2025年释放AI时代的潜力报告（英文版）-高盛

Group 1 - The core viewpoint of the report is that artificial intelligence (AI) is driving a historic transformation in computing, requiring significant capital and infrastructure support while facing challenges such as surging power demand [1][2][3] - AI is fundamentally changing the computing paradigm, with AI data centers being more complex and resource-intensive than traditional cloud computing, projected to require 50 times more power by 2027 [1][2][29] - The demand for AI infrastructure is expected to lead to an investment of $1 trillion by hyperscale companies by 2027, with a significant focus on GPUs and data centers [1][31] Group 2 - Power demand is identified as a critical bottleneck for AI development, with global data center power demand expected to surge by 160% by 2030, necessitating new capacity to meet this growth [2][46] - The aging U.S. power grid infrastructure, with an average age of 40 years, is struggling to accommodate the increasing demand, leading companies to explore diverse power supply solutions [2][3] - Data centers are becoming strategic assets in geopolitical contexts, with countries leveraging them for influence, and companies like Amazon and Google expanding their presence in regions with favorable energy conditions [2][3] Group 3 - The capital requirements for AI infrastructure are immense, with projected needs of $2 trillion for digital infrastructure and $3 trillion for power by 2030 [3][46] - Innovative financing solutions are emerging, including joint ventures and asset securitization, to meet the capital demands of AI data centers [3][38] - The report emphasizes the necessity for cross-industry collaboration, policy support, and innovative capital tools to address the challenges posed by power and geopolitical factors [3][46]