Goldman Sachs, PwC Project Massive AI Infrastructure Spending Wave

Goldman Sachs Asset Management argues that agentic AI systems will drive over 90% of future digital infrastructure demand, creating investment opportunities in power, cooling, connectivity, and compute. Goldman Sachs projects US data center power demand to nearly double from 31 gigawatts in 2025 to 66 gigawatts by 2027. The IEA projects global data center electricity consumption to roughly double from 485 TWh in 2025 to 950 TWh by 2030, with AI-focused consumption tripling. PwC estimates global data center spending will total $31.6 trillion through 2050, rising to nearly $50 trillion if AI adoption accelerates, and finds that every dollar spent on data center construction commits approximately $12 more in future ICT equipment costs. US data center construction spending surged nearly 60% year-on-year in July per Census Bureau data. One analysis puts the AI infrastructure buildout at potentially $11 trillion through 2029, though that figure is contested. Goldman Sachs identifies the largest investment opportunity as infrastructure providers, chip makers, networking firms, utilities, and data center operators, rather than model creators. Firms named as relevant include Nvidia, Broadcom, Arista Networks, TSMC, Micron, SK Hynix, Samsung, Vertiv, Eaton, Constellation Energy, Vistra, Dominion Energy, NextEra Energy, CoreWeave, Nebius, Equinix, and IREN.