Goldman Sachs projects AI capital expenditures will reach $1.2 trillion by 2027, reflecting accelerating hyperscaler investment in compute infrastructure.
Goldman Sachs has significantly raised its forecast for AI infrastructure spending by America's largest technology companies. The bank now projects US hyperscaler capital expenditures could reach as high as $1.4 trillion by 2027—surpassing its previous estimate of roughly $1.1 trillion and well above the broader Wall Street consensus of approximately $920 billion.
According to Goldman's latest trajectory, hyperscaler AI capex is estimated at roughly $405 billion in 2025, rising to approximately $750 billion in 2026, before reaching the $1.2 trillion range in 2027, with the latest revision pushing that 2027 figure even higher. The investment wave is being driven by the largest tech firms: Microsoft, Amazon, Alphabet, Meta, and Oracle, alongside notable investment in OpenAI. Goldman characterizes these companies as transitioning from an experimental phase of AI deployment into full-scale commercial implementation.
Looking further ahead, Goldman estimates cumulative AI infrastructure spending could reach approximately $7.6 trillion from 2026 through 2031. The breakdown points to roughly $5.1 trillion for compute, $2.1 trillion for data centers, and $358 billion for power infrastructure. Goldman's analysts have drawn comparisons to historical technology waves, particularly the buildout of railroads and automobiles, both of which required massive upfront capital deployment before downstream economic benefits emerged.
More than one-third of the projected 2027 capital expenditure—roughly $400 billion—is expected to come through investment-grade bond issuance, marking a significant shift from the self-funded, cash-flow-driven investment model that characterized big tech's first two decades. Goldman points to advertising and subscription models as the primary monetization channels for AI-driven tools, betting that consumer AI agents and enterprise AI applications will generate revenue streams large enough to justify the infrastructure investment.
Goldman notes that median AI infrastructure stock valuations sit at around 26 times forward price-to-earnings ratios. The bank also flags several supply-side constraints that could slow the buildout regardless of capital availability. Power availability tops the list, with land accessibility presenting a similar bottleneck. Memory chip affordability rounds out Goldman's concerns, as AI workloads are extraordinarily memory-intensive. The shift toward debt financing also creates a secondary market dynamic worth monitoring, with over $400 billion in new investment-grade bond supply requiring absorption by fixed-income markets.