行 半导体 2026 年 03 月 01 日 业 研 究 半导体 关注 AI 模型迭代对端云飞轮的加速作用 ——大科技海外周报第 6 期 投资要点： 关注 AI 模型迭代对端云飞轮的加速作用。 行 业 定 期 报 告 我们认为年初以来国产 AI 大模型营销加速了端云飞轮的转动。1） 云端算力方面：我们认为算力需求=用户规模×调用频率×单次复杂 度，年初以来的大模型营销显著提升了用户规模和调用频率，用户增 多带动数据回流增加→模型参数扩张→推理模型规模提升，形成：模 型升级→端侧用户数提升→数据回流→模型再升级的飞轮，端侧 AI 的 用户数提升将持续提升云端算力需求。2）端侧 AI 方面，当前市场上 AI 眼镜/AI 玩具/具身智能机器人等端侧 AI 产品层出不穷，公众对一个 足够聪明、会聊天、能办事的 AI Agent 是非常期待的，而目前这个市 场需求仍未被很好满足，我们认为端云飞轮的加速转动、大模型的持 续迭代，会让 AI Agent 变得越来越好用，端侧 AI 作为 Agent 硬件载体 也将迎来新的市场机遇。 千问 AI 眼镜即将发布，看好 AI 眼镜产业链投资机会。 建议关注： 证 券 研 端侧 A ...

计算机 2026 年 3 月研究观点：中国模型领 跑全球调用量，软件吞噬情绪缓和 [Table_Invest] 评级： 增持 | [姓名table_Authors] | 电话 | 邮箱 | 登记编号 | | --- | --- | --- | --- | | 杨林(分析师) | 021-23183969 | yanglin2@gtht.com | S0880525040027 | | 吕浦源(分析师) | 021-23183822 | lvpuyuan@gtht.com | S0880525050002 | | 钟明翰(研究助理) | 021-38031383 | zhongminghan@gtht.com | S0880124070047 | 本报告导读： 2 月，中国模型调用量首超美国站至第一；模型应用层面，Grok 与谷歌多模态模型 加速迭代，Claude Cowork 插件体系扩容，企业级 Agent 向职能场景纵深落地，软 件吞噬情绪缓和。 投资要点： 请务必阅读正文之后的免责条款部分 风险提示：技术发展不及预期，公司业务拓展不及预期。 [Table_Report] 相关报告 计算机《计算机周观点第 ...

策略研究 | 证券研究报告 — 总量周报 2026 年 3 月 1 日 策略周报 中东地缘冲突升级，节奏与影响 A 股短期或受地缘政治波动及避险情绪干扰，但外部冲击幅度有限，关注资 源品及 AI 国产算力方向。 中银国际证券股份有限公司 具备证券投资咨询业务资格 策略研究 证券分析师：王君 (8610)66229061 jun.wang@bocichina.com 证券投资咨询业务证书编号：S1300519060003 证券分析师：徐沛东 (8621)20328702 peidong.xu@bocichina.com 证券投资咨询业务证书编号：S1300518020001 证券分析师：郭晓希 | 观点回顾 4 | | --- | | 市场热点聚焦 5 | | 中观行业与景气 8 | | 一周市场总览、组合表现及热点追踪 11 | | 风险提示 13 | (8610)66229019 xiaoxi.guo@bocichina.com 证券投资咨询业务证书编号：S1300521110001 证券分析师：徐亚 （8621）20328506 ya.xu@bocichina.com 证券投资咨询业务证书编号：S1300 ...

计算机 2026 年 3 月研究观点：中国模型领 跑全球调用量，软件吞噬情绪缓和 [Table_Invest] 评级： 增持 | [姓名table_Authors] | 电话 | 邮箱 | 登记编号 | | --- | --- | --- | --- | | 杨林(分析师) | 021-23183969 | yanglin2@gtht.com | S0880525040027 | | 吕浦源(分析师) | 021-23183822 | lvpuyuan@gtht.com | S0880525050002 | | 钟明翰(研究助理) | 021-38031383 | zhongminghan@gtht.com | S0880124070047 | 本报告导读： 2 月，中国模型调用量首超美国站至第一；模型应用层面，Grok 与谷歌多模态模型 加速迭代，Claude Cowork 插件体系扩容，企业级 Agent 向职能场景纵深落地，软 件吞噬情绪缓和。 投资要点： 请务必阅读正文之后的免责条款部分 风险提示：技术发展不及预期，公司业务拓展不及预期。 [Table_Report] 相关报告 计算机《计算机周观点第 ...

持续关注 AI Infra 及商业航天 计算机周观察 20260301 TMT 及中小盘/计算机 本篇报告梳理了过去一周科技相关热点，建议持续关注 AI Infra 及商业航天。 OpenAI 完成新一轮融资估值达到 7300 亿美元；英伟达发布 FY26Q4 业绩，营 收与利润创历史新高，数据中心贡献超九成；SpaceX 正加快推进史上最大规 模 IPO 计划，预计估值将达到 1.75 万亿美元；中国 Token 调用量首超美国， 算力云服务是最确定方向。 推荐（维持） 行业规模 | | | 占比% | | --- | --- | --- | | 股票家数（只） | 285 | 5.5 | | 总市值（十亿元） | 4529.7 | 3.9 | | 流通市值（十亿元） | 4022.3 | 3.8 | 证券研究报告 | 行业定期报告 2026 年 03 月 01 日 行业指数 % 1m 6m 12m 绝对表现 -2.1 2.7 11.0 相对表现 -2.2 -4.7 -7.7 资料来源：公司数据、招商证券 -30 -20 -10 0 10 20 30 40 Mar/25 Jun/25 Oct/25 Jan/ ...

