elasticity, while the Chip ETF (512760) is suggested for balanced holdings [3]