今日，市场情绪继续高涨，沪深两市成交额超2.34万亿元，同类规模领先的云计算ETF汇添富(159273)探底回升，盘中成交额超1.6亿元。资金面上，云计算 ETF汇添富（159273）溢价高企达0.48%，盘中大举吸金超1.28亿元，截至8月25日，上市14天已有12天获净流入，累计"吸金"超4亿元！同类规模、流动性持 续领跑！ | 分时 多日 1分 5分 15分 30分 ▼ | | | | | 综合屏 F9 前复权 超级叠加 画线 工具 令 ② > | | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | 159273.SZ[云计算ETF汇添富] | | | | 2025/08/15 收1.064 幅0.66%(0.007) 开1.052 高1.069 低1.052 均应 | | | | l ol | | MA5 1.041↑ | | | | | 2025/08/06-2025/08/26(15日)▼ ■ | | | SZSE | | 1.245 | | | | | | 1.240 | | 净值元 | | | | | | | | | ...

点击上方"公众号"可订阅哦！ 蒋佳霖/孙乾/杨本鸿/陈鑫/张旭光/杨海盟/桂杨 兴业证券计算机小组 周观点 风险提示： 板块业绩不达预期；下游需求存在景气度不稳定的风险；科技创新可能带来短期业绩下降。 本周观点聚焦 加仓核心科技龙头，重点布局国产算力链 1、本周 观 点： 加仓核心科技龙头，重点布局国产算力链 加仓核心科技龙头，继续加大板块配置。 下周板块迎来半年度业绩的密集披露，板块赛道龙头一方面具备业绩优势，同时核心受益于国内科技产业景气发展，在风险偏 好上行阶段建议重点加仓。同时展望 9 月份，迎来重要事件关口，科技自强，科创崛起有望成为重要风口，计算机板块科技属性占优，建议继续加大板块配置。 2、深度跟 踪 ： 全联接大会在即，关注华为产业链机遇 Deepseek 加持国产芯片，重点布局国产算力链。 本周 Deepseek-V3.1 发布，其使用 UE8M0 FP8 参数精度是针对即将发布的下一代国产芯片设计，是对国产芯片生 态完善的重要助力，同时近期国产新一代 AI 芯片性能、开发生态等均有明显提升，核心龙头厂商高成长可期，建议围绕国产算力产业链进行重点布局。 本文第二章，对华为产业链进行了专题更新 ...

今日早盘芯片股持续走高，截至午间收盘，海光信息涨超17%。寒武纪涨超12%，中芯国际、盛科通信纷纷大涨，盛美上海涨超17%，北方华创、华海清 科、芯源微、东芯股份等多股涨超5%。其中寒武纪股价一度逼近1200元关口，总市值一度逼近5000亿，截至午间收盘，该股7月以来涨幅已近94%。 指数方面，截至午间收盘，科创50指数大涨超5%。场内ETF方面，多只芯片相关ETF涨超6%。 大爆发的原因是什么？ 一是从市场交易层面看，4月中旬以来，AI硬件持续走强，带头的是北美算力链，胜宏科技、中际旭创、新易盛等龙头股这段时间迭创新高。与此同时， A股指数不断创出反弹新高，风险偏好提升。在此背景下，国产算力链存在较大的补涨需求。 二是寒武纪带来的示范效应。8月12日，寒武纪"20CM"涨停，股价创历史新高。8月20日，公司股价收盘站上1000元。今天上午，股价继续大涨，与贵州 茅台的差距不断缩小。在这个过程中，市场重塑对国产算力链的认知。 三是DeepSeek公众号8月21日宣布，正式发布DeepSeek-V3.1。DeepSeek-V3.1使用了UE8M0 FP8 Scale的参数精度。DeepSeek官微在置顶留言中透 ...

