全球进入新瓦特时代 电力设备与新能源 马晓明 Xiaoming.ma@nomuraoi-sec.com SAC执证编号：S1720522060001 李沛雨 Peiyu.li@nomuraoi-sec.com SAC执证编号：S1720525040001 本PPT内容来自于外发报告《全球进入新瓦特时代-电力设备行业2026年年度策略报告》 请务必阅读报告正文后各项声明 一、电力需求进入超级周期 资料来源：Grid Strategies，野村东方国际证券 2 欧洲：电力需求有望重回加速增长 1 美国：AI数据中心建设推动电力需求高增长 图1：美国十年的电力增长 图2： 2030年峰值需求预测不断上调 图3：不同场景驱动负荷增长 3.7% 2.8% 0.9% 0.6% 0.6% 0.7% 2.4% 3.0% 4.7% 7.4% 8.8% 2025-2030E 2024 2023 2022 2010 2000 1990 1980 1970 1960 1950 24 38 64 120 166 0 20 40 60 80 100 120 140 160 180 GW 2022预测 2023预测 2024预测 2025 ...

野村- 全球进入新瓦特时代 20251229 摘要 美国用电量增速自 2020 年起显著回升，预计 2025-2030 年 CAGR 达 3.7%-4%，主要驱动力为制造业回流和数据中心建设，数据中心用电需 求占比高达 90 吉瓦。 欧洲用电趋势滞后美国 2-3 年，未来 5 年预计增速达 2.4%，主要由工 业化、电气化和数据中心发展推动，预计到 2030 年数据中心将带来 35 吉瓦的需求。 中东地区过去 20 年用电需求增长三倍，受城镇化、制冷需求和淡水供 应驱动，天然气发电占比超 70%，新能源渗透率达 11%，数据中心建 设将进一步推高能源需求。 中国是全球电力需求增长最快的国家之一，2024 年全社会用电量约 9,100 亿度，工业用电占比 64%，2025 年前 10 个月同比增长约 10%，制造业和出口是主要驱动力。 美国数据中心主要集中在弗吉尼亚、德州和加州，但接入难度增加，项 目容量增长受限，居民反对、电网接入点紧张和矿工转型带来供需矛盾。 德州电网支撑能力较弱，但用电需求增长迅速，未来电价上涨可能性高， 华尔街对德州缺电趋势判断乐观，能源公司已积极布局。 碳酸锂市场预计 2026 年迎来 ...

▲图表1： 2024年全球电力需求同增4% 资料来源：Ember，野村东方国际证券 ▲图表2： 2024年全球电力需求分布 资料来源：Ember，野村东方国际证券 1）美国：据Grid Strategies数据，2024年美国电力负荷增速2.8%（2022-2023年增速低于1%）， Grid Strategies预测2025-2030年美国电力负荷增速将达到3.7%，并指引未来5年（2025-2030年） 美国电力夏季峰值负荷需求增量将达到166GW，较2025年4月预测数据120GW进一步上调，增 量主要来自数据中心90GW（占比55%）。美国已经正式进入新一轮的瓦特时代。 2）欧洲：我们认为在新能源车渗透率提升、新能源新增装机量有望持续增长的背景下，欧洲未 来电力需求前景向好。据麦肯锡预测，2023-2030年欧洲多国（德国、法国、英国等）电力需求 复合增速有望重回增长态势，带动欧洲电力需求增量达到294-461TWh。主要驱动行业包括数据 中心（增量需求70-91TWh）、绿氢（增量需求79-101TWh）、以及交通（增量需求70-128TW h）。 3）中东：我们认为在清洁能源转型和经济持续发展的背 ...

Core Viewpoint - The article discusses the dual narratives of the current era, emphasizing the rise of AI and the critical importance of energy as a fundamental resource for sustaining AI technologies. It argues that the future of AI is not just about computational power but also about energy supply, leading to a new paradigm termed the "New Watt Era" [5][6][8]. Group 1: The Shift in Understanding Energy and AI - The concept of "computational power" is evolving from an asset to a liability, as the energy demands of AI systems grow exponentially [7][12]. - Data centers are transforming from mere storage facilities into "digital furnaces" that consume vast amounts of electricity to produce intelligence [11][12]. - The energy required for AI operations is becoming a critical factor for entry into the market, shifting from a cost to a capital requirement [15][19]. Group 2: The New Landscape of Energy Acquisition - The new competition among tech giants is focused on securing energy sources rather than proximity to users, leading to a "new land grab" for energy hubs [16][17]. - Power purchase agreements are evolving into "energy treaties," reflecting the strategic importance of securing energy supply for AI operations [18]. - The fear of "computational stranding" is emerging, where expensive AI infrastructure becomes worthless without access to energy [19][21]. Group 3: The Role of the Power Grid - The power grid is being redefined as the "sovereign artery" of AI civilization, essential for transporting energy to support AI systems [22][23]. - The aging infrastructure of the power grid poses a significant risk, as it struggles to meet the increasing demands of AI technologies [23][24]. - The mismatch in manufacturing cycles between GPUs and power transformers creates a bottleneck in energy supply for AI operations [26][27]. Group 4: The New Power Hierarchy - A new power hierarchy is forming, with energy and grid controllers at the top, followed by computational power providers and AI application developers [36][40]. - The dynamics of power are shifting, as AI sovereignty must now seek permission from energy authorities, highlighting the interdependence of energy and AI [40][41]. Group 5: Future Energy Strategies of Tech Giants - Major tech companies are investing heavily in green energy sources, not solely for sustainability but to secure political and energy quotas [44][46]. - The pursuit of nuclear fusion and small modular reactors (SMRs) indicates a strategic move towards energy independence, allowing companies to bypass traditional energy constraints [46][48]. - The vision of a self-sufficient AI ecosystem powered by private energy sources reflects a significant shift in how tech giants view their operational frameworks [49][50]. Group 6: Reevaluation of Physical Resources - The article calls for a reevaluation of the physical limitations that govern AI development, emphasizing that energy is becoming the most critical resource [55][56]. - The illusion of unlimited information is being challenged by the reality of finite physical resources, necessitating a new understanding of value in the AI landscape [52][53][58].

