麦肯锡报告预测，2028年全球AI推理市场规模将达1500亿美元，2025-2028年复合年增长率超40%。智 算作为承接推理爆发的核心基础设施，将迎来前所未有的市场增量空间。金山云星流平台此次通过提升 平台效率、突破行业边界、加速推理布局的三维进阶，为迎接AI应用爆发做好了充分准备。申万宏源 称，AI 拉动与生态协同效应将在未来持续释放。 智通财经APP获悉，周一，金山云(KC.US)盘前涨6%，报13.08美元。消息面上，金山云高级副总裁刘涛 在1月21日举办的金山云年度Tech Talk上表示，智算平台金山云星流已完成从资源管理平台向一站式AI 训推全流程平台的战略升级。从训推平台、机器人平台到模型API服务，升级后的金山云星流平台构建 了从异构资源调度、训练任务故障自愈到机器人行业应用支撑、模型API服务商业化落地的全链路闭 环。 ...

广发证券在研报中表示，金山云AI业务维持高增长。预计2026年金山云将延续高资本开支，超过100亿 元，主要考虑需求依然旺盛。预测金山云2026年收入同比增长28.8%达122亿元，其中AI公有云收入同 比增长85%达53.7亿元，占总收入44%，预测行业云稳健增长；预计毛利率维持，费用率收窄，经调整 OP转盈利，经调整EBITDA率进一步提升。 广发证券认为，金山云拥有高可见度的强劲需求，小米金山加大AI投入，基础大模型、智能汽车、 WPS均需要大量AI训练和推理算力，非小米客户AI需求亦旺盛。 上证报中国证券网讯（记者 李五强）1月26日，金山云盘中一度涨逾9%，截至10时30分，该股报6.84港 元/股，涨5.88%。 金山云高级副总裁刘涛在1月21日举办的金山云年度Tech Talk上表示，智算平台金山云星流已完成从资 源管理平台向一站式AI训推全流程平台的战略升级。从训推平台、机器人平台到模型API服务，升级后 的金山云星流平台构建了从异构资源调度、训练任务故障自愈到机器人行业应用支撑、模型API服务商 业化落地的全链路闭环。 来源：上海证券报·中国证券网 ...

Group 1 - The core viewpoint of the news is that Kingsoft Cloud has made significant advancements in the AI computing sector, with its platform, Kingsoft Cloud Star Stream, transitioning to a comprehensive AI training and inference platform, indicating a strong demand for AI applications [1] - Kingsoft Cloud's stock price increased by 6.5% in Hong Kong, which positively influenced its US stock price, rising by 5.5% in pre-market trading, marking a third consecutive day of gains [1] - According to Guangfa Securities' latest research report, the company is expected to experience strong demand driven by increased AI investments from Xiaomi and other clients, with significant capital expenditures anticipated in Q4 2025 and throughout 2026 [1] Group 2 - The stock performance details include a closing price of 12.340, a pre-market price of 13.020, and a trading volume of 1.0918 million shares [2] - The total market capitalization of Kingsoft Cloud is reported at 3.728 billion, with a total share count of 302 million [2] - The stock has a 52-week high of 22.260 and a low of 10.290, indicating significant volatility in its trading history [2]

每经AI快讯，金山云(03896.HK)早盘涨超9%，截至发稿，涨9.13%，报7.05港元，成交额5.09亿港元。 每经AI快讯，金山云(03896.HK)早盘涨超9%，截至发稿，涨9.13%，报7.05港元，成交额5.09亿港元。 ...

金山云(3896.HK)大涨超9%，最高触及7.05港元。 本文源自：金融界AI电报 ...

刘涛披露，公司智算云业务Q3账单收入达7.8亿元，同比增速近120%，主要受益于行业内具身智能场景 及AI推理场景爆发。其中，小米金山生态的贡献显著，2025年上半年相关收入占年度关联交易总额的 40%。除了小米金山生态这个基本盘，金山云的外部客户用量也在齐增。 1月26日，金山云早盘大涨超9%，最高触及7.05港元。消息面上，在上周举办的金山云年度Tech Talk 上，金山云高级副总裁刘涛宣布，智算平台金山云星流已完成从资源管理平台向一站式AI训推全流程 平台的战略升级。据悉，金山云星流训推平台提供从模型开发、训练到推理的完整生命周期管理，具备 开发、训练、推理和数据处理四大模块能力，通过降低多模块协同复杂度，能实现"开箱即用"的AI开发 体验。 ...

