作为一种不依赖电网的解决方案，现场发电（Onsite Power Generation）正成为破解AI供电难题的关键。黄仁勋上周表示，行业需要建设独立于电 网的发电设施，以快速满足需求并保护消费者免受电价上涨的影响： 资产管理巨头Brookfield与燃料电池制造商Bloom Energy达成一项合作，旨在通过部署现场发电技术，为AI数据中心提供新的能源解决方案。 根据周一发布的公告，Brookfield Asset Management将投资至多50亿美元，用于部署Bloom Energy的燃料电池技术。这是Brookfield为支持大型AI 数据中心电力及计算基础设施战略的首笔投资。受此消息提振，Bloom Energy股价在盘前交易中一度飙升超过30%。 当下，AI行业庞大的数据中心计划正与老化的美国电网发生冲突，后者往往难以快速部署新增电力容量。Bloom Energy的燃料电池提供了一种可 快速部署的现场发电方案，因为它不依赖于现有电网，从而有效绕开了传统电力基础设施的限制。 两家公司表示，他们正在全球范围内合作设计和交付所谓的"AI工厂"，并将在今年年底前宣布一个位于欧洲的具体项目地点。此举标志着 ...

证券日报网讯立昂技术（300603）10月13日在互动平台回答投资者提问时表示，公司以西南地区的立昂 技术AI算力园区、广州南沙算力中心为核心，围绕国家八大算力枢纽节点进行网络布局，构建"算力 +算法+应用"的智算创新应用生态体系，通过地域系统和资源融合复用，实现算力资源的高效分配，打 造立昂云数据全国调度中心，助力全国一体化算力网络建设。未来将结合自身在IDC领域较为明显的业 务优势，在京津冀、杭州湾、大湾区、成渝、武汉及新疆等城市开展数据中心业务。公司密切关注新疆 国际融合算力中心项目的建设与发展，并具备参与此类项目的技术与服务能力，可提供算力中心建设、 系统支撑及运营服务等相关支持。截至目前，公司暂未就该项目一期建设签订正式合作协议，后续将积 极把握项目相关合作机会。若有实质性进展，公司将按规定及时履行信息披露义务。 ...

瑞银认为，目前估值已具备吸引力，加之政策预期升温，市场资金有望重演越跌越买的趋势。 瑞银中国策略主管王宗豪在报告中表示，鉴于该指数自4月份关税战升级后的低点已反弹约36%，短期 内或出现部分获利回吐。当前市场走势与4月份情形高度相似，当时指数在74点附近吸引了大量资金回 流。 瑞银最新发布的中国股市策略报告指出，若MSCI中国指数进一步下跌至74点水平，市场或将获得显著 支撑，投资者可能再度逢低入市。目前，MSCI中国指数报约85点。 报告同时指出，在4月份的抛售中表现较弱、此后出现明显反弹的板块，本轮或将面临更大抛压，包括 数据中心、互联网、科技硬件、汽车及零部件以及生物科技等行业。 ...

Core Viewpoint - The stock price of *ST Yushun (002289.SZ) experienced significant volatility, influenced by the progress of its major asset restructuring, which is seen as a critical move for the company's survival [2][3]. Group 1: Stock Performance - *ST Yushun's stock price fluctuated dramatically, hitting a limit down early in the trading session but later rebounding to a limit up, closing at 34.7 CNY per share with a gain of 3.09% [2]. - The stock has seen an extraordinary increase of over 1000% within the year, with a current year-to-date gain of 845.51% despite recent adjustments [5]. Group 2: Asset Restructuring - On October 10, *ST Yushun completed the first phase of its major asset restructuring, paying 335 million CNY, which marks a significant advancement in the restructuring process [2]. - The restructuring involves acquiring 100% equity of three companies—Zhong'en Cloud Technology, Shenhui Biyuan, and Zhong'en Cloud Information—operating a data center project with approximately 8,000 cabinets [2][3]. - The total cash payment for the acquisition is set at 3.35 billion CNY, funded through loans from the controlling shareholder, the company's own funds, or self-raised funds [2]. Group 3: Financial Performance - *ST Yushun has reported continuous losses over five years, with a cumulative loss exceeding 120 million CNY, and faces delisting risk due to its financial performance [3]. - In contrast, the target companies are projected to generate revenues of 735 million CNY, 815 million CNY, and 218 million CNY for the years 2023, 2024, and the first quarter of 2025, respectively, with net profits significantly surpassing *ST Yushun's current performance [3]. - If the restructuring is successful, *ST Yushun's annual profit could potentially turn from a loss to a profit exceeding 150 million CNY, aiding in its financial recovery [3]. Group 4: Regulatory Attention - The unusual trading activity of *ST Yushun has attracted the attention of regulatory authorities, with the Shenzhen Stock Exchange monitoring the stock closely due to abnormal trading behaviors [5].

