Exclusive: Have We Reached an Inflection Point in Data Centers? CAPRE Texas Summit Coverage
AUSTIN, TX – The dynamics of the data center markets are changing, the talks of recession are on the rise, and an election is a year away. So, where are the most attractive domestic investment opportunities? One of the signature panels at CAPRE’s Text Data Center Summit last week, the 10:00 am session titled “Domestic Investment Opportunities: Have we reached an inflection point?” tackled these questions and more.
Moderator Miles Loo, Jr., Executive Vice-President at Newmark Knight Frank kicked off the session with a question for panelist Sami Badri, Senior Equity Analyst at Credit Suisse. “After a long economic expansion, no one knows when the recession will hit or how severe it will be,” offered Loo. “So what is your view on the timing of the next downturn – how will it affect the DC market?”
Badri replied that there’s no easy answer. Rather, all of the major players are telling different stories. “You could say that not a single data center operator is in line with the other operators’ results. That’s the biggest takeaway from Q3 2019 results.” However what is most interesting, according to Badri, is that the global economic recession narrative seemed like it would impact the data center industry by now, but that hasn’t been the case universally. “When it comes to consumption of equipment, you’re seeing a slowdown, but for the actual data center infrastructure, you’re not….it’s a mixed picture.”
Loo then pivoted to the recent acquisition of Interxion by Digital Realty as an example of unprecedented activity, and looked to Tony Wanger, an independent data center developer and industry veteran, to talk about why that’s the case at this moment in time. “What jumps off the page for me is the availability, cost, and number of participants supplying capital in this space. That’s changed dramatically. We couldn’t find a number in 1998,” he recalled. “There’s just money coming from everywhere. Pension funds, infrastructure funds, real estate funds, non-bank lenders, syndicated and high-yield lenders – there’s just a ton of capital.”
Rashad Kawmy, Partner at Boundary Street Capital then offered a different take on this topic, saying that the change in perception of real estate-oriented assets is key to the activity. “This previously didn’t exist. A lot of investors got confused about data centers – are they technology or are they real estate? How will you leverage it?” he offered. “That market has become much more efficient.”
“There is a lot more capital,” concurred Patrick Fear, Managing Director, AB Private Credit Investors, who then offered a word of caution. “But how permanent is that capital? Because you’re seeing spreads back up in certain parts of the market. You’re seeing people get tripped up on questions like, is the public cloud going to obliterate the need for colocation? And you’re seeing unable to differentiate between retail and wholesale models…you’re seeing people struggle for a services mix, rather than just charting for space and power.” In sum, Fear thinks that together, these factors could lead to a potential slow-down.
“A lot of us in this industry have wanted to know when data centers will be recognized as an asset class for themselves – as data centers,” shared Scott Widham, President of Widham Capital. “They’re not real estate with higher TIs, and higher rents, or towers with higher operational complexity. Or telecom networks that are finite….we’re not there yet.”
“The institutional real estate got in and out of the state relatively quickly,” he continued. “It’s a little bit more management-intensive than your traditional passive assets. Now they’ve landed with infrastructure funds, which seems to be going well, but there’s definitely a huge bias toward being creditworthy-type of operators. Which has left the rest of the space probably under-capitalized.”
Loo then asked the panel to fast forward a bit. “Will capital flows over the next five years favor different sectors of the infrastructure market than they have in the last few years?” he asked.
“I think the jury is a little bit still out on how these large infra-funds will fare as owners of these assets,” offered Fear. “I’m sure they’ll do fine, but whether or not they’ll actually achieve their targeted returns is still uncertain. Just as we saw with the telcos, while this is an asset class that has elements of real estate, it’s also technology. And we’ve seen that these buildings, by the time they get five to ten years in, there are often going to be generational assets that will occur over time. And as with telcos, they will require significant capital investments to stay relevant.”
Kawmy then chimed in quickly, pointing out that his team has seen big capital flows from Asia and Europe driving infrastructure investment in the United States.
“The way the market has stabilized since the late 2000s, is that the capital went to very specific underwrites,” Fear recalled, taking back the spotlight for another minute. “Since we’ve had this long period of expansion, it’s gotten more aggressive. We now have people spending money where the returns are going to be, no doubt, less good. And the impact that under-achievement still has on these big activity centers has yet to be seen.”
“What are the ingredients that make business models in our sector financeable?” asked Loo as a follow-up query, as the panel geared up toward its conclusion and really honed in on the nuts and bolts of the business.
“Well, for a lot of wholesale operators, their colocation businesses have stalled,” replied Widham, asking a follow up question of his own. “They’ve begun to initiate managed services businesses. So do you bifurcate those revenue streams?” He also shared that, by his observations, there are more data center operators looking for third-party capital, outright sales than he’s ever seen before.
“That’s a good point,” shared Badri, referring to the managed services point. “A lot of private operators have added, acquired or merged with MSPs. Here’s the fundamental in the private market – MSP revenue streams have higher churn rates and less visibility, making their creditworthiness and return on investment of capital a much tougher conversation. That’s in addition to the health of the assets, the original intent, and utilization of the actual facility which in most cases are sitting in Tier II or Tier III markets. So I’d say even though it’s increasing in growth, it’s happening at the expense of KPIs becoming a little more shaky.”
“We do delineate,” offered Kawmy, enhancing Badri’s response. “We don’t think all services are bad. We look at complexity and what’s being delivered, along with utilization. If you have a complex service that is closer to the application layer, or you’re dealing with complex applications where there’s a shortage of resources in house to deal with it, and you’re able to get a two to three year contract on a fixed-fee basis, that’s pretty financeable.”
“It sounds like markets for both data center and office tenants are pretty similar,” observed Loo, preparing to ask one final question, again focused on the future direction of an important sector in the industry — this time hyperscale. “You’re looking for good corporate tenancy, long-term durable income. But with data centers, as it relates to hyperscale, data centers can be owned or leased. Do you anticipate hyperscale providers who are leasing capacity today to ultimately turn off their leased facilities to build larger, owned data centers in the future?”
“I personally would be shocked if Microsoft, for example, renews many of their turn-key leases. I would expect them to turn to a model made of massive campuses in the middle-of-nowhere, rural areas,” predicted Kawmy, who had perhaps the most succinct response. “And they’ll likely maintain Tier I market campuses, effectively, as their Edge. They’ll still have 10-20 megawatts in Ashburn, but instead of that being their core compute for Azure or 365, that will likely be an on-ramp back to a campus somewhere else.”