In investing, as with most things in life, we all have our winning moments and we also have our not-so-glorious moments. And while there is a strong human tendency to keep make bombastic remarks about our winning moments, there is usually very little commentary around our mistakes. Why so? Well, why would anyone want to admit that they were wrong. There is, of course, the optics of the situation. It is hard to look smart while talking about your mistakes. On the contrary, you appear to be a genius when talking about your winners. For content creators, there is also the risk of losing your subscriber-ship or viewership or followers etc.
I want to take a different stance. Investing is hard. Individual stock picking is even harder. And as an investor, you are bound to make mistakes. This is a fact and there is no way to circumvent this. And we need to learn from our mistakes in order to become better at investing. This adage is true about most things in life. And the only way to learn is to be open and frank about where you think you might have gone wrong with your decision making process. Refine your strategy and ensure you never repeat this mistake again.
So with background out of the way, let me talk about one of my biggest investing flops. Lets talk about Digital Realty Trust (ticker: DLR)
What is DLR?
Since I have not yet covered DLR in a deep-dive post on this blog, it is only fair that I spend a few sentences explaining what is this business all about. DLR is a leading provider of data center and interconnection solutions for customers. Its primary focus lies in owning, acquiring, developing and operating data centers. Per their latest 10-K report, as of Dec 2022, they own a total of 316 data centers around the world, of which 132 are in the USA, 114 in the Europe and the rest scattered across Asia, Africa and Australia.
The company boasts of attracting a high-quality and diversified mix of customers (over 4000) which includes names such as IBM, JPMorgan Chase, AT&T, Verizon, Oracle, Meta etc.
The data-center narrative
If you consider any major technological advancement in the last 5-10 years, or in the upcoming 5-10 years, it would be hard to disassociate this from some form of technology in data processing. This includes areas such as internet traffic, artificial intelligence, communication networks etc. the amount of data generated in each of these areas is vast. And enterprises that have an expertise in data management are considered to be the technological power-houses of today and are in an enviable position for the next decade.
Research indicates that cloud computing market is expected to grow at an astonishing CAGR of nearly 15.8% over the next 6-7 years. It is hard to argue against these numbers given the evidence today.
DLR, and DLR like businesses, have benefited from this narrative. If you are investor and you strongly believe in this narrative, you can play this in the real-estate sector through investing in REITs such as DLR and Equinix, two of the largest players in this industry.
The rationale is pretty straightforward. It is incredibly hard for a new player entering this market to steal market share away from the big-boys. Owning and maintaining such data centers is an expensive proposition so it becomes very hard for new players to compete.
The data-center business landscape
Lets first take a step back and explore the landscape of the data-center business. If you study the large software enterprise companies, there are three primary ways in which they can manage their data:
- Manage your data in-house: Yes, software companies can choose to store and manage their data in-house on their own servers. This is the best-case scenario because such companies have full control over their data. There are no concerns regarding security, as the IT team that manages these servers, the servers themselves and everything else related to the data management are fully controlled by the company in question.
- Relying on co-located data centers: The in-house solution, while great, is not a scalable solution. Obviously, once the data set grows to a large enough size, maintaining them on in-house servers becomes problematic. Then, the cost associated with maintaining these servers i.e. physical real-estate, cooling solutions, backups, recovery time, server maintenance etc. is a huge headache. Enter the co-located data centers i.e. players like DLR and Equinix. This option allows the enterprise firms to use the servers at a server farm maintained by the co-located data centers. The legacy data centers, in turn, collect a rent and are responsible for providing the high-bandwidth network connections, performing routine maintenance etc.
- Hyperscalars or Cloud-solution providers: The third approach, and the one that is seeing massive growth especially in the last few years, are cloud-solution providers like Google (Google Cloud), Microsoft (Azure), Amazon (AWS) etc. This allows enterprises to host their data on the servers owned by these cloud-providers in exchange for a “rent”.
Per the recent trends, while the co-located data centers generated peak revenues around the 2016-2017 time frame, returns on invested capital since then have been very poor. On the contrary, the returns from hyperscalars such as AWS, Azure etc. have been extremely impressive. It is therefore clear that hyperscalars are gradually stealing market share from the co-located data centers.
There is an interesting dynamic here: While the cost associated with spinning up a new cloud-based data center is orders of magnitude cheaper than a brick-and-mortar co-located data center, there are situations where the hyperscalars would choose to use a physical data-center provided by a legacy data-center company because of the cost associated with ramping a new cloud data center vs the returns generated at a particular location, makes it more favorable to go with the physical co-located data center at that location.
This makes the hyperscalars some of the largest tenants of the co-located data centers. But when it comes to pricing power, it appears that the pricing power favors the hyperscalars, NOT the co-located data centers!
