Advanced Search

$ 0 to $ 1,500,000

More Search Options
we found 0 results
Your search results

Dallas offers Under the Hood View to Larger Data Center Trends / Themes

Posted by Natalie Steinwand on September 20, 2018
| 0

September 19, 2018 ; Equity Research

Image result for wells fargo securities Communications Infrastructure & Telecom Services

Over the past 2 days, we visited 7 different data center companies, both public and private, in the Dallas market. Our itinerary included visits to Digital Realty Trust, CyrusOne, QTS, Compass Data Centers, T5 Data Centers, DataBank, and Flexential. Dallas is one of the more diverse data center markets in the US and offers an interesting “petri dish” view into broader trends. We list below our takeaways: 

DALLAS FEELS SLIGHTLY OVERSUPPLIED, BUT EMERGING HYPERSCALE DEMAND COULD FLIP SWITCH – The various players we spoke with viewed Dallas as a strong longer-term market, particularly with diverse enterprise demand and connectivity options. However, new supply delivered in the first half of 2018 has outstripped absorption, with only ~10-15 MW of absorption YTD trailing typical annual absorption of 35-40 MW. According to broker reports, vacancy in the broader Dallas market is ~20%. All companies felt confident that Dallas will become a more attractive market capable of supporting hyperscale deals of +10 MW. In fact, we heard from multiple sources that Facebook is looking at ~10 MW of capacity in Dallas. We believe QTS and DLR will both be competitive in bidding for that business. 

HYPERSCALE DEALS STILL COMPETITIVE, BUT WILLING TO INK LONG TERM DEALS – Not overly surprisingly, the market for large hyperscale deals remains extremely competitive. Several private players we spoke with indicated that cash yields are often in the high single-digit range for these projects, although we note this does range depending on build costs. While these yields are down from 2-3 years ago, it is relatively stable from our checks over the past year. While we acknowledge these yields are lower than historical levels, importantly these deals are still being signed with 10+ year terms (and 2-3% annual escalators in most cases). We believe this data point perhaps more than any speaks loudly to the long term runway for this important and growing customer vertical. For instance, CONE estimates the cloud companies will have capacity demands of ~5-10 gigawatts over the next 5 years. In some ways, we believe that building out supply quickly enough, and not underlying demand, will be a bigger issue over the next few years. 

INDUSTRY RIPE FOR CONSOLIDATION – Our contacts expect consolidation will continue among the various public and private companies remaining in the space. Some described it as an “accordion” type of trend where it will get bigger and smaller at different cycles especially given all the private equity infrastructure funds entering the space. We continue to believe some of the US players have larger global ambitions – with CONE as the next player to make a major international play. As we wrote last week, we believe Europe is a prime area of interest for many existing data center cos and infrastructure PE funds. 

MORE TAKEAWAYS IN THE NOTE FROM CONE ANOver the past 2 days, we visited 7 different data center companies, both public and private, in the Dallas market. Our itinerary included visits to Digital Realty Trust, CyrusOne, QTS, Compass Data Centers, T5 Data Centers and DataBank. Dallas is one of the more diverse data center markets in the US and offers an interesting “petri dish” view into broader trends. We list below our takeaways:

DALLAS FEELS SLIGHTLY OVERSUPPLIED, BUT EMERGING HYPERSCALE DEMAND COULD FLIP SWITCH – The various players we spoke with viewed Dallas as a strong longer-term market, particularly with diverse enterprise demand and connectivity options. However, new supply delivered in the first half of 2018 has outstripped absorption, with only ~10-15 MW of absorption YTD trailing typical annual absorption of 35-40 MW. Vacancy rates in the market are ~20% according to recent broker estimates. All companies felt confident that Dallas will become a more attractive market capable of supporting hyperscale deals of +10 MW. In fact, we heard from multiple sources that Facebook is looking at ~10 MW of capacity in Dallas. We believe QTS and DLR will both be competitive in bidding for that business.

