Part 1: Data Center Series
A fundamental understanding of data centers is essential to understand the investment case.
In this three-part series we will examine the data center investment thesis and its impact on emerging market real estate investors. The parts are:
👉 Introduction to Data Centers
Drivers of the Investment Thesis
Cautionary Note (i.e., 🐻 Bear Case)
What are Data Centers?
Data centers occupy physical rooms, within buildings, which house IT infrastructure for building, running, and delivering digital applications and services. Below is an image of the first data center from the 1950s which resulted from a partnership between IBM and American Airlines to create a passenger reservation system.
Data center design was, and still is, built around safely supporting the IT infrastructure which stores, processes and transmits data. The racks provide physical protection, and are designed to create sufficient spacing between the component parts for cables and air to pass. The huge bundling of cables is required to power and connect all the network elements. They must be carefully positioned so as not to hinder airflow required to cool the sensitive equipment.
The more advanced and powerful the data center, the more important cooling becomes. High-end servers are generally elevated from the ground, and cooled by specialty air conditioners which strictly regulate temperature and humidity. Modern data centers (i.e., 30kW+) utilize water cooling infrastructure as an additional layer of protection.
Modern Trend of Co-Locating
In the past, data centers were tightly-controlled, privately-owned and housed on-premises of the company utilizing the infrastructure. Today, a growing number of companies make use of remote facilities, or a network of facilities, shared by several customers. This trend is what is driving data center adoption and next week we discuss that, as well as the three models of data center ownership.
When you hear the word “cloud”, you should think of data centers, not clouds. “Cloud” was used early on as a marketing gimmick to entice more users to adopt (and pay for) the service. “Cloud” sounds better than what it actually is which is a shared offsite server, managed and maintained by a third-party for a fee. A data center.
If the “experts” are right, by 2025, 100 zettabytes (i.e., 100 trillion gigabytes) of public and corporate data will be stored in the cloud. This will represent about 50% of the total data anticipated to be stored in 2025, up from 25% in 2015. More companies are storing data in the cloud primarily to save money. For small- and medium-sized businesses, using a cloud services will be around 40x more cost-effective than storing and processing on-premises. Additionally, data security is usually enhanced and the risk of expensive downtime reduced.
Measuring Capacity
Capacity of data centers isn’t measured by the volume of data which go through the servers as one might expect. Instead, capacity is measured by the megawatts (“MW”) available and utilized by the servers operating within.
America accounts for about 40% of global data center power consumption (i.e., demand), which is expected to reach 35,000 MW in 2030, up from 17,000 MW in 2022, according to McKinsey. 2022 was a record year for the industry with 2,500 GW worldwide absorbed by new and expanded data centers.
Next Saturday we’ll look at the investment case for data centers, including the drivers of capital flows to develop data centers in the emerging markets.