"硬科技"上午全面大爆发，除了寒武纪，海光信息、中科曙光、中芯国际、盛美上海等个股大涨，北方华创、中兴通讯、中微公司、澜起科技、浪潮信 息、芯原股份等龙头股均上涨。 在"硬科技"带动下，今天上午，上证指数延续涨势，逼近3800点。截至上午收盘，上证指数上涨0.67%，报3796.36点，深证成指上涨1.32%，创业板指上 涨2.56%。市场半日成交超1.53万亿元。 "硬科技"爆发 今天上午，"硬科技"爆发，半导体产业链全面走强，AI应用端和硬件端联手上涨。科创50指数上涨5.25%。 | ■ 688256 | | | | | --- | --- | --- | --- | | 杰华特 | 36.76 | 10.32% | 97亿 | | 融 688141 | | | | | 盛科通信-U | 126.01 | 9.36% | 254亿 | | 融 688702 | | | | | 华峰测控 | 162.45 | 9.32% | 220亿 | | 融 688200 | | | | | 恒炼股份 | 47.66 | 9.26% | 22.9 乙 | | 融 688416 | | | | | 中心国际 | 98. ...

Group 1 - The market is strengthening, with large-cap stocks leading the rally, and domestic computing power chains showing strong performance [1] - Attention is drawn to two aspects: the growth rate of financing balances, which may indicate that leveraged funds are in an unfavorable position, and the potential for low-position sectors to rebound [1] - The consumer sector, currently at a low position, may also experience a rebound due to favorable policies on consumer loan interest subsidies [1] Group 2 - If the rebound in low-position sectors is accompanied by a decline in the anti-involution and infrastructure mainline, caution is advised regarding potential market volatility [1] - The market turnover remains at a level of 1.9 trillion yuan, and if the market continues to rise but turnover decreases, it may signal the nearing end of the rally [1]

Group 1 - The A-share market continued its "small spring" trend in July, with the Shanghai Composite Index surpassing 3600 points and a year-to-date increase of over 6.6%, led by sectors such as building materials, rare earths, and innovative pharmaceuticals [1] - Despite the overall market performance, some funds experienced significant losses, with certain products down nearly 20% year-to-date, highlighting a stark contrast to the market's gains [1] - The performance of actively managed funds has been disappointing, with fund managers failing to demonstrate effective operational capabilities in the face of market fluctuations [1] Group 2 - The Qianhai Kaiyuan AI-themed mixed fund reported a year-to-date loss of 19.15% as of the end of July, ranking low among its peers [2] - This fund underwent a significant portfolio adjustment at the beginning of the year, shifting from established AI leaders to smaller chip companies, which has been viewed as a "dark horse gamble" [2] - The fund's strategy has been criticized for not including leading AI companies, and its performance continued to decline despite further changes in the second quarter [2] Group 3 - Star fund manager Qu Yang stepped down in June after managing the fund for nine years, with the fund's assets shrinking from 600 billion yuan at its peak in 2021 to 144 billion yuan [3] - The fund's return during the dual management period with Wei Chun was -41.32%, contrasting sharply with the 93.3% return during Qu Yang's sole management [3] Group 4 - The Jianxin China Manufacturing 2025 fund, managed by Sun Sheng, also faced a loss of over 15% year-to-date, attributed to poor timing in its investment strategy [4] - The fund made significant changes to its top holdings, reflecting a shift towards computing infrastructure, but suffered from a market pullback in the first quarter [4] - The fund's performance continued to lag in the second quarter, with a net asset value decline of 6.43% due to weaker-than-expected domestic AI development [4] Group 5 - The Vanguard funds managed by Liu Zhiqiang also revealed inconsistencies between strategy and performance, with both funds experiencing net value declines exceeding 14% in the first quarter [5] - The funds claimed to maintain a flexible strategy for stable returns, yet their actual performance significantly lagged behind the benchmark [5] - Many of these funds are labeled as "thematic funds," but their performance benchmarks are tied to broad market indices, raising questions about their investment focus [5] Group 6 - Industry experts noted that aggressive portfolio adjustments can lead to repeated mistakes, particularly for funds that have not aligned their strategies with market trends [6] - Many underperforming funds made extensive adjustments in the first quarter, attempting to follow market shifts, but ended up underperforming their benchmarks [6] - The trend of frequent and aggressive adjustments has resulted in further declines in net asset values, illustrating the risks of misjudging market directions [6] Group 7 - Despite the challenges faced by some active funds, there are still a number of successful actively managed funds that have generated significant excess returns through deep industry insights and precise stock selection [7] - Investors are advised to adopt a more rational "core-satellite" strategy, combining broad market index ETFs for stability with selectively chosen active funds for potential alpha returns [7] - This structured approach can help mitigate risks while allowing for a more measured response to market fluctuations and the short-term volatility of certain active funds [7]

