Digital Realty Trust, Inc. (DLR-PK)

Think of Digital Realty this way: companies generate enormous amounts of data and need places to put it. They need secure, climate-controlled buildings full of servers and networking equipment. They need backup power in case of outages. They need redundancy across multiple locations so if one site fails, another takes over. Digital Realty owns and operates those buildings. The company leases space and power to everyone from hyperscale cloud providers like Amazon and Microsoft to smaller enterprises that cannot afford to build data centers themselves.

The business is simple on the surface. Build or acquire a data center. Fill it with racks of servers. Charge customers monthly to rent space and power. Collect the revenue. The customers stay put because moving is expensive and disruptive. Digital Realty has grown by doing this repeatedly, across the United States and then globally—from California to London to Singapore to Sydney.

Why size matters in data centers

Being big is almost everything in this business. A customer needs redundancy; they want their data in multiple data centers so no single failure brings them down. A large operator like Digital Realty with facilities in dozens of cities can offer that. A small operator with one or two sites cannot. Hyperscale companies like Amazon and Google can build their own data centers—they have the capital and the engineering talent. But even they often rent space from Digital Realty because it is cheaper and faster than owning and operating everything themselves.

Digital Realty’s size means it can absorb the enormous fixed costs of running a data center—real estate, power infrastructure, cooling systems, security, 24/7 staffing. A small operator struggles to spread those costs across enough paying tenants. With hundreds of thousands of servers installed across the world, Digital Realty achieves economies of scale. It can negotiate better rates from power suppliers. It can install networking gear that smaller competitors cannot afford.

Size also means Digital Realty can offer connectivity. Large data centers sit at the intersection of fiber-optic cables that carry the internet’s traffic. The company can connect its customers to each other, to major cloud providers, and to the internet backbone itself—a service that small data centers simply cannot provide.

How Digital Realty makes money

Revenue comes primarily from leasing space and power. Customers pay for rack space—physical room for their servers—and for the electrical power those servers consume. Power is the bigger cost driver. A typical server consumes as much electricity as a household. A data center with hundreds of thousands of servers consumes as much power as a small city. Customers pay not just for the megawatts they use but also for redundancy—backup power generators and batteries that keep systems running if the grid fails.

A secondary but growing revenue stream is cross-connects. Customers often want to connect servers in one rack to equipment in another. Digital Realty sells that connectivity service—a direct fiber connection between two points in the same data center or between two cities. High-frequency traders pay premium rates for this because microseconds of latency matter.

The beauty of the business is that revenue is mostly recurring. Once a customer rents a rack, they do not want to move. The switching cost is high. Revenue stays stable even when economic conditions weaken, because companies need to keep their data online no matter what. During the 2008 financial crisis, when plenty of companies went under, data center occupancy stayed strong.

The constraints of being big

Being big also brings constraints. Digital Realty is a real estate company in the regulatory sense—specifically, a Real Estate Investment Trust or REIT. REITs are required to distribute at least 90 percent of their taxable income as dividends to shareholders. That rule exists to ensure shareholders get consistent payouts rather than the company holding cash. But it also means Digital Realty cannot reinvest its earnings the way a traditional company can. Growth must come from borrowing, issuing new shares, or external capital.

Large data centers are also extremely capital-intensive. Building a new facility costs hundreds of millions of dollars. The company must obtain land, pour concrete, install electrical infrastructure, run fiber-optic cables, install cooling systems, and set up security. All of that happens before the first customer ever plugs in a server. For a company committed to returning 90 percent of earnings as dividends, that capital requirement is a perpetual constraint.

Real estate is also geographically bound. Digital Realty cannot operate a data center in a city where land is prohibitively expensive or where power is unreliable. The company must choose locations with nearby sources of cheap, abundant electricity—proximity to hydroelectric facilities in the Pacific Northwest, nuclear plants in France, or renewable wind farms in Texas. Once a facility is built in the wrong location, it cannot be moved.

Competition and the scale trade-off

The largest cloud providers—Amazon Web Services, Google Cloud, Microsoft Azure—can build their own data centers at scale and price customers aggressively because they use the space themselves. They can cross-subsidize compute and storage against real estate. For customers who want to use multiple cloud providers or who want independence from a single vendor, Digital Realty and its peers offer neutrality—data centers that welcome all tenants equally.

Other large REITs like Equinix and CoreWeave compete for the same customers. The advantage goes to whoever can offer the most locations, the most power, the most connectivity, and the lowest price. Scale helps, but it does not guarantee victory. A smaller, nimbler operator with strong real estate in high-value locations can outcompete a larger rival that has squandered its real estate choices.

Pressure from power and climate

Digital Realty’s biggest future constraint may be power availability. Data centers consume vast amounts of electricity—often enough that a single large facility draws as much power as a mid-sized city. In regions where electricity grids are congested, getting permission to build a new data center is a multi-year battle against regulators and neighbors. Climate change is reshaping where power is reliable and available; some historically cheap power regions are becoming unreliable, while others are emerging as stable.

The company is also under growing pressure to run on renewable energy. Customers, particularly technology companies, want to lease space in data centers powered by wind and solar. That pushes Digital Realty to invest in renewable energy contracts and to build facilities in locations with abundant clean power. Those locations may not be the cheapest overall.

How to research Digital Realty

Start with the annual 10-K filing (SEC CIK 0001297996). Look at how many data centers the company operates, where they are located, and what percentage of space is leased. Track the occupancy rate—the percentage of available capacity already rented. High occupancy means pricing power and growth constraints; low occupancy means room to grow but also unused costs. Watch the power consumption per megawatt and the prices customers are paying per kilowatt-hour. Those metrics reveal whether the company is handling supply and demand well in its markets.

Pay attention to where new data centers are being built and why. Are they in expensive coastal cities, in cheap power regions, or in locations where customers are clustering together? The quarterly earnings calls reveal management’s thinking about supply and demand. Are they optimistic about cloud spending? Are they seeing pressure from competition? Are they managing to raise prices or are they locked into long-term discounts from hyperscale tenants?