算力平权，国产 AI 力量崛起 20250512 摘要 • 腾讯和阿里巴巴加速数据中心建设，腾讯已下单 10 万台 910C 服务器， 阿里巴巴租用 1GW 数据中心并推动巴拿马电源项目，利好中恒电气等相 关企业。 • 新元股份否认三星流片限制传言，已与龙头云厂商签署合作协议，涉及设 计费 5-7 亿元和 IP 授权 5-10 亿元，并在汽车芯片设计服务方面与理想汽 车等合作，定增计划值得关注。 • 中芯国际先进制程稳步推进，从 14 纳米到 N+2 节点，逐步接近台积电 7 纳米工艺。2024 年第一季度少数股东权益数据显示先进制程需求崛起， 预计 2025 年第二季度开始持续乐观。 • 升腾系列产品性能提升显著，C 型号 FP16 算力接近 H100 水平，大幅增 强 AI 训练及云计算应用能力，预计 2025 年出货表现良好，有望满足市场 需求。 • DMS 模型优化提升训练效率 40%，缩小与国际高端算力差距。2025 年 第一季度少量出货，预计第二季度末至第三季度初大规模出货，达到万卡 级别，主要客户包括字节跳动等 CSP 厂商。 Q&A 中美关税和解对海外和国产算力市场的影响如何？ 中美关税和解对 ...

Summary of Conference Call Notes Industry or Company Involved - The discussion primarily revolves around the public fund industry in China and its impact on the broader financial market. Core Points and Arguments - **Impact of New Public Fund Regulations**: The new regulations aim to address the issue of active funds underperforming, potentially leading to a shift towards passive investment strategies and reducing market volatility. The main benchmarks are the CSI 300 and the CSI 800 indices [1][3][4]. - **Market Sentiment and Activity**: Positive market sentiment is noted, with an increase in total trading volume across all A-shares, a rise in the proportion of strong stocks, and an increase in margin trading. The micro-cap stock index has reached new highs, indicating a favorable market environment for structural opportunities [1][5]. - **Investment Recommendations**: Three main investment directions are highlighted: 1. Self-controlled sectors within the Sci-Tech 50, particularly in military aircraft supply chains and the Harmony operating system [2][7]. 2. Traditional core assets represented by the Shanghai Stock Exchange 50, benefiting from the return to performance benchmarks, including finance, insurance, and public utilities [2][3]. 3. New consumption sectors, focusing on platform economies, elderly care, and maternal and infant products [2]. - **Sectoral Fund Flows**: Following the new regulations, sectors such as banking, non-bank financials, and coal are expected to see significant short-term inflows, while high-tech companies are projected to gradually increase their market capitalization share [3][11]. Other Important but Possibly Overlooked Content - **Growth Themes**: Key growth themes include solid-state batteries, AI, and low-altitude economy, with upcoming technology exhibitions expected to catalyze market interest [8][9][10]. - **Long-term Market Trends**: The long-term outlook suggests a gradual increase in the market capitalization of high-tech companies, particularly in the TMT (Technology, Media, and Telecommunications) sector, which is worth exploring for growth potential [12]. - **Current Market Dynamics**: The current market environment is characterized by a recovery in risk appetite, with the U.S. delaying interest rate cuts and ongoing trade negotiations between China and the U.S. contributing to a stabilizing global equity market [5][6]. This summary encapsulates the key insights from the conference call, focusing on the implications of regulatory changes, market sentiment, investment strategies, and growth opportunities within the public fund industry and related sectors.

