One could argue that there are more colocation providers and data center facilities than at any other time in history. M&A is up, data center expansion is strong, construction is up, land acquisition is strong. Investor appetite for data center real estate is near all-time highs.
There is just one problem...
Supply & Demand for Data Centers
There is literally no supply of available data centers that meet investor criteria. What is that criteria? Investors interested in participating in the data centers want a credit-worthy, anchor tenant, or a large number of retail customers included in the sale. Right now, that supply simply does not exist. Think about it this way. Why would a data center owner sell a cash-flowing property with creditworthy tenants and upside for continued growth? It would have to be a pretty big number if you ask me.
Do not get me wrong. There are plenty of data centers for sale. Some built in the early 1990s. Others built in the 2000s. Some built for retail colocation. Others built as enterprise data centers. In nearly all cases, the data centers for sale have 100% vacancy, single-tenant planning to exit, and relocate to a colocation facility or cloud. Some data centers for sale may have a handful of retail colocation customers. However, these are not the type of data centers that institutional investors, pension funds, or VCs want to purchase. They are looking for a turn-key operation that provides predictable cash-flows. The so-called rent roll and creditworthiness of the tenants is mission-critical.
Outdated Data Centers Are a Hard Sell
Another major challenge with these types of data centers is that they’re outdated by today’s standards. Less than 500kW of total power in a data center is not going to cut it in today’s highly competitive colocation environment. Plus, antiquated HVAC will not handle the high-density requirements of today’s colocation customers. Another factor is the curb appeal. Yes, I just said curb appeal. Newly constructed data centers have all of the features and amenities that retail and wholesale colocation clients want – office space, conference rooms, break rooms, staging areas, high ceilings, storage areas, loading docks, and more. Not to mention advanced security systems, personnel, and monitoring. The Mission Impossible presentation if you will.
Cloud Adoption Creates Market Change
There is one other factor at play here. The cloud. IT workloads are moving to the cloud whether you like it or not. Take a second and think about how technology is changing. Do you have an exchange server? What about a database server? What about servers that run your applications and host your website? It is true that some workloads run better on physical servers and certain software vendors make it hard to license their software on virtual infrastructure. That will be the case for years to come.
When you look at who is buying colocation services and why you see a clear picture of what is truly going on in the market. Business and enterprise clients are adopting the cloud in a big way. They are moving more and more IT workloads to the cloud. It could be public, private, or hybrid cloud. It could even be multi-cloud deployments. At the same time, they need to maintain on-premise or collocated physical servers for certain circumstances. The physical footprint for IT infrastructure is shrinking.
So, who are the future buyers of colocation services? Are you ready? It is the hyperscalers such as AWS, Microsoft, Google Cloud, IBM Cloud, Oracle, Alibaba Cloud, and others. It makes sense right? If business and enterprise clients are moving IT workloads to the cloud, cloud service providers (CSPs) require more data center space to support those workloads. You are probably thinking that CSPs build their own private data centers. That is correct. However, they often leverage wholesale colocation until they get to a critical mass where it makes more sense economically to build a data center than lease one. Every CSP is different when it comes to the threshold for building a data center.
Retail Colocation Continues Growth Path
Is there a play for retail colocation? Absolutely. Business and enterprise clients are building their own private clouds in colocation facilities. There are many instances where physical servers, private cloud servers, and storage make more economic sense than going with one of the big three cloud providers. Like I said before. There will always be a need for physical servers based on the workload type and software licensing.
I do offer this word of caution for those concerned about the cost of cloud services. The big hyperscalers have economies of scale, talent, and workforces that simply cannot be matched by a single business or enterprise. Look for this trend to continue. They are simply going to get larger and more powerful. They will find ways of increasing operational efficiencies to drive down the cost of cloud services. Think five to ten years from now. What will the data center industry look like?
Future Technologies Drive Colocation Demand
There is another future play for colocation. Drumroll, please. It is called the edge or edge computing. It is not here yet but it is coming with a number of other technological innovations like 5G, internet of things (IoT), autonomous vehicles, and artificial intelligence (AI). The edge is essentially pushing IT workloads closer to the end-user wherever they are located. The main concept behind the edge is that there will be regional and centralized data center locations. Smaller, regional data centers will act as a bridge in communicating between end-users and central data centers.
Coming around full circle. The market for smaller, older, outdated data centers is essentially non-existent today from a commercial real estate perspective. At best, these are powered shells in a strategic location than need to be acquired and built out by a colocation provider. The amount of capital required to retrofit and operate one of these data centers is not even a consideration for 99.999% of potential data center investors.
That is today. However, could there be a play for these data centers in the future with edge computing. Wide-ranging adoption of edge is still five to ten years out from now. The demand for edge computing and edge data centers is still a big unknown. If I was a betting man, I would place my chips on regional data centers in rural or remote locations with the potential to become an edge data center in the future. There is already enough competition for colocation data centers in large metros.