The bear case
Obviously, there is more to this story than meets the eye. And I was lucky to actually have a fellow member of the dividend community, and someone whom I respect immensely, @EuropeanDGI point this out. It all started with a comment from him on one of my posts here on this blog, where he asked me what I thought about DLR’s inability to grow their FFO/share over the last few years. He was right in asking a very pertinent question.
It got me thinking and revisiting my investment thesis in this company. I decided to look at the period between 2018-2022.
Indeed, if you look at the FFO growth for DLR, it has basically gone nowhere in the last few years, growing at a laughably slow rate.
Another useful metric to look at when evaluating REITs is NOI (Net Operating Income). The trends here are also not encouraging.
From a dividend safety point of view, I always want to consider the generated free-cash-flow from a business. All else being equal, this is a definitive indicator of the quality of a business. FCF/share has also gone nowhere during this period.
While a dividend investor can rely on the EPS payout ratio for a traditional business, REITs are a little different. REITs are expected to pay a large portion of their earnings out as distributions to their shareholders to (a) not be taxed on these earnings, and (b) be qualified as a REIT in the first place. So the EPS payout ratio is a little misleading on that front. That said, there are no such gotchas with the FCF payout ratio. As one would expect, the FCF based payout has been trending upwards and well above the range for where the dividend can be considered as safe.
If this was not concerning enough, I also read in the news that Jim Chanos has a short stake in the legacy data center business. Chanos is of course legendary for having first raised concerns about Enron before it collapsed completely. Most of what he saying seems to match up with my own research. But he also raised concerns about the CapEx being reported by DLR which I found very interesting. Per Jim, DLR is deliberately understating their maintenance CapEx in order to paint a false picture of economic profitability. I wanted to dig into this claim further, so I looked into the numbers.
Indeed, CapEx during this period has grown by an astonishing amount, but what is interesting is that DLR is claiming that over 80% of this CapEx is being directed towards “Developmental Projects”. Interestingly, only ~10% of this CapEx is directed towards maintenance of the existing data center farms. In my conversations with peers from the industry, maintenance for these data centers is a very capital intensive proposition. This involves expenses such as running the air-conditioners, the physical location i.e. geography of the data center itself and external climatic conditions etc. I find it hard to believe that DLR while growing the number of data center locations around the world, has managed to keep their maintenance CapEx nearly constant during this duration. Something just does not seem to add up here.
Indeed with revenue generation being as slow as it is, and DLR burning an astonishing amount of ~$2.5B in capital expenditure, this business is burning way more cash than what is being brought in. Given this run rate per data center location, with a growth in number of data centers around the world, the business is actually shrinking in size rather than growing. The return on invested capital here is far less than the cost of capital, a recipe for disaster as far as an investment.
To their credit, DLR has been able to secure debt at a lower fixed interest rate, but in an environment where interest rates are rising, and given the capital intensive nature of this business, I was having some serious concerns about the safety of the dividend. A growing dividend seems completely out of question for the next few years.
And so I bailed out…
I submit that it would be hard to imagine companies like DLR and Equinix going obsolete anytime soon. But I got into this business expecting that the secular growth trends in the data center business would propel businesses like DLR. Hence I had this business as one of my “growth like stocks” in the REIT space. What I am learning now is that this is more like a mature slow-growing business that will struggle in rocky times. There are far better quality businesses that I can invest in that fit that description instead of DLR.
I also do not want to invest in businesses which seem to be a direct competition to the likes Amazon, Microsoft and Google in the cloud business. That seems like a money-losing proposition any which way I look at it. If I want to invest in the cloud business, I am better off investing in these big-tech giants rather than the likes of DLR and Equinix.
For these reasons, I sold out of DLR completely. I made a loss on this investment, but I did not want to waste a minute after I realized what a jackass I had been for misunderstanding this business completely.
- Investing in individual stocks is hard as is. Investing in REITs is even more harder. Frankly, I had a migraine after reading through DLR’s 10-K because it is not a straightforward business to analyze.
- Businesses evolve. DLR had a great run around the 2016-2017 time-frame. But it has been a terrible business to invest in since then. It is important to not be suckered into the narrative around a business, because that is only part of the entire picture. Numbers never lie.
- DLR continues to be touted as a “great” dividend growth stock on #fintwit/#divtwit. However, I have yet to come across a cohesive and coherent bull thesis for DLR that can refute the findings of my own research or that of Jim Chanos.
- Don’t get defensive when someone raises a pertinent question about your investment decision. In fact, welcome such questions, because they will force you to go back and deeply introspect and think about your investment. You will either come out with a stronger thesis OR change your mind completely and avert a potential disaster.
Investing is no different than life. You win some and you lose some. This was one in the latter category. And just like I love to discuss my winners, I should also discuss my losers because both are a part of my journey as an investor.