在国家战略布局的宏大棋盘上，"东数西算""西电东送""西电西用"正共同勾勒出资源跨域调配的新格 局。一场以数据流重组能源流、以比特赋能瓦特的智慧革命席卷南北西东。 针对这些结构性矛盾，亟须构建系统性的解决方案，推动实现算力、电力、数据流三者的高效协同与价 值循环，将挑战转化为机遇，真正释放"东数西算"的战略潜能。 第一，推动"源网荷储算"一体化发展。在西部新能源富集区配套建设算力中心，把实时性要求低的模型 训练、图片渲染、数据备份等业务留在西部，实现新能源就地消纳。宁夏、内蒙古等地的实践表明，这 种模式能够将数据中心电能利用效率（PUE）降至1.1~1.2，显著提升能源利用效率。放眼全球，算力与 电力从分离到耦合正成为新赛道，谷歌把数据中心搬进丹麦海上风场，用风机直流母线直供GPU，省去 升压、降压、并网等三级损耗，度电成本下降18%。 第二，在国家层面建立"算电协同项目库"。把数据中心、风光配套、调节电源、外送通道打包核准，用 一张甘特图锁死工期，谁拖延谁买单。建立算力负荷参与电力现货市场的规则，特别是在八大算力枢纽 节点开展协同试点，探索分布式新能源参与绿电交易的具体路径。目前我国算力平台已接入山西、青 ...

重要事件：国务院发布《固体废物综合治理行动计划》。2026年1月4日，国务院印发《固体废物综合治 理行动计划》（以下简称"固废十条"）的通知。"固废十条"要求到2030年，重点领域固体废物专项整治 取得明显成效，固体废物历史堆存量得到有效管控，非法倾倒处置高发态势得到遏制，大宗固体废弃物 年综合利用量达到45亿吨，主要再生资源年循环利用量达到5.1亿吨，固体废物综合治理能力和水平显 著提升。 国信证券近日发布公用环保2026年1月投资策略：12月沪深300指数上涨2.28%，公用事业指数下跌 2.46%，环保指数下跌0.04%，月相对收益率分别为-4.74%和-2.32%。申万31个一级行业分类板块中，公 用事业及环保涨幅处于第27和第19名。电力板块子板块中，火电下跌5.17%；水电下跌3.08%，新能源 发电下跌0.39%；水务板块下跌2.55%；燃气板块下跌0.18%。 以下为研究报告摘要： 核心观点 市场回顾：12月沪深300指数上涨2.28%，公用事业指数下跌2.46%，环保指数下跌0.04%，月相对收益 率分别为-4.74%和-2.32%。申万31个一级行业分类板块中，公用事业及环保涨幅处于第27 ...

AIDC 时代，供配电面临的不是"更大功率"这么简单，而是三件事同时发生： 01 算力增长从线性变成跳跃，建设节奏被压缩； 02 在 AI 驱动的数据时代，供配电系统正处于重构的亮点线上。面对更高密度、更高功率、更高韧性的挑战，伊顿与行业伙伴一起，正在回答一个核心问 题：智算时代，我们该如何供电？ 2025 年 12 月，在中国 IDC 产业年度大典上，伊顿以"从电网到芯片，供配电系统的重构之路"为主题，联合来自云计算、运营商、解决方案厂商重量级， 共同探讨下一代数据中心供配电架构的演进趋势与技术路径。 【从现在到未来】 AC 到 DC 是可持续演进的过程 不是一蹴而就的突变 电力资源从"可获得"变成"受约束"，需要协同与互动； 03 容错空间被持续挤压，任何一次中断都更接近业务级事件。 在这种全新的边界条件下，伊顿的判断清晰而坚定：动力，是为 IT 算力服务的。没有动力，哪来的算力。 这意味着供配电系统不再只回答"能不能供上电"，更要精准回应"能不能跟得上算力节奏、匹配资源约束、把风险降维"三大核心命题。 适配当下 以可演进架构 跟上算力的动态变化节奏 服务算力的首要前提，是跟上其动态变化。如今 AI 负载 ...

