Goldman Sachs(US:GS) Hua Er Jie Jian Wen·2025-11-14 03:54Core Insights - The report from Goldman Sachs highlights the critical role of electricity supply in the AI race, indicating that disparities in power availability may hinder the expansion of AI infrastructure in the U.S. while providing an unexpected competitive advantage to China [1][5]. Group 1: U.S. Electricity Supply Challenges - The U.S. is facing an imminent electricity supply crisis, with effective reserve power capacity projected to fall below the critical 15% threshold by 2030 due to rapid growth in data center demand [1][2]. - Currently, data centers account for approximately 6% of total U.S. electricity demand, expected to rise to 11% by 2030, putting additional strain on the grid [2]. - The effective reserve power capacity in the U.S. has decreased from 26% to 19% over the past five years, with eight out of thirteen regional grids already at or below the critical level [2][4]. Group 2: China's Electricity Supply Advantage - In contrast, China's electricity supply is robust, with an expected effective reserve power capacity of around 400 GW by 2030, significantly exceeding its own demand and the projected global data center electricity demand of approximately 120 GW [1][3]. - China's substantial investments in renewable energy, coal, and nuclear power have led to a significant increase in electricity capacity, providing a solid foundation for future AI-driven data center expansion [3][4]. Group 3: Implications for the AI Race - The report emphasizes that sufficient and reliable electricity supply is a key factor in the AI competition, with the potential for U.S. electricity infrastructure bottlenecks to slow down AI development until at least 2030 [5]. - Conversely, China's ample and cost-effective electricity supply may attract data center investments, potentially altering the global distribution of data centers and impacting related industries such as energy equipment and high-energy-consuming sectors [5].