Major cloud computing and hyperscale providers are investing heavily in their data center infrastructure to meet the growing demand for digital services and the rapid adoption of AI technologies, with capital expenditures (CapEx) expected to reach record levels in 2024.

In total, the annualized capital expenditures (CapEx) from Amazon Web Services (AWS), Microsoft, Google (Alphabet), Meta (Facebook), and Oracle reached $166 billion. This spending, which includes investments in data centers and servers, is expected to increase to $185 billion over the coming year.

As the cloud computing industry continues its rapid growth, understanding the CapEx of major players like Amazon, Microsoft, Google, Meta, and Oracle is crucial for investors, suppliers, IT professionals, and government agencies around the world. Dgtl Infra reviews the key components of each company’s global CapEx and provides commentary on their forecasted spending for 2024 and beyond.

Understanding Cloud and Hyperscale CapEx

The capital expenditures (CapEx) of top cloud service providers (CSPs) and hyperscale companies include investments in technology infrastructure such as data centers, servers, networking equipment, graphics processing units (GPUs), and fiber optic cable systems. Additionally, the CapEx of these “hyperscalers” encompasses property and equipment acquired through finance lease arrangements, corporate offices, research and development facilities, and building improvements.

#HyperscalerAnnualized CapExFuture CapEx
1Amazon Web Services (AWS)$28.6 billion$30+ billion
2Microsoft$56.0 billion$60+ billion
3Google (Alphabet)$48.0 billion$50+ billion
4Meta (Facebook)$26.9 billion$37.5 billion
5Oracle Corp$6.7 billion$7.5 billion
Total$166 billion$185 billion
Note: Annualized CapEx = extrapolated based on latest quarter and Future CapEx = projected over the next year.

1. Amazon Web Services (AWS)

In the latest quarter, Amazon.com reported capital expenditures (CapEx) of $13.9 billion. Of this total, Amazon Web Services (AWS) accounted for approximately $7.1 billion, implying annual CapEx of $28.6 billion. In comparison, over the past twelve months, AWS’s total capital expenditures amounted to approximately $25.3 billion, indicating that AWS is accelerating its pace of CapEx.

Amazon Web Services AWS Cloud Capital Expenditures CapEx Overlaid on Digital Infrastructure

Components of AWS’s CapEx

AWS’s capital expenditures primarily consist of investments in technology infrastructure, which include property and equipment acquired through finance leases and build-to-suit (BTS) lease arrangements. Specifically, these technology infrastructure costs cover data centers, servers, and networking equipment.

In total, AWS operates 38.2 million square feet of facilities, including 20.4 million leased square feet and 17.8 million owned square feet.

READ MORE: Amazon Web Services (AWS) – Data Center Locations

CapEx Forecast for 2024

Amazon anticipates a significant increase in capital expenditures in 2024, primarily driven by higher technology infrastructure spending to support growth in AWS. This includes investments in generative AI and large language models (LLMs).

AWS plans to add capacity for cloud region expansions and prioritize work on generative AI projects, which will contribute to the rise in CapEx for 2024. Initially, Amazon expects Q1 2024 to be its lowest CapEx quarter of the year, with spending increasing in the subsequent quarters.

2. Microsoft

In the latest quarter, Microsoft reported capital expenditures (CapEx) of $14.0 billion, implying annual CapEx of $56.0 billion. In comparison, over the last twelve months, Microsoft’s total capital expenditures amounted to $47.4 billion, indicating that the company is accelerating its pace of CapEx.

Microsoft Cloud Azure Capital Expenditures Digital Technology Data Center Infrastructure Connectivity

Components of Microsoft’s CapEx

Microsoft’s capital expenditures, including assets acquired under finance leases, support the demand for its Azure cloud and AI offerings, as well as services, including Bing, Microsoft 365, Microsoft Teams, LinkedIn, and Skype. These expenditures cover additions to property and equipment, such as new data centers, computer systems, corporate offices, research and development facilities, Microsoft Experience Centers, and various other equipment.

Microsoft Azure’s more than 300 data centers span over 75 million square feet and are connected with a diverse and global fiber network that traverses over 175,000 miles (280,000 kilometers).