新华财经北京5月9日电四大证券报内容精华摘要如下： 中国证券报 ·降息落地引导实体经济综合融资成本下行 据中国人民银行网站消息，5月8日，人民银行以固定利率、数量招标方式开展了1586亿元逆回购操作， 下调政策利率0.1个百分点，即公开市场7天期逆回购操作利率从目前的1.5%调降至1.4%。因当日无逆回 购到期，因此实现净投放1586亿元。同日，根据《中国人民银行关于下调个人住房公积金贷款利率的通 知》，下调个人住房公积金贷款利率0.25个百分点。专家表示，此次降息，不仅有利于提振市场情绪， 还有利于保持商业银行净息差稳定，同时通过利率传导，有效降低实体经济综合融资成本，巩固经济基 本面。下调个人住房公积金贷款利率，将更好满足住房消费需求，支持房地产市场持续健康发展。 ·一季报数据显示 4084家公司实现盈利回升向好态势巩固 上海证券报 ·国新办新闻发布会解读民营经济促进法新法如何护航民营企业发展多部门划重点 5月20日，《中华人民共和国民营经济促进法》将正式实施。在国新办5月8日举行的新闻发布会上，多 部门相关负责人全面介绍制定出台民营经济促进法的背景、意义，以及在推动该法落地实施、助力民营 企业持续健康高 ...

2025年一季报业绩进一步印证了算力产业链的高景气程度，特别是国产算力核心企业，净利润普遍大幅 增长，这反映了算力市场的旺盛需求。在业界人士看来，国产算力正处于政策利好持续释放、市场需求 爆发、技术创新突破、业绩开始兑现四重共振的关键节点，其价值实现的奇点正加速到来。 政策红利持续释放 政策作为产业发展的"方向盘"和"助推器"，在推动国产算力崛起的过程中，发挥着不可或缺的关键作 用。 2020年4月，中共中央、国务院印发《关于构建更加完善的要素市场化配置体制机制的意见》，"数 据"作为一种新型的生产要素首次被列入。2020年底，国家发展改革委等部门发布《关于加快构建全国 一体化大数据中心协同创新体系的指导意见》，提出到2025年，全国范围内数据中心形成布局合理、绿 色集约的基础设施一体化格局。 2021年5月，《全国一体化大数据中心协同创新体系算力枢纽实施方案》确定了在京津冀、长三角、粤 港澳大湾区、成渝，以及贵州、内蒙古、甘肃、宁夏等地布局建设全国一体化算力网络国家枢纽节点， 提出加快实施"东数西算"工程。2021年12月到2022年2月，"东数西算"涉及的8个国家算力枢纽节点和10 个国家数据中心集群全部 ...

Investment Rating - The report maintains a "Buy" rating for several key companies in the semiconductor and AI sectors, including 中芯国际 (SMIC), 海光信息 (Haiguang), and others, indicating strong growth potential in the domestic AI and computing landscape [5][6]. Core Insights - The domestic AI landscape is witnessing significant advancements with the emergence of models like 豆包 (Doubao) and DeepSeek, which are leading the charge in multi-modal and lightweight AI model development, respectively [1][2]. - The report highlights a shift towards domestic computing power solutions, with chip manufacturers rapidly adapting to the evolving AI ecosystem, particularly through advancements in semiconductor processes and AI training capabilities [2][3]. - There is a notable increase in capital expenditure among cloud computing firms, driven by the rising demand for AI computing infrastructure, which is expected to lead to a "volume and price rise" scenario in the cloud computing market [3][4]. Summary by Sections Section 1: Breakthroughs in Domestic AI Models - 豆包 has emerged as a leading multi-modal model, enhancing capabilities in speech, image, and code processing, with a significant release of its visual understanding model in December 2024 [1][11]. - DeepSeek focuses on lightweight model upgrades, achieving a remarkable cost-performance ratio with its DeepSeek-V3 model, which has 671 billion total parameters and costs only 557.6 million USD, positioning it among the world's top models [1][12]. - The rapid iteration of domestic models, including updates from 通义千问 and others, reflects a competitive landscape that is accelerating the development of AI applications [1][34]. Section 2: Advancements in Domestic Computing Power - 中芯国际 is advancing its semiconductor processes, with N+1 and N+2 technologies being developed to support the growing demand for AI chips, achieving significant performance improvements [2][56]. - The report notes that the domestic chip industry is evolving, with companies like 昇腾 (Ascend) and others making strides in AI training and inference capabilities, thereby reducing reliance on international competitors [2][59]. - The cloud computing sector is experiencing a capital expenditure boom, with companies like 华勤 and 浪潮 rapidly deploying servers that are compatible with domestic computing power solutions [3][4]. Section 3: Infrastructure and Supply Chain Developments - The report emphasizes the need for enhanced computing infrastructure to meet the surging demand for AI applications, with significant investments being made in server and power supply innovations [3][4]. - Innovations in power supply and cooling systems, particularly the shift from traditional air cooling to liquid cooling, are becoming essential to support the increasing power density in data centers [4]. - The report identifies key players in the supply chain, including companies in power supply, cooling, and server manufacturing, that are poised to benefit from the growth of the AI and computing sectors [5].