今年3月，同力天启与天启鸿源一起同甘肃省庆阳市政府成功签署战略合作框架协议。根据协议内容， 三方将围绕储能装备制造基地和储能电站建设展开合作，一期项目包括建设储能装备生产线和独立储能 电站。项目建成后，储能电站将为庆阳"东数西算"产业园区的数据中心直供绿色电力，实现算电协同， 并在保证供电可靠性的前提下，实现部分备用电源的功用，提高算力中心绿电占比，进一步降低算力中 心能耗成本。 7月，同力天启与甘肃移动签订战略合作协议。甘肃移动作为中国移动全资子公司及国 家"东数西算"工程庆阳节点与集群的主要建设方，在庆阳"东数西算"产业园规划了140亩智算机房，具 备承载不低于16个万卡集群的强大能力。结合甘肃移动的属地资源、数字化能力与同力日升的项目全流 程经验，双方将重点开发"通信基础设施新能源化"项目。 12月26日，国家发展改革委、工业和信息化部、国家能源局发布《国家级零碳园区建设名单(第一批)》 的通知，共纳入北京经济技术开发区、甘肃庆阳东数西算产业园等52个园区，并在其中提到，各地区发 展改革委、工业和信息化主管部门、能源局要积极支持本地区国家级零碳园区建设，推动绿电直连、新 能源就近接入增量配电网等绿色电力 ...

当AI（人工智能）能够以毫秒级速度预测风电、光伏发电的波动，当智能算法将风电场弃风率压降至 3%以下，当虚拟电厂精准调度5万个分散充电桩与空调负荷——我们蓦然发现，能源世界的主导者，已 不再仅仅是传统资源与硬件，而是数据、算法与算力。 人工智能重构能源权力的全球格局 从2022年ChatGPT横空出世，到2024年DeepSeek等大模型迅猛崛起，人工智能的进化速度已突破人类想 象力边界——它的迭代周期从年缩短至月、周，参数规模每9个月翻10倍，认知能力在一些领域甚至超 越人类。 这场静默却汹涌的革命，正以摧枯拉朽之势重构全球能源秩序的底层逻辑。 传统能源秩序以"资源—设施—市场"为核心，谁占有油田、煤矿、风光、管线、电站，谁就拥有主导 权，而AI时代的秩序则以"数据—算法—算力"为核心，谁掌握高质量连续数据、谁能训练出低能耗高泛 化模型、谁能在毫秒级响应中完成最优决策，谁才是新世界的王者。 千亿级大模型应运而生，国家电网光明电力大模型、国家能源集团擎源大模型、中国石油昆仑大模型、 南方电网大瓦特大模型、三峡集团大禹大模型……这些"超级大脑"正在接管各类核心业务，大幅提升系 统工作效率，深刻改变能源行业运行模 ...

相较于传统数据中心，耗电更多、用电负荷波动更大的人工智能数据中心（AIDC）对于供能产业链提 出新要求，安全、经济、绿色成为能源产品迭代的主要方向。专家认为，绿色能源、节能降碳是解决 AIDC用能问题的两大路径，虚拟电厂、储能、燃气轮机等能源解决方案提供商迎来发展机遇，相关领 域已成为企业重点布局的赛道。 ● 本报记者 郑萃颖 企业积极布局 AIDC独特的能源需求正在重塑能源产业链发展格局。阳光慧碳科技有限公司总裁周文闻表示，相较于 传统数据中心，AIDC更专注于提供算力服务，其能源消耗呈倍数增长；与此同时，随着算力密度不断 提高，对能源设备的功率密度、散热方式提出了更高要求。 国际能源署今年4月发布的报告显示，到2030年，全球数据中心的电力需求将增长1倍以上，达到约945 太瓦时（TWh）；AI是这一增长的主要驱动力，预计到2030年，AIDC的电力需求将增长四倍以上。业 内人士认为，AIDC的高速发展，始终离不开电力供应安全、经济、绿色三大核心需求。周文闻表示， AI应用的普及正推动AIDC能源需求向多维度升级。"在此过程中，AIDC能源供应的经济性成为市场首 要考量目标。绿色低碳发展的大趋势，则利好风 ...