READ MORE: Microsoft Azure’s Data Center Locations

Microsoft has also committed to constructing new buildings and making improvements, primarily related to data centers. These data centers require permitted and buildable land, power, and networking supplies. They house servers, specialized components like graphics processing units (GPUs), and network equipment.

CapEx Forecast for 2024

Microsoft is currently experiencing AI demand that slightly exceeds its available capacity, particularly within its Azure infrastructure. To address this, the company began accelerating its capital expenditures more than a year ago, focusing on purchasing servers and new data center expansions.

Microsoft plans to continue increasing CapEx in the upcoming quarters to add capacity in response to its growing pipeline. The company anticipates higher CapEx in fiscal year 2025 compared to fiscal year 2024, primarily driven by investments in its cloud offerings and AI infrastructure.

3. Google (Alphabet Inc)

In the latest quarter, Google’s parent company Alphabet reported capital expenditures (CapEx) of $12.0 billion, implying annual CapEx of $48.0 billion. In comparison, over the last twelve months, Google’s total capital expenditures amounted to $38.0 billion, indicating that the company is accelerating its pace of CapEx.

Google Alphabet Cloud Data Center CapEx Digital Background and Technology-Themed Infrastructure

Components of Google’s CapEx

Google’s capital expenditures primarily consist of investments in technical infrastructure. This infrastructure includes IT assets such as servers and network equipment for computing, storage, and networking needs. It also encompasses data center land and buildings to support the growth of its business and long-term initiatives, particularly in AI products and services. The largest component of Google’s CapEx is servers, followed by data centers.

Google’s capital investments also include office facilities, ground-up development projects, and building improvements (also known as “fit-outs”).

Google invests in land, buildings for data centers and offices, and IT assets through property and equipment purchases and lease arrangements. These investments provide capacity for the growth of its services and products, such as Gmail, Google Drive, Google Maps, Search, and YouTube.

READ MORE: Google Cloud’s Data Center Locations

For instance, the company’s data center construction projects are generally multi-year endeavors with multiple phases, involving the acquisition of land and buildings, construction of facilities, and installation of IT assets.

CapEx Forecast for 2024

Google has indicated that the company’s quarterly capital expenditures (CapEx) are expected to be at or above $12 billion throughout the year, implying annual CapEx of $48.0 billion. As a result, Google anticipates that its CapEx investment in 2024 will be significantly higher than in 2023.

4. Meta Platforms (Facebook)

In the latest quarter, Facebook’s parent company Meta Platforms reported capital expenditures (CapEx) of $6.7 billion, implying annual CapEx of $26.9 billion. In comparison, over the last twelve months, Meta’s total capital expenditures amounted to $27.7 billion, indicating a similar pacing of CapEx.

Meta Platforms Facebook Capital Expenditures Computing Overlaid Cloudy Digital Infrastructure

Components of Meta’s CapEx

Meta Platforms makes capital expenditures, including principal payments on finance lease right-of-use (ROU) assets, to support its technical infrastructure. This infrastructure is essential for delivering its products, including Facebook, Instagram, Messenger, and WhatsApp, and is used to support production traffic. Meta invests in its infrastructure primarily through data centers, buildings, servers, and network infrastructure, such as subsea and terrestrial fiber optic cable systems.

Meta has offices or data centers in approximately 40 countries and 90 cities across North America, Europe, the Middle East, Africa, Asia Pacific, and Latin America. This includes 21 data center locations globally. The company has designed and built its own data centers and key portions of its technical infrastructure, although a substantial portion of its technical infrastructure is also provided by third parties through colocation facilities.

CapEx Forecast for 2024

Meta anticipates that its capital expenditures for the full year of 2024 will range from $35 to $40 billion, an increase from its previous guidance of $30 to $37 billion (up $4 billion at the midpoint). While Meta has not provided guidance for years beyond 2024, it expects CapEx to continue rising in 2025 as it aggressively invests in AI research and product development.

This increase is driven by Meta’s acceleration of infrastructure investments to support its generative artificial intelligence (AI) training and inference needs, including Meta AI, which utilizes the Llama 3 model. Additionally, this CapEx supports the development of more advanced and compute-intensive recommendation models.

Meta plans to “invest significantly more in infrastructure over the coming years” to build advanced models and the world’s largest scale AI services. The company is growing its infrastructure capacity to ensure it has the training clusters needed to support the next generations of foundational AI models, including those required to train Llama 4 through Llama 7.

