Goldman Sachs said AI capital expenditure is still not close to a sudden contraction point, even as markets begin to examine returns more closely and ask how much cloud revenue and profit each dollar of spending can actually produce.
According to BlockBeats, the bank said on Sept. 26 that U.S. equities have remained resilient despite pressure from higher oil prices and elevated yields, while the AI theme continues to support the tech sector.
Goldman’s revenue math points to a high bar
In Goldman’s latest estimates, leading U.S. AI cloud providers would need to generate about $300 billion in annualized AI revenue over the next few years to cover the scale of current investment.
If cloud providers are to earn attractive returns and application-layer companies are to maintain relatively high profit margins, end users would need to spend close to $1 trillion a year on AI applications, the bank said.
Capex expectations remain large through 2027
Goldman Sachs analyst Ryan Hammond expects hyperscale cloud providers to reach about $800 billion in capital spending in 2026. Market consensus for 2027 is around $1.1 trillion.
Goldman’s base case is that actual 2027 investment could still come in above that consensus level, although both growth and the degree of outperformance are likely to ease over time.
Markets are watching monetization more closely
For AI trades that already rely on data centers, GPUs, storage and networking equipment, Goldman said capex is not yet near a sudden pullback point. What changes now is the focus: markets are likely to watch monetization efficiency more closely, especially how spending converts into cloud revenue and profit.
The report did not recast the AI narrative in bearish terms. Goldman said second-quarter hyperscale cloud revenue was already about $70 billion above the pre-AI trend on an annualized basis, while disclosed revenue backlog has surpassed $1.5 trillion.
The bank also said enterprise AI procurement is still in its early stages. With corporate spending accelerating recently, AI’s impact on company profits may become clearer over the next several quarters.