Group 1 - The A-share market has returned to 4000 points, with a rotation of market hotspots, particularly in the new energy sector, including energy storage, hydrogen energy, electricity, and charging piles [1] - Several stocks in the electric power and grid sector have seen significant gains, with companies like Hailu Heavy Industry and Moen Electric achieving consecutive trading limits [1] - The global AI wave is highlighting the urgent issue of power shortages, as tech leaders warn that electricity is becoming a critical constraint for AI development [2][3] Group 2 - Major investments in AI infrastructure are being made by tech giants in North America, driven by the ongoing electricity shortage [5] - Collaborations such as the $80 billion partnership between the U.S. and Westinghouse for nuclear reactors, and Google's construction of a 50 MW nuclear power plant, indicate a strong push towards enhancing power supply for data centers [6] - The aging U.S. power grid infrastructure requires significant upgrades, with 30% of power lines and transformers needing replacement in the next decade [7] Group 3 - Goldman Sachs reports that electricity supply is a major obstacle to AI development, predicting a 50-fold increase in power consumption by AI servers by 2027 compared to five years ago [8] - The International Energy Agency forecasts that global annual investment in power grids will rise to $500 billion by 2030, with a compound annual growth rate of 12.6% [8] - UBS anticipates a 10-year power supercycle in China, with electricity demand growth expected to double from previous estimates, driven by new energy vehicles and data centers [8] Group 4 - Domestic companies are expected to benefit from increased orders and export demands, with significant growth in transformer and switch exports reported [9] - CITIC Securities highlights a structural demand in the power transmission and transformation equipment sector, indicating a potential reversal in market conditions for high-voltage and smart grid segments [10]

Core Viewpoint - The article highlights a significant surge in the electric equipment sector, driven by the increasing demand for electricity to support AI infrastructure, with several stocks experiencing substantial gains [2][3]. Group 1: Electric Equipment Sector Performance - Electric equipment stocks, including companies like Shuangjie Electric and TBEA, saw collective surges, with many stocks containing the character "电" (electric) leading the gains [2]. - Specific stock performance includes N Xingbei with a 172.60% increase, and other notable mentions like Canchuang and Yichuang Electric with increases of 29.96% and 20.79% respectively [3]. Group 2: AI and Electricity Demand - Microsoft CEO Nadella and OpenAI CEO Altman indicated that the AI industry faces a power shortage rather than an excess of computing power, emphasizing the need for infrastructure close to power sources [4]. - The U.S. Department of Energy warned that if power suppliers cannot increase generation during peak demand, power outages could double in the next five years [4]. Group 3: Future Electricity Consumption Trends - The EIA predicts that U.S. electricity consumption will reach historical highs in 2025 and 2026, primarily driven by AI and data center expansions [4]. - Goldman Sachs forecasts that AI data centers will increase global electricity demand by 175% by 2030 compared to 2023 [4]. Group 4: Impact on Electricity Pricing - The imbalance between electricity supply and demand may lead to higher electricity prices, with reports indicating that U.S. power suppliers are seeking significant rate increases, totaling $29 billion in rate hikes approved for 2025 [4][5]. Group 5: Equipment Replacement and New Technologies - The U.S. is entering a "New Watt Era," with a significant portion of electrical grid equipment needing replacement, including 30% of power lines and transformers over the next decade [5]. - Companies like Eaton and Hitachi are experiencing growth in orders, particularly in the electrification sector, with Eaton reporting a 55% year-on-year increase in data center orders [5]. Group 6: Emerging Energy Solutions - New energy technologies such as fuel cells, solar storage, and nuclear power are expected to fill the electricity gap for data centers, with SOFCs projected to have an annual installation capacity of 0.5-1.25 GW in North America from 2026 to 2030 [6]. - Nuclear power is gaining attention from tech giants, with significant investments planned, including a $200 billion commitment in the U.S. for nuclear projects [6]. Group 7: Long-term Energy Strategy - Haitong Securities predicts that the AI narrative will accelerate U.S. electricity system construction, leading to a delay in coal power retirement and an increase in solar storage and SOFCs, with a long-term focus on nuclear power [7].