Core Viewpoint - Huawei has successfully trained a trillion-parameter model, marking a significant advancement in AI capabilities and reducing reliance on Nvidia's technology [1][4][74]. Group 1: Challenges in Training Large Models - Training trillion-parameter models faced several challenges, including load balancing difficulties, high communication overhead, and low training efficiency [3][10]. - The architecture optimization, dynamic load balancing, distributed communication bottlenecks, and hardware adaptation complexities were identified as the four main challenges [10]. Group 2: Huawei's Solutions - Huawei's Pangu team utilized over 6,000 Ascend NPUs to achieve stable training of a 718 billion parameter MoE model, implementing breakthrough system optimization techniques [4][5]. - The team developed a model simulation tool that accurately predicts performance, achieving over 85% accuracy in matching actual test data [17]. Group 3: Load Balancing and Efficiency - A new EP group load balancing loss algorithm was introduced, which balances task distribution without excessive constraints, thus saving communication costs [24][25]. - The training efficiency of the Pangu Ultra MoE model improved significantly, with a Model FLOPs Utilization (MFU) of 30.0%, a 58.7% increase compared to previous optimizations [33]. Group 4: Communication Optimization - The team designed a hierarchical EP communication strategy to reduce inter-node communication volume, enhancing overall training efficiency [42][44]. - An adaptive pipe overlap mechanism was implemented to mask communication delays, further improving performance [48]. Group 5: Model Performance and Benchmarking - The Pangu Ultra MoE model demonstrated competitive performance across various benchmarks, achieving high scores in general understanding and reasoning tasks [61][62]. - The model's architecture allows for significant specialization among experts, enhancing its overall expressiveness and performance [64][66]. Group 6: Future Implications - The advancements in Huawei's technology signify a shift in the global AI landscape, showcasing China's capabilities in leading AI innovations [74]. - The ongoing development and application of the Pangu Ultra MoE model are expected to drive intelligent transformation across various industries, contributing to China's technological leadership [74].

Group 1 - The core viewpoint of the articles highlights the positive performance of the data ETF and its underlying index, indicating a growing interest in the data industry and AI-related sectors [1][2] - As of May 8, the CSI Big Data Industry Index rose by 0.86%, with notable increases in constituent stocks such as *ST Dongtong (up 15.62%) and Donghua Software (up 9.96%) [1] - The data ETF (516000) experienced a significant increase, reaching a price of 0.96 yuan, with a 5-day cumulative rise of 6.12% as of May 7 [1] Group 2 - Market sentiment has improved in May, with a focus on technology growth, particularly in domestic computing power and AI Agent sectors, driven by investments from major companies like Google and Microsoft [2] - The emergence of the DeepSeek-R1 model is seen as a foundational development for domestic AI Agent creation, indicating a shift towards deeper thinking models in AI [2] - The data ETF and its associated funds closely track the CSI Big Data Industry Index, focusing on key areas such as data elements, data security, and the East Data West Computing initiative, with major stocks including iFlytek and Inspur Information [2]

