Goldman Sachs Outlook Cites Accelerating Cloud Growth and Large Hyperscaler Order Backlogs
Редакция ЦОД:Главное
A Moomoo market review, citing a Goldman Sachs report dated October 2, says combined cloud-revenue growth at AWS, Microsoft Azure, Google Cloud and Oracle rose from 24% in the first quarter of 2025 to 48% in the second quarter of 2026. The review also cites nearly $1.7 trillion in cumulative cloud orders across AWS, Azure and Google Cloud.
- Combined cloud-revenue growth across AWS, Azure, Google Cloud and Oracle was reported at 48% in Q2 2026, versus 24% in Q1 2025.
- Goldman Sachs reportedly forecasts 55% growth in the third quarter and 56% in the fourth quarter, although the forecast year was not specified.
- AWS, Azure and Google Cloud were said to hold nearly $1.7 trillion in cumulative cloud orders.
- The cited AI-compute revenue benchmarks indicate that utilization, contract tenor and financing costs are material variables in hyperscaler infrastructure returns.
- No specific data-center project, capacity addition or customer contract was identified in the source material.
According to Moomoo’s review of a Goldman Sachs report, combined cloud-revenue growth for AWS, Microsoft Azure, Google Cloud and Oracle accelerated to 48% in the second quarter of 2026, from 24% in the first quarter of 2025.
Goldman Sachs reportedly expects the growth rate to reach 55% in the third quarter and 56% in the fourth quarter. The article does not specify the year for those quarterly forecasts. It also states that AWS, Azure and Google Cloud have accumulated nearly $1.7 trillion in cloud orders.
For data-center operators and infrastructure investors, the reported growth and order backlog suggest that demand for AI-oriented cloud capacity could remain strong enough to support continued hyperscaler expansion. This is an inference from the reported market indicators, rather than a confirmation of specific data-center projects or capacity commitments.
The review presents wide revenue benchmarks for AI-compute leasing. Short-term AI-compute rentals are cited at $40 billion-$50 billion per GW annually, compared with $20 billion-$30 billion per GW annually for long-term contracts. Goldman Sachs is also cited as estimating that a hyperscaler would require approximately $11.6 billion of annual AI revenue per GW to achieve a 15% return on invested capital.
Those benchmarks highlight the sensitivity of AI infrastructure economics to utilization, contract mix and financing costs. The article further suggests that lower funding costs could improve the economics of future data-center investment. Separately, it cites expectations from some institutions that the US 10-year Treasury yield could decline from about 5.2% to 4.7% by early 2027; this is a market expectation, not an achieved outcome.
The cited GW-based commercial and return benchmarks are industry-level estimates, not terms from a specific customer contract or data-center project. The source also does not clarify whether the GW reference is IT load or facility power.
The figures point to sustained demand for cloud and AI infrastructure that could support further hyperscaler data-center investment. However, the cited economics also underline that returns on capital-intensive AI capacity remain sensitive to utilization, contract duration and the cost of capital.
- — / 12 октября 2026 г. · КитайскийПервоисточник →
Материал подготовлен редакцией ЦОД: Главное на основе указанных первоисточников.