Core Insights - The rise of Artificial Intelligence Data Centers (AIDC) is creating new energy demands, emphasizing safety, economy, and sustainability in energy supply chains [1][3] - Green energy and carbon reduction are identified as two main pathways to address AIDC energy consumption issues, leading to opportunities for energy solution providers [1][4] Industry Developments - Companies are actively investing in virtual power plants, AIDC power support, and energy storage solutions to adapt to the new energy demands of AIDC [1][2] - China Energy Construction has made significant technological breakthroughs in virtual power plants, enhancing system flexibility and reducing operational costs [1] - Yangdian Technology plans to invest 50 million yuan to establish a subsidiary focused on comprehensive power solutions for data centers and AIDC [1] Energy Storage Innovations - Energy storage is crucial for the stable operation of AIDC, with companies like Haicheng Energy launching lithium-sodium collaborative storage solutions to increase green energy usage [2] - Jerry Holdings has made progress in the data center power generation sector, signing contracts with global AI giants for generator sales, indicating a move into the North American high-end power market [2] Energy Demand Projections - The International Energy Agency projects that global data center electricity demand will more than double by 2030, reaching approximately 945 terawatt-hours (TWh), with AIDC demand expected to quadruple [3] - AIDC's unique energy requirements are reshaping the energy industry, necessitating higher power density and cooling solutions [2][3] Policy and Regulatory Framework - The Chinese government is increasing guidance on green and low-carbon development for data centers, aiming for an average energy efficiency of 1.5 or lower by the end of 2025 [4] - New policies will require over 80% of green electricity in newly built national hub data centers by 2025, promoting the use of renewable energy [4] Technological Innovations - Companies are exploring direct current power supply solutions to minimize energy losses and transitioning from traditional diesel generators to diverse energy storage solutions [5] - Various energy storage technologies are being developed, including lead-carbon batteries and modular lithium and sodium batteries, focusing on safety and economic viability [6] Collaborative Energy Management - The concept of "compute-power synergy" is emerging, where data centers can optimize power dispatch and improve overall efficiency through AI-driven load scheduling [6][7] - Major internet companies are utilizing virtual power plant technologies to manage computing power across multiple data centers, optimizing based on electricity costs and carbon footprints [7] Future Outlook - The energy reform focus is shifting towards energy consumption, with expectations for the emergence of unicorn companies in virtual power plants and new energy systems during the 14th and 15th Five-Year Plans [8]

Core Insights - The article emphasizes the rapid transformation of the global technology landscape driven by artificial intelligence (AI), highlighting the evolution of data centers towards higher density, efficiency, and intelligent management [1] - The concept of "computing power infrastructure" is redefined in the context of AI, necessitating continuous adaptation and innovation to meet changing demands [3][4] Group 1: Understanding "More Knowledgeable" Infrastructure - The essence of being "more knowledgeable" is defined as continuous evolution, recognizing that static understanding quickly becomes outdated in the fast-paced AI era [3] - The company has accumulated extensive insights from its deep involvement in the transition from cloud computing to intelligent computing, allowing it to capture rapidly changing industry needs [3] - The company’s nearly 40 years of experience in power electronics and over a decade in data center lifecycle services contribute to its strong product and project implementation capabilities [3] Group 2: Challenges and Solutions in AI Era - AI's rapid iteration, particularly in GPU technology, has fundamentally altered data center design logic, shifting from standardization to meeting high demands for infrastructure [4] - The company is addressing cooling challenges by adopting new cooling technologies like liquid cooling, integrating them with power supply and control systems [4] - The scale of computing clusters is expanding dramatically, necessitating the introduction of AI management to enhance operational efficiency and energy savings [4][6] Group 3: Technological Innovations and Energy Efficiency - The company has launched a groundbreaking 200 kW power supply product, utilizing third-generation semiconductor devices, achieving an efficiency of up to 99.2% [5] - Enhancements in energy efficiency can save millions in electricity costs, especially in regions facing power shortages, allowing for more effective computing power utilization [6] - The trend towards prefabrication and modularization is driven by the need for rapid deployment, significantly reducing delivery times from one year to three to four months [6] Group 4: Green and Low-Carbon Development - The company emphasizes the necessity of green and low-carbon development in data centers, addressing the growing conflict between computing power and energy consumption [7] - Innovations in product design, such as using silicon carbide (SiC) materials, have led to power supply efficiencies exceeding 98%, reflecting the company's commitment to energy conservation [7] - The concept of "computing power and electricity synergy" is proposed, leveraging energy storage systems to provide flexible power support for computing centers [7] Group 5: Reliability and Global Expansion - The reliability of systems is critical as data center complexity increases, with proactive maintenance strategies being implemented through AI technology [8] - The integration of sensors in products allows for real-time monitoring of equipment health, facilitating predictive maintenance and enhancing system availability [8] - The company aims to expand globally, with prefabricated power modules designed for easy transport and rapid deployment, showcasing advancements in China's high-end manufacturing capabilities [8]

