独家洞察 | 当算力遇上电力：AI繁荣正在重塑美国能源版图

Core Insights - The article discusses the transition of the US AI industry from a "technological breakthrough phase" to a "deepening infrastructure phase" by 2026, emphasizing the shift in competitive focus towards acquiring computing power more efficiently and cost-effectively [2] Group 1: Energy Infrastructure Challenges - Major tech companies are investing heavily in AI, leading to the construction of energy-intensive data centers that require stable power supply, making energy a critical variable in the AI industry [4] - PJM Interconnection warned of a potential power supply gap of up to 60 gigawatts over the next decade due to rapid data center expansion, significantly exceeding the previously identified gap of 6.5 gigawatts [4] - The declining redundancy in the power grid and increasing risk of outages are structural challenges facing the US power system, affecting nearly one-fifth of the US population [4] Group 2: Political and Regulatory Landscape - The Trump administration supports data center investments to maintain US technological leadership, but rising electricity demand is increasing residential electricity prices, leading to political backlash [5] - A new emergency intervention plan was announced, requiring tech giants to fund new power plants to mitigate the impact of data center expansion on residential electricity prices, marking the first direct government intervention in regional power market pricing [5] Group 3: Industry Growth and Investment Opportunities - The construction of data centers is significantly increasing electricity demand in the US, with the EIA projecting a 2.4% increase in commercial electricity sales in 2025, and further growth of 2.4% and 4.3% in 2026 and 2027 respectively [6] - Companies in the gas turbine sector are experiencing a surge in orders, with GE Vernova reporting a 67% year-on-year increase in gas turbine orders to 10.2 gigawatts in Q4 2025 [6] - Major manufacturers are expanding production capacity to meet rising demand for natural gas power generation, with Caterpillar planning to double its gas turbine capacity and Siemens Energy investing $1 billion to expand local production [7] Group 4: Long-term Implications of AI on Energy Demand - AI is reshaping the energy demand structure in the US, indicating a long-term reconfiguration around "computing power—energy—infrastructure" rather than a short-term disruption [7]