从数据港十大流通股东角度 来源：新浪基金∞工作室 10月13日，数据港跌5.33%，截至发稿，报33.02元/股，成交1.91亿元，换手率0.80%，总市值237.21亿 元。数据港股价已经连续6天下跌，区间累计跌幅13.56%。 资料显示，上海数据港股份有限公司位于上海市静安区江场路1401弄14号1601室，成立日期2009年11月 18日，上市日期2017年2月8日，公司主营业务涉及数据中心服务器托管服务及网络带宽服务。主营业务 收入构成为：IDC服务业99.31%，IDC解决方案0.69%。 南方中证1000ETF（512100）基金经理为崔蕾。 截至发稿，崔蕾累计任职时间6年342天，现任基金资产总规模949.76亿元，任职期间最佳基金回报 177.68%， 任职期间最差基金回报-15.93%。 风险提示：市场有风险，投资需谨慎。本文为AI大模型自动发布，任何在本文出现的信息（包括但不 限于个股、评论、预测、图表、指标、理论、任何形式的表述等）均只作为参考，不构成个人投资建 议。 数据显示，南方基金旗下1只基金位居数据港十大流通股东。南方中证1000ETF（512100）二季度增持 130.24万股 ...

每经AI快讯，有投资者在投资者互动平台提问：2025年6月中标新疆国际融合算力中心一期项目，负责 建设2000P算力集群，重点支持卫星遥感数据处理和跨境物流智能调度。请问公司是否参与新疆国际融 合算力中心项目？ （记者 胡玲） 立昂技术（300603.SZ）10月13日在投资者互动平台表示，公司以西南地区的立昂技术AI算力园区、广 州南沙算力中心为核心，围绕国家八大算力枢纽节点进行网络布局，构建"算力+算法+应用"的智算创 新应用生态体系，通过地域系统和资源融合复用，实现算力资源的高效分配，打造立昂云数据全国调度 中心，助力全国一体化算力网络建设。未来将结合自身在IDC领域较为明显的业务优势，在京津冀、杭 州湾、大湾区、成渝、武汉及新疆等城市开展数据中心业务。公司密切关注新疆国际融合算力中心项目 的建设与发展，并具备参与此类项目的技术与服务能力，可提供算力中心建设、系统支撑及运营服务等 相关支持。截至目前，公司暂未就该项目一期建设签订正式合作协议，后续将积极把握项目相关合作机 会。若有实质性进展，公司将按规定及时履行信息披露义务。 ...

2024年，全国数据中心用电量高达1660亿千瓦时，几乎相当于两个三峡水电站的发电量。"在当今时 代，高算力需求与高能源消耗紧密相连，能源供应与碳排放之间的矛盾愈发尖锐，如同天平的两端，亟 待找到平衡的支点。"科华数据（002335）相关负责人接受证券时报记者采访时表示。 高算力需求与高能耗之间的矛盾，正成为人工智能时代的核心挑战之一，而科华创新提出的"智算+光 储"融合创新理念，正在为行业提供一种新的平衡路径。从一家电力电子技术企业，成长为在"智算中 心""智慧电能""清洁能源"三大赛道融合布局的解决方案提供商，科华数据正在算力爆发与能源转型的 双重浪潮下打开新的成长空间。 "算电协同"：重新定义能源与算力关系 据介绍，针对AI业务场景的高密度算力POD液冷微模块，科华数据实现了模块化、高弹性、解耦型的 灵活交付模式，基于高可靠，高安全的液冷产品设计理念与"AI+智能化"运维的加持下，实现了绿色节 能。 在智慧电能领域，科华数据持续拓展应用场景。上述负责人举例称，自2008年科华数据进入民航领域以 来，经过十多年的深耕沉淀，已成为民航数智化、绿色化建设的领军企业，被评为"2013—2024中国机 场建设优秀 ...