HYPERSCALE DEALS STILL COMPETITIVE, BUT WILLING TO INK LONG TERM DEALS – Not overly surprisingly, the market for large hyperscale deals remains extremely competitive. Several private players we spoke with indicated that cash yields are often in the high single-digit range for these projects, although we note this does range depending on build costs. While these yields are down from 2-3 years ago, it is relatively stable from our checks over the past year. While we acknowledge these yields are lower than historical levels, importantly these deals are still being signed with 10+ year terms (and 2-3% annual escalators in most cases). We believe this data point perhaps more than any speaks loudly to the long term runway for this important and growing customer vertical. For instance, CONE estimates the cloud companies will have capacity demands of ~5-10 gigawatts over the next 5 years. In some ways, we believe that building out supply quickly enough, and not underlying demand, will be a bigger issue over the next few years.

INDUSTRY RIPE FOR CONSOLIDATION – When the 5th largest player in the industry is less than 20% the size of the largest player (in terms of square feet), it feels like consolidation is close at hand! That was one of the main takeaways from our visit to the 7 companies in our Dallas adventure. To put this in perspective, DLR (by far the largest player in terms of square footage) has 8.7MM net operational square feet (and may be getting bigger in South America if recent press reports are to be believed – DLR has not commented) and Flexential – which we visited is the fifth largest player with 1.612MM net operational square feet (EQIX, CONE and SWCH represent players #2-4 in order on this same metric). When we asked the management teams if they expected consolidation to continue in the space, most agreed and pointed to past telecom segments – wireless, RLECs etc. But the timing of this remains a question as to how quickly such a wave could come through. Some described it as an “accordion” type of trend where it will get bigger and smaller at different cycles especially given all the private equity infrastructure funds entering the space. We continue to believe some of the US players have larger global ambitions. As we wrote last week, we believe Europe is a prime area of interest for many existing data center cos and infrastructure PE funds.

SECONDARY MARKETS GETTING MORE INTERESTING ON ENTERPRISE SIDE – Some of the smaller players we met with (DataBank, Compass) spoke much about the importance of growing a footprint in some second tier cities. While both these cos also have exposure in some of the large data center markets, there seems to be strong demand in some secondary markets such as Minneapolis and Salt Lake City. We think this demand is increasingly coming from enterprise customers with footprints in those cities. On the hyperscale cloud side, it seems demand continues to concentrate in the major data center markets, notably Northern Virginia, Phoenix, Santa Clara and Chicago.

TAKEAWAYS FROM CONE MANAGEMENT MEETING – We met with Gary Wojtaszek (CEO), Diane Morefield (CFO), Jonathan Schildkraut (Chief Strategy Officer) and Michael Schafer (VP of IR). CONE remains very bullish on longer-term demand trends, as their customer conversations and internal projections estimate potential cloud capacity demand of ~5-10 gigawatts over the next 5 years. The top cloud service providers are increasingly receptive to leasing space and power vs. building data centers on their own. The near-term focus remains on integrating the Zenium acquisition and expanding their European footprint, which should be ~250 MW by the end of 2019. CONE noted the Zenium acquisition should be completed in about a month. Longer-term, CONE would like to expand into the Asia-Pacific region, with Australia, Japan and India as potential new markets. CONE would likely look for a joint venture with local partners for this international expansion. Regarding its credit rating, CONE is “hopeful” that they could get to investment-grade within the next 12 months, although noted the exact timing remains uncertain. CONE estimates that their credit spreads could compress ~50-100 bps as an investment grade company and that they could push leverage up to ~6x once they achieve those ratings. However, we would expect CONE to continue opportunistically issuing equity to keep leverage in the ~5x range over the next year, particularly with a likely increase in 2019 capex over 2018 levels.

TAKEAWAYS FROM QTS MANAGEMENT MEETING – We met with David Robey (COO), Stephen Douglas (VP of IR), and Aaron Hicks (Dallas Site Director) at QTS’s Irving Dallas center facility. QTS noted that its core colocation demand has actually strengthened in recent quarters, with enterprise customers that historically bought in ~100kW increments going up to 500kw+ in terms of deployment size. On an apples-to-apples basis, QTS sees colocation pricing on new deals remaining roughly flat. QTS stated that they fully expect to sign at least one more hyperscale deal this year, with the longer-term expectation of 1-3 hyperscale deals each year, with deal sizes in the range of 5-40 MW. QTS is actively engaging in conversations with anchor tenants for the 32 MW shell that they’re building in the Dallas/Irving facility.

Leave a Reply

Your email address will not be published.

Compare Listings