Core Viewpoint - The performance of Jianxin Inherent Power A fund has significantly lagged behind the market, with a year-to-date net asset value decline of 12.6% compared to a 4.68% increase in the CSI 300 index, indicating poor investment decisions and high turnover rates [1][7]. Fund Performance - Jianxin Inherent Power A has experienced a net asset value decline of 12.6% as of July 22, 2025, underperforming its benchmark by approximately 16.4 percentage points [7]. - The fund's performance has been volatile, with a decline of over 40% since the current fund manager, Sun Sheng, took over in October 2021 [5]. - Historical performance shows the fund had losses of 29.79% in 2022 and 21.94% in 2023, with a slight gain of 17.05% in 2024 [6]. Investment Strategy - The fund has a high turnover rate, with stock turnover rates of approximately 851.64%, 1140.08%, and 619.19% for the years 2022, 2023, and 2024, respectively, indicating aggressive trading strategies [8]. - As of the end of Q2 2025, the fund's stock holdings accounted for 85.72% of total assets, with the top ten holdings representing 54.94% of net asset value [8]. - The fund's investment style is aggressive, focusing heavily on the domestic computing power chain sector, which has underperformed in the second quarter of 2025 [9]. Manager Overview - Sun Sheng has managed Jianxin Inherent Power A since October 2021 and oversees a total of four funds, all of which have shown declines in net asset value this year [2][11]. - The overlapping stock holdings among the funds managed by Sun Sheng have contributed to the overall underperformance, with significant commonality in top holdings [12].

Group 1: Humanoid Robots and AI Solutions - UBTECH announced the launch of a commercial humanoid robot solution centered around the Walker C robot, integrating advanced AI models and navigation algorithms for applications in exhibitions, supermarkets, and transportation hubs [2] - The emergence of AI companies like DeepSeek is driving the development of general-purpose humanoid robot models, indicating a strong trend towards industrial applications of humanoid robots [2] - The humanoid robot industry is entering a phase of rapid development, with commercial applications becoming increasingly viable, suggesting potential investment opportunities in domestic component manufacturers [2] Group 2: AI Infrastructure Investments - Major tech companies such as Amazon, Microsoft, Oracle, and Meta are making significant investments in AI infrastructure, with total investments reaching up to tens of billions of dollars [3] - The demand for digital infrastructure driven by AI is leading to a sustained increase in global data center capacity, with the domestic computing power sector expected to recover due to policy support and technological upgrades [3] - The long-term growth of AI-driven infrastructure is anticipated, with opportunities arising from technological upgrades and domestic replacements in sectors like optical modules, switches, and cooling systems [3] Group 3: Robotics and AI Market Potential - Google DeepMind introduced the Gemini Robotics On-Device model, which can run locally on robotic devices, enhancing their adaptability to new tasks without needing constant internet connectivity [4] - The competition among major tech firms in the field of embodied intelligence is expected to unlock a trillion-dollar market, as robots transition into the embodied intelligence era [5] - The intersection of humanoid robots and AI represents a critical point in technological advancement, with significant breakthroughs anticipated in application, cost, and software development [5] Group 4: Tungsten Market Dynamics - The strategic value of tungsten is increasing, with supply-demand tightness expected to persist, leading to a bullish market trend and potential price increases [6] - Tungsten is essential in high-end manufacturing and is considered a strategic resource in China, with strict controls on its mining [6] - The global tungsten supply is projected to grow at a CAGR of 2.57% from 2023 to 2028, while demand is expected to rise due to emerging industries like photovoltaics and robotics [6]

