EB Securities: Data center construction reshaping the American electricity landscape, optimistic about the benefits of gas turbines, power equipment, and energy storage.

EB SECURITIES released a research report stating that the structural problem of power shortage in the United States is transforming into long-term and high certainty investment opportunities, with gas turbines, power equipment, and energy storage as the benefiting sectors. 1) Gas turbines: In the context of a highly prosperous market, there is a bottleneck in the production capacity of overseas gas turbine leading companies. It is expected that Chinese companies will continue to increase their market share. Recommended companies to watch include Dongfang Electric Corporation (01072, 600875.SH) and Shanghai Electric Group (02727, 601727.SH). 2) Power equipment: The growth in infrastructure demand in the US power grid is expected to benefit the transformer segment with supply constraints. Recommended companies to watch include Hainan Jinpan Smart Technology (688676.SH), Sieyuan Electric (002028.SZ), Eaglerise Electric & Electronic (002922.SZ), and others. Upgrading AI power architectures can effectively improve power efficiency. Continued volume growth is expected for HVDC and SST solutions. Companies to watch include Shenzhen Sinexcel Electric (300693.SZ), Beijing Sifang Automation (601126.SH), Shenzhen Megmeet Electrical (002851.SZ), and Shenzhen Hopewind Electric (603063.SH). 3) Energy storage: Energy storage can effectively enhance the reliability of power systems in the short term. Recommended companies to watch include Sungrow Power Supply (300274.SZ) and CSI Solar Co., Ltd. (688472.SH). EB SECURITIES' main views are as follows: What are the causes of power shortages in the United States? The research firm believes that the core reason for power shortages in the United States is the continuous increase in capital expenditure expectations for data centers. Based on this, Grid Strategies has raised its forecast for peak load growth in the US from 64GW in 2025 to 166GW in 2030. There is a mismatch between capital expenditure expectations for data centers and actual demand, as well as a mismatch between actual demand and infrastructure capacity, leading to uncertainty in the pace of data center deployments. Energy storage, SOFC, and other adjustable power sources can effectively fill the gap in load demand. The future addition of power generation in the US is expected to be primarily gas-fired, with EIA projecting an addition of gas-fired capacity of 7/7/16/8/7GW from 2026 to 2030, while other stable power sources will see minimal additions. Coal-fired power plants will also face significant retirement pressures. By mid-October 2025, planned data center backup capacity in the US has reached 245GW since January 2023. As data centers continue to come online, peak load demands will increase, making it difficult to meet load demands solely with stable power sources. The research firm has calculated the extent of power shortages in the US based on different scenarios of data center deployment. If adjustable power sources are not considered, the projected load gaps in 2030 would be 2.35.65.157GW for scenarios of data center construction paces of 5 years 90GW/10 years 245GW/8 years 245GW/5 years 245GW. With adjustable power sources considered, all four scenarios can meet load demands. Significant load growth is concentrated in regions with high data center construction density. The future growth in peak load demand in the US will be primarily concentrated in transmission areas such as ERCOT and PJM, driven by data centers. PJM and ERCOT are the primary regions for data center construction in the US. As of January 2026, planned data center capacities in Virginia for PJM and in Texas for ERCOT are 34.19GW and 26.60GW, respectively. PJM has a network advantage with low latency, while ERCOT has a power advantage with abundant natural gas resources. PJM: The surge in data center demand is projected to increase summer peak load from 156GW in 2026 to 222GW in 2036. Additionally, with a decrease in effective installed capacity, PJM's power reserve margin fell to 18.6% in 2025-2026, below the 20% safety threshold. This has led to a sharp increase in capacity prices from $28.92/MWh in 2024-2025 to $269.92/MWh in 2025-2026. PJM is adjusting policies to expedite power construction, prioritizing stable power source development. By June 2025, PJM has signed grid connection agreements for 46GW of power projects, with 11.8GW mainly gas-fired projects in the fast approval pipeline. ERCOT: With abundant natural gas resources, lower electricity prices, and faster grid connection approval, ERCOT has become the primary choice for new data center planning. As of November 8, 2025, ERCOT has received applications for 226GW of large load connections, including 164GW for data centers. This will drive its summer peak load from 87GW in 2025 to 138GW in 2030. ERCOT predicts a reliability shortage in its electricity system by 2028. To address this, ERCOT is accelerating the development of energy storage, gas-fired, and other sources that can effectively enhance the reliability of the power system. By the end of 2025, ERCOT has 441GW of power projects in the queue for grid interconnection, including 175GW of energy storage and 54GW of gas-fired capacity. Risk analysis: Risks include potential delays in AI data center construction, trade risks related to exporting to the US, and fluctuations in raw material prices.