BlockBeats news, September 26 — U.S. stocks remained resilient under pressure from high oil prices and high yields, with AI themes continuing to support the technology sector's performance. Goldman Sachs' latest estimates push the market's focus further toward the return side: leading U.S. AI cloud providers need to achieve approximately $300 billion in annualized AI revenue over the next few years to cover current large-scale investments; for cloud providers to achieve substantial returns while the application layer also maintains high profit margins, end users would need to spend nearly $1 trillion annually on AI applications.
Goldman Sachs analyst Ryan Hammond expects hyperscaler capital expenditure to reach approximately $800 billion in 2026, with the market consensus for 2027 at approximately $1.1 trillion. Goldman Sachs' baseline judgment is that actual investment in 2027 may still exceed consensus, though the growth rate and magnitude of upside surprises will gradually moderate. For AI trades already reliant on data centers, GPUs, storage, and networking equipment, capital expenditure is not near a point of sudden contraction, and the market will more closely watch how much cloud revenue and profit each dollar of investment can generate.
The report also does not simply turn the AI narrative pessimistic. Goldman Sachs noted that second-quarter hyperscaler cloud business revenue has already annualized approximately $70 billion above the pre-AI trend, with announced revenue backlog exceeding $1.5 trillion. Enterprise AI procurement remains at an early stage, and the recent acceleration in enterprise spending will make AI's impact on corporate profits clearer over the coming quarters.