Meta also plans to ramp up construction on sites featuring its newly announced data center architecture, with expectations of significant future build activity.

5. Oracle Corp

In the latest quarter, Oracle reported capital expenditures (CapEx) of $1.7 billion, implying annual CapEx of $6.7 billion. In comparison, over the last twelve months, Oracle’s total capital expenditures amounted to $6.0 billion, indicating that the company is accelerating its pace of CapEx.

Oracle Corp Cloud Capital Expenditures CapEx Data Center with a Digital Circuit Board Infrastructure

Components of Oracle’s CapEx

Oracle’s capital expenditures enable the company to expand its existing data center capacity and establish new data centers in different geographic locations. This expansion supports the growth of Oracle Cloud, meeting both current and anticipated customer demand.

Oracle delivers its cloud services through global data centers, nearly all of which are leased from colocation providers. These data centers are equipped with servers, graphics processing units (GPUs), network equipment, and various improvements.

READ MORE: Oracle Cloud’s Data Center Locations

Additionally, Oracle’s properties include both owned and leased office facilities dedicated to sales, support, research and development, services, manufacturing, cloud operations, and administrative functions. In total, Oracle owns or leases over 30 million square feet of space in various locations across the U.S. and internationally.

CapEx Forecast for 2024

Oracle anticipates that its capital expenditures for fiscal year 2024 will be between $7 billion and $7.5 billion. This projection is driven by the accelerating need to expand cloud capacity, given the significant backlog and pipeline Oracle is currently facing.

To meet this demand, Oracle plans to build 100 additional cloud data centers. The company is experiencing billions of dollars in contracted demand that exceed its current supply capabilities. Consequently, Oracle is constructing some of the world’s largest AI data centers, including one in the U.S. that is large enough to accommodate eight Boeing 747s “nose to tail” in a single facility.

Additionally, Oracle is building 20 new cloud data centers colocated with and connected to Microsoft Azure. These data centers will house an initial 2,000 full racks of Exadata database machines, providing sufficient capacity to serve tens of thousands of customers. This expansion underscores the strong demand for Oracle’s services and the growth of multi-cloud agreements.

Frequently Asked Questions

Digital Glowing Compute Instance Composed of Lights Hovering Above a Networked Infrastructure

What is the Difference Between CapEx and OpEx in Cloud Computing?

In cloud computing, CapEx (Capital Expenditure) and OpEx (Operational Expenditure) represent two different approaches to financial resource allocation and accounting.

From the customer perspective, the main differences between CapEx and OpEx in cloud computing are:

  • CapEx involves upfront investments in hardware, software, and infrastructure, which are then depreciated over time. In contrast, OpEx is a pay-as-you-go model where expenses are incurred based on the actual consumption of cloud resources
  • With CapEx, the customer owns and maintains the infrastructure, requiring a dedicated IT team for management and support. OpEx shifts the responsibility of infrastructure management to the cloud provider, allowing the customer to focus on its core business activities
  • CapEx requires significant initial investments and long-term commitments, making it less flexible for scaling resources up or down. OpEx provides greater flexibility and scalability, as customers can easily adjust their resource consumption based on demand, paying only for what they use

From the cloud provider perspective, CapEx and OpEx represent different aspects of their business model and cost structure:

  • CapEx for a cloud provider involves significant upfront investments in building and maintaining the physical infrastructure, including data centers, servers, networking equipment, and software licenses. These investments are long-term and are depreciated over several years
  • OpEx for a cloud provider includes the ongoing costs of running and maintaining the cloud infrastructure, such as electricity, cooling, staff salaries, and network connectivity. These expenses are incurred on a regular basis and are necessary to keep the cloud services operational
Mary Zhang covers Data Centers for Dgtl Infra, including Equinix (NASDAQ: EQIX), Digital Realty (NYSE: DLR), CyrusOne, CoreSite Realty, QTS Realty, Switch Inc, Iron Mountain (NYSE: IRM), Cyxtera (NASDAQ: CYXT), and many more. Within Data Centers, Mary focuses on the sub-sectors of hyperscale, enterprise / colocation, cloud service providers, and edge computing. Mary has over 5 years of experience in research and writing for Data Centers.

LEAVE A REPLY

Please enter your comment!
Please enter your name here