Summary of Key Points from the Conference Call Industry Overview - The conference call focuses on the **domestic computing power industry chain** in China, particularly highlighting the **switching equipment sector** which has recently reached a profitability inflection point [1][2][5]. Core Insights and Arguments - The **domestic data center sector** remains stable despite a slowdown in new data center tenders, with major internet companies showing steady rack installation rates [2][4]. - The **demand for optical modules** is rapidly increasing, particularly for 800G and 1.6T modules, driven by major players like BBAT and Huawei, indicating strong investment in AI computing power [3][10][11]. - The **profitability of the domestic computing power chain** has improved due to scale effects, with significant increases in shipment volumes and revenues for domestic servers, optical modules, and switches [2][5]. - The **Ethernet switch market** in China is primarily driven by demand from large internet companies, which prefer SDN (Software Defined Networking) and white-box switches over traditional IB (InfiniBand) switches [6][7]. - **White-box switches** offer configurability and rapid iteration capabilities, making them more suitable for the dynamic traffic needs of data centers [8]. - The **high gross margins** in the switching industry are attributed to hardware production and collaborative R&D with internet companies, which involves complex software technologies [9]. Additional Important Insights - The **data center REITs projects** have been approved, providing a benchmark for asset valuation, with the REITs market offering around 15 times EV/EBITDA valuation levels [3][19][20]. - The **AI training and inference demand** is a key driver for the growth of the domestic computing power chain, with major companies like Alibaba and Tencent increasing their AI computing investments [11]. - The **domestic data center industry** faces challenges such as chip shortages, which have slowed down order fulfillment, but long-term prospects remain positive as these issues are expected to be resolved [17]. - **Policy support** for the data center industry includes project approvals and energy consumption indicators, which helps maintain a balanced supply-demand dynamic [18]. - The **current market conditions** suggest a favorable outlook for the domestic computing power chain, with expectations of improved performance in the upcoming reporting season [21]. Recommendations - Companies such as **Hua Gong Technology** are recommended for investment due to their leading position in the optical module market and expected improvements in profitability as demand for 400G and 800G modules rises [14][15]. - Attention is also drawn to **Xingwang Ruijie and Ruijie Networks** in the switching sector, which are expected to benefit from the industry's profitability improvements [9].

Core Viewpoint - The article discusses the impact of the renewed tariff war initiated by the Trump administration on global trade, emphasizing that the essence of this conflict is a technological battle. It highlights a significant breakthrough in China's tech industry with the DeepSeek-R1 model, which reduces inference costs to 30% of the industry average while maintaining a parameter scale of 175 billion, shifting the narrative of the global AI competition from a mere technological arms race to a comprehensive contest of engineering capabilities [2][3][4]. Group 1: Fund Performance Overview - The article presents an overview of the top-performing technology-themed funds for the first quarter, noting that the average return for these funds was 4.1%, with the top 10 achieving an average return of 20.1%, significantly outperforming the Shanghai Composite Index and the CSI 300 Index [8][9]. - The top-performing fund, China Europe Intelligent Manufacturing Mixed A (015143.OF), achieved a return of 26.72% in the first quarter, with a one-year performance of 71.15%, ranking 26 out of 4217 in its category [7][9][10]. - Other notable funds include Jianxin Yuli Flexible Allocation Mixed (002281.OF) and Ping An Research Preferred Mixed A (017532.OF), with returns of 20.89% and 19.64% respectively [18]. Group 2: Fund Manager Strategies - Fund manager Shao Jie of China Europe Intelligent Manufacturing Mixed A has increased investments in AI hardware and applications, focusing on companies like Langqi Technology and Tencent Holdings, which contributed significantly to the fund's performance [11][12][13]. - Jianxin Yuli Flexible Allocation Mixed has shifted its top holdings to include companies in AI computing and hard technology, indicating a bullish outlook on manufacturing upgrades and long-term development in these sectors [25][22]. - The article notes that the performance of these funds is under scrutiny, as many stocks are trading at high valuations, and any underperformance in AI applications could lead to volatility in tech stocks [17]. Group 3: Consumer Fund Performance - The article contrasts the performance of consumer funds, which generally faced challenges, with technology funds, highlighting that the main consumer industry index fell by 1.1% in the first quarter [51][52]. - Despite the overall decline, two consumer-themed funds, Yuanxin Yongfeng Consumption Upgrade (004934.OF) and Zhongyin Xin New Consumption Growth (010965.OF), managed to achieve returns of 8.9% and 7.4% respectively by heavily investing in technology stocks [53][57]. - This trend of consumer funds investing in technology stocks raises concerns about potential misalignment with their stated investment objectives, indicating a broader industry issue where fund managers may be compelled to chase performance at the expense of adhering to their mandates [61][63].