According to the latest research from Morgan Stanley, global cloud capital expenditure is projected to reach USD 735 billion by 2026, representing an approximate year-on-year growth of 60% and marking the third consecutive year of rapid expansion. This growth is primarily driven by the significant upward revision in capital expenditure guidance by the four major U.S. cloud service providers (Meta, Alphabet, Amazon, and Microsoft), reflecting a sustained excess of demand over supply for AI computing power. Analysts believe that cloud service providers are facing long-term supply constraints and must accelerate investments to align with growing demand, which will prolong a high-intensity capital expenditure cycle.
According to the latest research by Morgan Stanley, global cloud computing capital expenditure is expected to surge to approximately USD 735 billion in 2026, marking a year-on-year increase of about 60% and achieving a growth rate exceeding 60% for the third consecutive year. This forecast represents a significant upward revision of 22 percentage points compared to three weeks ago, with an additional scale of approximately USD 120 billion.
The core driver behind this upward revision is the robust capital expenditure guidance issued by the four major U.S. hyperscale cloud service providers during the latest earnings season. $Alphabet-A (GOOGL.US)$ 、 $Amazon (AMZN.US)$ 、 $Microsoft (MSFT.US)$ and $Meta Platforms (META.US)$ All have significantly increased their investment plans for 2026. Although the growth rate has slowed compared to 2025, it is unprecedented in industry tracking history for the world's top 11 cloud service providers to maintain such high-intensity capital expenditure growth for three consecutive years.
The research suggests that against the backdrop of AI-driven computing power demand consistently outpacing supply, the high-intensity investment cycle in cloud infrastructure will persist. This trend is explicitly beneficial for suppliers with significant capital expenditure exposure within the cloud computing supply chain.
Capital expenditure intensity reaches an all-time high
The capital expenditure intensity of the world's top 11 cloud service providers has climbed to a historical peak of over 26% of total revenue, tripling the average level from 2014-2023. Just one year ago, market forecasts for global cloud capital expenditure in 2025-2026 were approximately USD 710 billion, whereas current expectations have surged to USD 1.2 trillion, indicating a cumulative upward revision of nearly USD 500 billion.
Morgan Stanley’s forecast for global cloud capital expenditure in 2026 has reached USD 795 billion, which is approximately 8% higher than the current market consensus. Given that the capital expenditure guidance from major cloud service providers has been sequentially revised upward for at least nine consecutive quarters, analysts believe there is still room for further upward adjustments to current market expectations.
U.S. hyperscale cloud service providers significantly increase investments
In the latest earnings season, the four major U.S. hyperscale cloud service providers further clarified the accelerating trend of AI-driven capital expenditure. Meta set its 2026 capital expenditure guidance at USD 115-135 billion, reflecting a year-on-year increase of 73% at the midpoint, with funds primarily directed toward Meta Super Intelligence Labs and broader AI infrastructure needs.
Alphabet announced its 2026 capital expenditure guidance at USD 175-185 billion, marking a 97% year-on-year increase at the midpoint, with spending expected to rise quarter by quarter throughout the year. The expenditures will mainly support Google DeepMind, optimization of user-facing AI functionalities, and large-scale cloud infrastructure expansion. The company revealed that approximately 60% of spending will be allocated to server procurement, while the remaining 40% will go toward long-term assets like data centers and networks.
Amazon expects its 2026 capital expenditure to be around USD 200 billion, representing a 52% year-on-year increase, primarily directed toward AWS operations to address unexpectedly strong growth demands from both core non-AI businesses and AI cloud workloads. The company emphasized that its monetization pace is generally synchronized with the deployment progress of installed capacity.
Although Microsoft did not provide specific quantitative guidance for 2026, it noted that capital expenditures in the March quarter of this year will decline quarter-over-quarter due to fluctuations in the pace of cloud infrastructure construction and arrangements for lease financing deliveries. Management stated that short-term assets (CPU and GPU servers) will account for approximately two-thirds of capital expenditure, similar to the structure in the previous quarter.
Supply chain benefits significantly
According to Morgan Stanley analysis, among more than 50 technology companies covered, revenue contributions from cloud capital expenditures average 45%.
As the number of AI tokens processed monthly grows exponentially, total revenue growth rates for Google Cloud, AWS, and Azure are accelerating. Commitments to data center construction continue to expand, with key component suppliers generally reporting accelerated demand and extended visibility. Morgan Stanley believes these factors will collectively drive continued upward pressure on capital expenditure forecasts for hyperscale cloud service providers.
The growth rate of cloud revenue for the top four U.S. cloud service providers is expected to increase to a range of 30%-35% in the coming quarters, representing the strongest growth level since 2020. Based on the bank’s projections, non-AI-related cloud capital expenditures will maintain near-60% high growth in 2026, following an 80% year-over-year increase in 2025.
Market expectations continue to rise
Over the past year, market consensus expectations for cloud capital expenditures in 2026 have been significantly revised upward by approximately 100%, with a cumulative increase of about $370 billion. Morgan Stanley pointed out that as long as the industry maintains a positive qualitative outlook for cloud spending, related forecasts still have room for further upward revisions.
All four major U.S. hyperscale cloud service providers indicated that they will face supply constraints for key computing resources throughout 2026 and will need to address rapidly surging demand through accelerated investment and optimized capacity allocation. This persistent supply-demand imbalance provides structural support for maintaining high-intensity growth in capital expenditures.
Editor/Doris
