Introduction 

Data centers are among the most important pieces of infrastructure in the modern world, and among the least understood. They are the physical back bone of our digital lives, quietly enabling everything from cloud computing  and artificial intelligence to online banking, shopping, streaming media, GPS navigation, healthcare, and government services. Trillions of dollars of economic activity depend on them every day, yet for most people, even sophisticated investors, they remain opaque, inaccessible, and difficult to evaluate.

I am by no means a professional writer. I am a real estate and finance professional who happens to know more than most about data centers and the data center ecosystem. Over the years, it has become increasingly clear to me that there is a widespread lack of understanding of data centers in general, including as a real estate and infrastructure investment asset class. Most people can have a basic, if not solid, grasp of traditional real estate sectors. Anyone can walk into an office building, visit a shopping center, rent and live in an apartment, lease a self-storage unit, or tour an industrial warehouse and intuitively understand how the income, expenses, and general risks around these assets function. Their economics are relatively straightforward. Supply and demand are visible if you know where to look. Comparable data including land costs, construction costs, rents, vacancy rates, and stabilized yields are abundant. Banks and lenders have decades of experience underwriting these properties. 

Data centers are very different. 

They are typically hidden behind high-security fences, physically guarded around the clock. They can cost 10, 20, 30 times more to build and operate than conventional real estate assets. They are highly complex facilities, requiring sophisticated electrical and mechanical systems layered with multiple levels of redundant systems to minimize downtime and maximize uptime. The demand side is driven by a relatively small group of technology companies, roughly a few dozen companies referred to as hyperscalers, cloud service providers, enterprise users, public sector/government demand, and an emerging tenant class referred to as neo-cloud companies. Most of whose  requirements are shaped broadly by cloud computing and AI. Understanding all these dynamics, including whether these companies will build data centers for themselves, lease from third parties, or renew existing leases, is far  from intuitive. 

Even the basic economic metrics are unfamiliar. Rents are not quoted in the common per square foot or square meter metric; rather, they are quoted per kilowatt of delivered power. Data centers consume large amounts of electricity, in some cases placing real strain on local and regional power grids. They require connectivity to fiber-optic networks, and depending on design and location, they may also require significant water resources for cooling.  

To complicate matters further, comparable transaction and operating data are often difficult to obtain, as tenants, operators, and developers tend to guard information closely. And yet, despite this complexity, data centers have  become one of the fastest-growing real estate and infrastructure asset classes in the world. 

In this book, my aim is to unpack these issues in a clear and accessible way. My goal is to demystify how data centers are designed, built, powered, financed, leased, and operated. Whether you are merely curious, evaluating a  data center as an investment opportunity, sitting on an investment committee for a pension fund, or considering a career in the data center ecosystem, this book is intended to provide a practical framework for understanding what  drives value, risk, and long-term performance in this sector. 

A brief word about my background. I have spent more than 30 years working in corporate real estate and finance. In 2014, I was fortunate to join Amazon Web Services, where I spent eight years until 2022. During that time, I led real estate and land development teams across much of the world, supporting the growth of AWS’s global cloud infrastructure. My geographic responsibilities included regions spanning North America, Latin America, Europe, the Middle East, Africa, and Asia Pacific. I was directly involved in  the development of data centers from Stockholm to Frankfurt to Cape Town, from Tokyo to Sydney to Mumbai, and from Montreal to Silicon Valley to Sao Paulo. In my final years at AWS, I also led teams designing cloud infrastructure for US government agencies and foreign government customers. 

In 2022, I transitioned to the investment side of the industry, joining PGIM, the global asset management business of Prudential Financial. I have been extraordinarily fortunate throughout my career to work alongside exceptionally talented colleagues, external consultants, and service providers, whose expertise and insights have shaped my understanding of this industry. I am truly grateful to so many colleagues and friends who tolerated my incessant questions in search of a broader understanding. 

In my research, I found a number of books that dive deep into topics including data center engineering, construction, design, cooling, etc. These books, while extraordinarily thorough on their specific topics, are targeted  at a relatively narrow audience segment. I’ve intentionally written this book from a broader perspective and at the “every-person” level. While this could  easily be a 1,000-page book, diving deep into each topic, I don’t believe that’s what most readers want. Rather, I’m keeping this at an intermediate level;  basic enough to readily understand, but with enough detail and granularity to satisfy the somewhat curious and nerdy among us. My colleagues in the data  center sector largely know much of what is included in this book, as we speak this language nearly every day of our lives. Though I’m hoping this book helps all of us to explain this rapidly emerging sector to colleagues, current and future business partners, friends, our spouses/partners, children, and possibly even our parents. 

Regarding the structure of this book, it’s not a novel where the story builds to a twisting crescendo in the final chapter. Rather, each chapter covers a different topic or element of the data center ecosystem. Feel free to skip  around to chapters where you have a specific interest. Like in many sectors, the data center industry uses its fair share of unique terms and acronyms, and I have tried to explain those more thoroughly in the earlier chapters and less so in later chapters—fair warning. 

While this book is fairly “heavy” on the topics covered, I’ve tried to lighten it up by starting a number of chapters with a few personal stories from my travels. I hope you enjoy them. 

At their core, data centers are facilities that house racks and racks of computer servers that form the foundation of the internet, designed to do so as efficiently, securely, and reliably as possible. Critically, they are engineered to  operate nearly continuously. Most are designed to achieve 99.999% or greater availability, often referred to as “five nines” uptime, which allows for only minutes of unplanned downtime over the course of a year before contractual  financial penalties apply. When you send an email or a text message, shop online, use social media, stream a movie, book a flight, use digital maps, consult a medical professional online, or interact with an artificial intelligence  system, you are relying on data centers. They are invisible to most of us, but indispensable to all.