Goldman Sachs has raised its forecast for US hyperscaler AI capital spending to as much as $1.4 trillion by 2027, above its own prior estimate and Wall Street's consensus. More than a third of that spending is expected to come from debt rather than cash flow, and the bank flags power, land, and memory-chip supply as constraints on the buildout.
Goldman Sachs has raised its forecast for how much America's biggest tech companies will spend building out AI infrastructure, now projecting US hyperscaler capital expenditure could climb as high as $1.4 trillion by 2027. That tops the bank's own prior estimate of roughly $1.1 trillion and sits well above a broader Wall Street consensus that had been hovering closer to $920 billion.
A steeper spending curve
Goldman had previously modeled hyperscaler AI capex at roughly $405 billion in 2025, rising to about $750 billion in 2026, before hitting the $1.2 trillion range in 2027. The new revision pushes that final-year figure even higher.
The companies driving the wave are Microsoft, Amazon, Alphabet, Meta, and Oracle, with OpenAI also drawing notable investment activity. Goldman characterizes these firms as moving from an experimental phase of AI deployment into full-scale commercial implementation.
Debt takes on a bigger role
More than a third of the projected 2027 capex, roughly $400 billion, is expected to come through investment-grade bond issuance. That marks a shift from the self-funded, cash-flow-driven model that defined big tech's first two decades.
Goldman points to advertising and subscription models as the primary channels it expects to monetize AI tools, betting that consumer and enterprise AI applications will generate revenue large enough to justify the spending.
Valuations and supply constraints
Goldman puts median AI infrastructure stock valuations at around 26 times forward earnings. The bank's analysts flag power availability, land accessibility, and memory chip affordability as constraints that could slow the buildout regardless of how much hyperscalers are willing to spend, since AI workloads are extraordinarily memory-intensive.
Looking further out, Goldman estimates cumulative AI infrastructure spending could reach approximately $7.6 trillion from 2026 through 2031, split roughly between $5.1 trillion for compute, $2.1 trillion for data centers, and $358 billion for power infrastructure. Goldman's analysts compare the buildout to earlier waves of heavy infrastructure spending, such as railroads and automobiles, that required massive upfront capital before downstream economic benefits materialized.
Over $400 billion in new investment-grade bond supply will need to be absorbed by fixed-income markets.
Source: Crypto Briefing
Trading involves risk.
Risk Disclaimer