麦肯锡报告预测，2028年全球AI推理市场规模将达1500亿美元，2025-2028年复合年增长率超40%。智 算作为承接推理爆发的核心基础设施，将迎来前所未有的市场增量空间。金山云星流平台此次通过提升 平台效率、突破行业边界、加速推理布局的三维进阶，为迎接AI应用爆发做好了充分准备。申万宏源 指出，AI 拉动与生态协同效应将在未来持续释放。 fund （原标题：港股异动 | 金山云(03896)早盘涨超9% 星流平台完成战略升级 已进化为一站式AI训推全流程 平台） 智通财经APP获悉，金山云(03896)早盘涨超9%，截至发稿，涨9.13%，报7.05港元，成交额5.09亿港 元。 消息面上，金山云高级副总裁刘涛在1月21日举办的金山云年度Tech Talk上表示，智算平台金山云星流 已完成从资源管理平台向一站式AI训推全流程平台的战略升级。从训推平台、机器人平台到模型API服 务，升级后的金山云星流平台构建了从异构资源调度、训练任务故障自愈到机器人行业应用支撑、模型 API服务商业化落地的全链路闭环。 ...

刘涛披露，公司智算云业务Q3账单收入达7.8亿元，同比增速近120%，主要受益于行业内具身智能场景 及AI推理场景爆发。其中，小米金山生态的贡献显著，2025年上半年相关收入占年度关联交易总额的 40%。除了小米金山生态这个基本盘，金山云的外部客户用量也在齐增。(格隆汇) 金山云(3896.HK)大涨超9%，最高触及7.05港元。消息面上，在上周举办的金山云年度Tech Talk上，金 山云高级副总裁刘涛宣布，智算平台金山云星流已完成从资源管理平台向一站式AI训推全流程平台的 战略升级。据悉，金山云星流训推平台提供从模型开发、训练到推理的完整生命周期管理，具备开发、 训练、推理和数据处理四大模块能力，通过降低多模块协同复杂度，能实现"开箱即用"的AI开发体验。 ...

[Table_Page] 跟踪研究|软件与服务 证券研究报告 [Table_Title] 【广发计算机&海外】金山云（03896.HK） 前瞻：预计25Q4 AI云高增长，26年延续高资本开支 [Table_Summary] 核心观点： 盈利预测：备注：报告中汇率为 1 港元=0.893 人民币；货币若无特别说明均为人民币。 | [Table_ 单位:Finance] 人民币百万元 | 2023A | 2024A | 2025E | 2026E | 2027E | | --- | --- | --- | --- | --- | --- | | 营业收入（百万元） | 7,047 | 7,785 | 9,481 | 12,215 | 14,687 | | 增长率 （ % ） | -13.8% | 10.5% | 21.8% | 28.8% | 20.2% | | 经调整EBITDA（百万元） | -265 | 639 | 2,292 | 4,072 | 5,657 | | 经调整净利润（百万元） | -1,291 | -825 | -567 | -519 | -387 | | EPS（元/股） | -0.61 ...

Investment Rating - The report maintains a "Recommended" rating for the industry [4] Core Insights - The demand for Tokens is driving inflation across the AI industry chain, with price increases being observed from storage to CPUs and now extending to cloud services, marking a significant shift in the pricing dynamics of the cloud computing sector [14][30] - AWS has initiated a price increase for its EC2 machine learning capacity blocks by approximately 15%, breaking a long-standing trend of declining prices in the cloud services industry, which may lead to a revaluation of cloud computing and related service providers [14] - The report highlights potential investment opportunities in cloud computing, CPUs, and databases, suggesting companies such as Alibaba, Kingsoft Cloud, and Inspur for cloud computing; Haiguang Information and Longxin for CPUs; and companies like StarRing Technology and DM Database for databases [30] Summary by Sections 1. Cloud Computing Price Trends - The report indicates that the AI industry chain is experiencing a price transmission trend, with cloud computing being the next area to see price increases following storage and CPU price hikes [14] - AWS's price adjustment reflects anticipated supply-demand dynamics, suggesting that successful price increases could lead to further adjustments in the future [14] 2. CPU Market Dynamics - The report notes a persistent supply-demand imbalance in the CPU market, particularly with Intel facing production limitations that may lead to continued shortages into 2026 [16] - The importance of CPUs is expected to rise significantly due to the increasing demand from AI applications, with the report emphasizing the need for optimization across both CPU and GPU resources [24][25] 3. Database Sector Growth - The report anticipates a rapid increase in the number of database PCU nodes driven by the demand for AI agents, which could lead to significant revenue growth for database providers [26] - The emergence of AI-native databases is highlighted as a key trend, with companies like Alibaba's PolarDB gaining traction in the market [29] 4. Investment Recommendations - The report suggests focusing on companies in the cloud computing sector such as Alibaba, Kingsoft Cloud, and UCloud, as well as CPU manufacturers like Haiguang Information and Longxin, and database firms like StarRing Technology and DM Database [30]