Investment Rating - Investment rating: Outperform [1] - Target price: RMB 29.72 [1] Core Insights - The company reported a revenue of RMB 28.55 billion for 2024, a year-on-year decrease of 8.29%, while net profit attributable to shareholders was RMB 703 million, an increase of 39.05% [6][10] - The computing power business is experiencing rapid growth, with domestic revenue nearing RMB 6 billion and a compound annual growth rate (CAGR) exceeding 100% over the past two years [6][10] - The company has optimized its expense management, achieving a slight reduction in total expense ratio despite a decline in overall revenue [6][10] Financial Summary - Revenue forecast for 2025-2027: RMB 31.15 billion, RMB 32.97 billion, and RMB 34.57 billion respectively [4] - Net profit forecast for 2025-2027: RMB 1.01 billion, RMB 1.17 billion, and RMB 1.34 billion respectively [4][10] - Earnings per share (EPS) for 2025-2027: RMB 0.85, RMB 0.99, and RMB 1.13 respectively [4][10] Product Performance - Communication systems revenue: RMB 22.31 billion, down 6.19% year-on-year, with a gross profit margin of 19.46% [6][10] - Fiber and cable revenue: RMB 4.70 billion, down 12.77% year-on-year, with a gross profit margin of 25.35% [6][10] - Data network products revenue: RMB 1.14 billion, down 27.44% year-on-year, with a gross profit margin of 43.92% [6][10] Strategic Developments - The company has secured centralized procurement for 400G OTN, reinforcing its position as a core supplier [6][10] - Global expansion efforts include trial production at R&D and manufacturing bases in Thailand and Hungary, aimed at mitigating trade risks [6][10]

Investment Rating - Investment rating is upgraded to Outperform with a target price of RMB 29.72, reflecting a 12% upside from the current price of RMB 21.36 [1][5][9] Core Views - The company reported a revenue of RMB 28.55 billion for 2024, a year-on-year decrease of 8.29%, while net profit attributable to shareholders increased by 39.05% to RMB 703 million [5][9] - The computing power business is experiencing rapid growth, with domestic revenue nearing RMB 6 billion and a compound annual growth rate (CAGR) exceeding 100% over the past two years [5][9] - The company has secured a position as a core supplier for 400G OTN networks, enhancing its competitive edge in the optical communication sector [5][9] Financial Summary - Revenue projections for 2025-2027 are updated to RMB 31.15 billion, RMB 32.97 billion, and RMB 34.57 billion respectively, with corresponding net profits of RMB 1.01 billion, RMB 1.17 billion, and RMB 1.34 billion [3][5][9] - Earnings per share (EPS) estimates for 2025-2027 are RMB 0.85, RMB 0.99, and RMB 1.13 respectively [5][9] - The company’s gross profit margin for communication systems is 19.46%, for fiber and cable is 25.35%, and for data network products is 43.92% [5][9] Product Performance - Revenue from communication systems was RMB 22.31 billion, down 6.19% year-on-year; fiber and cable revenue was RMB 4.70 billion, down 12.77%; and data network products revenue was RMB 1.14 billion, down 27.44% [5][9] - The company has optimized its expense ratios, achieving a total expense ratio decrease of 0.11 percentage points despite a decline in overall revenue [5][9] Global Expansion - The company has initiated trial production at its R&D and manufacturing bases in Thailand and Hungary, which is expected to help mitigate trade risks [5][9]