Core Insights - The integration of computing power and electricity, termed "算电融合" (computing-electricity integration), is emerging as an innovative model that supports the green energy-driven digital economy, particularly benefiting the artificial intelligence (AI) industry [1][2][3] - The "算电融合" initiative aims to create a new infrastructure that combines energy flow, data flow, and value flow, enhancing the synergy between green electricity and intelligent computing [1][6] - The recent Huawei summit in Heilongjiang focused on promoting this integration and fostering collaboration among various industries to drive high-quality digital economic development in the region [2][5] Group 1: Industry Development - Heilongjiang is positioning itself as a leader in the integration of computing power and electricity, leveraging its energy-rich resources to accelerate economic growth in the digital era [1][6] - The establishment of the AI-enabled digital transformation industry alliance aims to bridge the gap between technology development and industrial application, facilitating the adoption of AI across various sectors [4][6] - Huawei's commitment to innovation in computing and energy sectors is expected to enhance the stability and efficiency of green electricity supply for computing needs, supporting the development of a new green intelligent computing center in Heilongjiang [7][8] Group 2: Strategic Initiatives - The national strategies "东数西算" (East Data West Computing) and "算电协同" (Computing-Electricity Collaboration) present significant opportunities for Heilongjiang as a demonstration province for computing-electricity integration [7] - The summit emphasized the importance of aligning computing power projects with green electricity supply, advocating for a coordinated approach to optimize energy and computing resources [6][7] - Future collaborations and project implementations are expected to unlock the potential of computing-electricity integration and AI, driving regional industrial upgrades and contributing to the high-quality development of Northeast China [8]

务求实效 推动"人工智能+能源"走深走实 ——三论学习贯彻2026年全国能源工作会议精神 从实践基础看，顶层设计已然明晰，路径探索初见成效。国家发展改革委、国家能源局联合发布《关于 推进"人工智能+"能源高质量发展的实施意见》，系统部署了"人工智能+"电网、新能源、新业态等应用 场景，着力培育人工智能驱动能源转型变革新生态。能源企业及开展融合实践，中国石油、国家电网、 南方电网、中国华电、中煤集团等企业已发布能源领域大模型超过30个，在带动产业链协同、促进行业 智能化升级方面积累了有益经验。从井下矿山的智能采掘，到跨区域电网的智慧调度，再到虚拟电厂灵 活聚合分布式资源，融合应用已在多领域展开，为下一步全面推广奠定了良好基础。 准确把握"抢占战略制高点"的实践要求，需深入理解能源与人工智能"融合发展、协同联动"的内在逻 辑。 一方面，以人工智能之"智"，重塑能源之"形"。我国能源市场大、应用场景广、数据基础好，与人工智 能融合发展前景广阔。在危险作业环境替代方面，人工智能可助力智能化煤矿建设、保障核电安全高效 运行，推动"增安提效"；在辅助管理决策方面，可优化电网运行、提高新能源功率预测精度，为新型电 力系统建 ...