Investment Rating - The industry rating is "Outperform the Market" [6][20]. Core Insights - The report emphasizes the acceleration of the end-cloud flywheel driven by the iteration of AI models, highlighting that the marketing of domestic AI large models has significantly increased user scale and call frequency since the beginning of the year [2]. - The demand for cloud computing power is driven by user scale, call frequency, and complexity of tasks, leading to a feedback loop that enhances model upgrades and increases cloud computing demand [2]. - The market for end-side AI products, such as AI glasses and intelligent robots, is rapidly evolving, with significant unmet demand for capable AI agents, suggesting new market opportunities [2][3]. - The upcoming release of the Qianwen AI glasses and other AI products is expected to drive growth in the AI glasses industry, with global shipments projected to exceed 23.687 million units by 2026 [3]. - The NVIDIA GTC conference is anticipated to showcase advancements in AI technology, with a focus on inference computing, indicating a growing demand in the computing power supply chain [4]. Summary by Sections Cloud Computing Power - The report outlines that the demand for cloud computing power is a function of user scale, call frequency, and task complexity, which has been positively impacted by the marketing of AI large models [2]. End-Side AI Products - The report notes the emergence of various end-side AI products and the public's expectation for intelligent AI agents, indicating a significant market opportunity that remains largely unmet [2]. AI Glasses Market - The report highlights the upcoming launch of Qianwen AI glasses and predicts a significant growth trajectory for the smart glasses market, with expected shipments in China to surpass 4.915 million units by 2026 [3]. Computing Power Supply Chain - The report mentions the upcoming NVIDIA GTC conference, which is expected to present new developments in AI technology and computing power solutions, reinforcing the positive outlook for the computing power supply chain [4].

Investment Rating - The report recommends a "Buy" rating for several companies within the technology sector, indicating a positive outlook for their stock performance over the next 12 months [31]. Core Insights - The report highlights key companies in the technology sector, including Jiangsu Technology (Electronics, Communication, Computer, Media) as top picks for March 2026 [5][9]. - The report emphasizes the growth potential in various sub-sectors, particularly in 3D printing, liquid cooling solutions, AI chips, and gaming [15][16][18][19][21]. Summary by Category Electronics - **Precision Electronics**: Expected to achieve a net profit of 0.8e-0.9e in 2025, a significant increase of 181.97%-192.21% year-on-year, driven by semiconductor orders and improved operational efficiency [15]. - **Han's Laser**: Anticipated explosive growth in 3D printing, with a potential market share increase due to demand from North American clients [15]. Communication - **Invech**: Projected net profits of 5.87 million, 10.47 million, and 15.35 million from 2025 to 2027, with a strong focus on liquid cooling solutions and overseas market expansion [16]. - **Rui Ming Technology**: Expected to recover from previous losses, with projected net profits of 3.81 million, 4.95 million, and 6.67 million from 2025 to 2027, driven by increased demand in commercial vehicle information systems [17]. Computer - **Haiguang Information**: Positioned as a leader in high-end CPUs and DCUs, benefiting from the growth of domestic AI computing power and collaboration with major tech firms [18]. - **Cambricon**: Anticipated to maintain strong growth in AI chip demand, supported by ongoing R&D and market trends [18]. Media - **Kaiying Network**: Expected to launch several new games in 2026, which are projected to drive revenue growth [19]. - **Perfect World**: Anticipated to release a new RPG game, "Yihuan," with high market expectations, contributing to future revenue growth [21].

证券研究报告 | 2026年2月29日 AI周报： AI快速发展，AI HALO资产价值彰显 行业研究 · 行业专题 计算机&通信 投资评级：优于大市（维持评级） 证券分析师：熊莉 021-61761067 xiongli1@guosen.com.cn S0980519030002 证券分析师：艾宪 0755-22941051 aixian@guosen.com.cn S0980524090001 请务必阅读正文之后的免责声明及其项下所有内容 AI HALO：AI领域的重资产、低淘汰率 图1：美国基建股指数价格波动（25年1月-26年2月） 资料来源：Wind，国信证券经济研究所整理 n AI HALO：AI领域的重资产、低淘汰率。 Ø 定义：HALO = Heavy Assets（重资产）+ Low Obsolescence（低淘汰率）。其中，重资产是指其商业模式建立在庞大的实物资本基础之上，具有很高 复制壁垒（例如成本、监管、建设时间、工程复杂性、网络整合难度等）的资产；低淘汰率是指资产的经济相关性能够穿越技术周期而持久存在（即 商业模式、资产是刚需，不会因AI革命而被取代）。 Ø 代表资产：HALO代表 ...

证券研究报告 | 2026年2月29日 AI周报： AI快速发展，AI HALO资产价值彰显 行业研究 · 行业专题 计算机&通信 投资评级：优于大市（维持评级） 证券分析师：熊莉 021-61761067 xiongli1@guosen.com.cn S0980519030002 证券分析师：艾宪 0755-22941051 aixian@guosen.com.cn S0980524090001 请务必阅读正文之后的免责声明及其项下所有内容 AI HALO：AI领域的重资产、低淘汰率 图1：美国基建股指数价格波动（25年1月-26年2月） 资料来源：Wind，国信证券经济研究所整理 资料来源：Wind，国信证券经济研究所整理 请务必阅读正文之后的免责声明及其项下所有内容 图2：美国SaaS指数价格波动（25年1月-26年2月） n AI HALO：AI领域的重资产、低淘汰率。 Ø 定义：HALO = Heavy Assets（重资产）+ Low Obsolescence（低淘汰率）。其中，重资产是指其商业模式建立在庞大的实物资本基础之上，具有很高 复制壁垒（例如成本、监管、建设时间、工程复杂性、网络整合难度 ...

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]