Core Viewpoint - The article emphasizes the urgent need to balance high computing power demands with energy consumption in the context of the AI era, highlighting the innovative "Intelligent Computing + Light Storage" integration proposed by Kehua Data as a potential solution to this challenge [1][3]. Group 1: Energy Consumption and Data Centers - By 2024, the total electricity consumption of data centers in China is projected to reach 166 billion kilowatt-hours, equivalent to the output of two Three Gorges Hydropower Stations [1]. - The report predicts that by 2030, China's data centers will consume approximately 525.76 billion kilowatt-hours, accounting for 4.8% of the total electricity consumption in society, indicating a more than twofold increase in six years [3]. - Kehua Data proposes a "Green Electricity + AI + Light Charging and Storage Integration" concept to provide clean energy support for high-energy computing needs and achieve refined energy management through AI technology [3]. Group 2: Technological Innovations - Kehua Data is focusing on liquid cooling technology to meet the high-density and high-efficiency requirements of AI computing, collaborating with leading domestic GPU manufacturers to develop high-density liquid cooling computing PODs [3][4]. - The company has achieved a modular, flexible delivery model for its high-density liquid cooling micro-modules, emphasizing green energy efficiency and intelligent operation [4]. Group 3: Clean Energy Solutions - In the clean energy sector, Kehua Data has developed a comprehensive grid-connected energy storage solution, responding to market demands and achieving breakthroughs across technology, products, and applications [5]. - The company has become the first in the industry to complete full-function verification of its grid-connected energy storage PCS, establishing a foundation for standardization and large-scale application [5]. Group 4: Global Market Positioning - Chinese companies are transitioning from "domestic substitutes" to "global standard definers" in fields like intelligent computing and clean energy, with Kehua Data leading this charge [7]. - The company has implemented an overseas development strategy to enhance its brand influence, successfully delivering data center products in regions such as Southeast Asia, Central Asia, Europe, and North America [7]. - Kehua Data has integrated data centers with "light storage" solutions, creating a unified "source-network-load-storage" system [7]. Group 5: Competitive Edge - The company emphasizes the importance of continuous innovation in technology and products to maintain long-term competitive advantages, focusing on excellence and ecosystem leadership rather than merely filling gaps in the market [8].

Investment Rating - The report suggests a focus on the nuclear power sector as a significant energy type for AI consumption, highlighting investment opportunities in companies like Entergy, Talen Energy, and Constellation Energy [5] Core Insights - The report indicates that the global infrastructure and construction equipment sector is seeing strong demand for data centers, with AMD supplying chips to OpenAI for AI infrastructure [1] - The International Energy Agency (IEA) has revised down its 2030 renewable energy capacity growth forecast for the US by 50% compared to last year's estimates, while globally, renewable energy generation capacity is expected to double by 2030 [1] - The report emphasizes the ongoing energy transition, with a balanced supply-demand scenario in the natural gas market, suggesting investment in companies like WMB and KMI [5] Summary by Sections Global Infrastructure and Construction Equipment - The vacancy rate for data centers in major North American markets has reached a historic low of 1.6%, indicating strong demand [8] - The average price for cabinets in data centers has increased by 2.5% for 250 to 500 kW cabinets and by 19% for those over 10 MW due to high demand and limited power supply [8] Global Electrical and Intelligent Equipment - The gas turbine price index increased by 3.43% year-on-year as of August 2025, while the electrical and special transformer production price index remained stable [17][31] - The report notes a significant increase in transformer exports from China, with a year-on-year growth of 18% in August 2025 [40] Global Energy Industry - The average spot price for electricity in major US regions decreased by 19.57% week-on-week, while natural gas futures prices rose by 3.7% [3] - The report highlights that the US electricity demand growth forecast has been revised upward, with an expected increase of 15.8% by 2029 [24] Global New Materials - The global spot price for uranium reached $82.63 per pound in September 2025, reflecting a 10% increase month-on-month [4] - The report tracks the dynamics of heavy rare earths, with prices for dysprosium and terbium remaining stable [4] Key Company Updates - GE Vernova has launched a new platform for utilities to monitor and control their infrastructure, supported by Verizon [45] - Hitachi has announced a strategic partnership with OpenAI to supply power distribution equipment for AI applications [45] - ABB has signed an agreement to provide automation solutions for the production and storage of green ammonia [45]

PUE是用于衡量整个数据中心用电量和IT设备用电量之间的比值。这个比值越接近1，就说明越节能、 省电。 中国信息通信研究院联合内蒙古和林格尔新区管委会共同编制的《中国绿色算力发展研究报告（2025 年）》显示，2024年，我国数据中心机架超900万标准机架，平均PUE降至1.46。 绿色算电协同发展 中西部地区绿色数据中心发展迅猛，入选"国家绿色数据中心"的数量占比从第一批的26.6%增长到第五 批的46%。作为全国首个开展绿色算电协同发展试点的省份，青海凭借得天独厚的绿能优势，积极融入 全国一体化算力网，在全国率先发布5项绿色算力地方标准。青海绿算产业如雨后春笋般拔节生长，全 面起势。 在全球数字经济发展与"双碳"目标驱动背景下，推动算力绿色发展已成行业共识和时代命题。国家发展 改革委等有关部门2024年发布的《数据中心绿色低碳发展专项行动计划》提出，到2030年底，全国数据 中心平均电能利用效率（PUE）、单位算力能效和碳效达到国际先进水平。 "最省电"的算力如何打造？近日，科技日报记者跟随"数字中国筑路行"主题采访活动来到青海，走访中 国电信（国家）数字青海绿色大数据中心等平台寻找答案。依托"东数西算" ...