1. Report Industry Investment Rating - No industry investment rating is provided in the report. 2. Core View of the Report - In the stage of strong macro - uncertainty, the low - volatility attribute of convertible bonds is prominent. In the short term, market pricing is expected to shift from macro - driven to industry - driven, and the combination of technology and dividend is still the best strategy [2][17]. 3. Summary by Related Catalogs 3.1 Market Focus Last Week (April 28 - April 30, 2025) - **Stock Market**: The equity market fluctuated narrowly before the holiday. The robot and AI - related industrial chains performed well, driving up sectors such as TMT, automobile, and machinery. Defensive sectors like banks and coal adjusted significantly. Most Shenwan primary industries closed down, with real estate (-3.04%), comprehensive (-2.75%), social services (-2.63%), and coal (-2.50%) leading the decline, while media (+2.69%), computer (+2.47%), beauty care (+2.46%), machinery and equipment (+1.59%), and electronics (+1.34%) performed well [8][9]. - **Bond Market**: The bond market rose significantly last week. With the central bank's support across months, the capital market was loose. Coupled with the expectation of weakening PMI and the lack of substantial progress in Sino - US tariff negotiations, bond market sentiment was positive. The 10 - year Treasury bond yield closed at 1.62% on Wednesday, down 3.63bp from the previous week [9]. - **Convertible Bond Market**: Approximately half of the convertible bond issues closed higher. The CSI Convertible Bond Index decreased by 0.07% for the whole week, the median price increased by 0.12%, the arithmetic average parity increased by 1.21%, and the overall market conversion premium rate decreased by 0.84% compared with the previous week. The arithmetic average conversion premium rates of convertible bonds in the par value ranges of [90,100), [100,110), and [110,120) changed by -0.09%, -0.04%, and -0.51% respectively, at the 36%, 48%, and 33% quantiles since 2021 [1][9]. 3.2 Views and Strategies (May 6 - May 9, 2025) - **Macro Environment and Market Trends**: Although the US initiated a tariff war in April, the probability of a marginal easing of tariff conflicts is higher in the short term. After the disclosure of annual and first - quarter reports, the impact of performance on the market has cleared. In May, the proportion of macro factors in stock market pricing may decrease, and the market is expected to return to being driven by industry logic. The "new economy" sector may be more cost - effective [2][17]. - **Bond Market**: The monetary policy was hawkish in the first quarter. Considering the long - term trade - war risk and the expectation of reserve requirement ratio cuts and interest rate cuts, the risk of a significant bond market adjustment is low. The low bond market interest rate means the opportunity cost of convertible bonds is extremely low, and the potential allocation power of convertible bonds is abundant [2][17]. - **Convertible Bond Market**: The average conversion premium rates in multiple par value ranges have returned to the levels in mid - January, significantly lower than during the spring market. The median price has returned to February 5, and the equal - weighted index of underlying stocks is around February 10. The low - volatility attribute of convertible bonds is prominent, which can reduce portfolio volatility while maintaining exposure to equity assets [2][17]. - **Investment Directions**: - **Dividend and Defensive Assets**: When the market tumbled on April 7, Hong Kong stock dividends were a direction for capital to increase positions. After the adjustment of the conversion price of some high - dividend convertible bonds, the par value increased by more than 5%. The power industry may see stock price increases from May to June [17]. - **Technology**: It includes opportunities in humanoid robots, semiconductor equipment localization, innovative drugs and related industrial chains, and the intelligent driving market [17]. 3.3 Valuation Overview - As of April 30, 2025, for equity - biased convertible bonds, the average conversion premium rates in different par value ranges are at different quantiles since 2010 and 2021. For debt - biased convertible bonds, the average YTM of those with a par value below 70 yuan is at the 31%/36% quantiles since 2010/2021. The average implied volatility of all convertible bonds is at the 52%/27% quantiles since 2010/2021 [18]. 3.4 Primary Market Tracking - **Last Week (April 28 - April 30, 2025)**: No convertible bond issuance was announced, and Weice Convertible Bond was listed. Weice Technology is a well - known third - party integrated circuit testing service enterprise. The issued convertible bond scale is 1.175 billion yuan, with a credit rating of AA [27]. - **Future Week (May 6 - May 9, 2025)**: No convertible bond issuance or listing is announced. As of April 30, there are 80 convertible bonds to be issued, with a total scale of 125.58 billion yuan, including 7 approved for registration with a total scale of 13.45 billion yuan and 6 approved by the listing committee with a total scale of 3.82 billion yuan [28].