Data Centers: Finding Opportunity in the Storm

Digital Realty’s guidance sent shockwaves throughout the data centers this week.  We had cautioned a month ago (and again earlier this week), that guidance was likely to come in below consensus (LINK).  Financial guidance came in as we expected, but two metrics were worse than we expected: 1) DLR lowered their return threshold for new data center builds; and 2) DLR guided to sharply declining lease renewal rates this year (LINK).  Both of these factors demonstrate the risk of pursuing hyperscale deals with little underlying asset differentiation, a risk which extends to the entire wholesale data center industry.  We believe the negative sentiment on the sector provides an opportunity to buy EQIX at depressed prices, since EQIX is largely insulated from pricing risk due to their focus on retail customers and their interconnection strategy.  We see potential for 40% upside to EQIX over the next 12 months.

DLR’s Return Thresholds Are Compressing...

DLR lowered their range of expected returns on new builds from 10-12% to 9-12%.  Returns are declining because customers are negotiating discounts as they sign larger deals (and deals are certainly getting larger: five years ago, a large deal was 5 megawatts; today, a large deal is 25 megawatts).  With a cost of capital around 6%, DLR can still earn a positive spread by targeting returns of 9%; however, as customers sign larger deals, returns will likely compress, which should compress the multiple on the equity as well.

…And Renewal Rates Are Falling

DLR expects lease renewal rates to be “down high-single-digits” in 2019, worse than guidance for them to be “slightly negative” in 2018.  There are two key drivers of the weakness: 1) downward pricing on legacy leases that were signed at above market rates; 2) a customer receiving discounted pricing for extending lease terms from 5 years to 15 years.  The first driver was well known and should not have been a surprise to investors, but the second driver was new.  Given DLR’s low cost of capital, giving one customer a price cut in return for long-term stability could make sense; however, it highlights DLR’s lack of pricing power (something that is associated with typical communications REITs that trade at lofty multiples).

Pricing Power: Those Who Have It, And Those Who Don’t

When we launched coverage of the data centers (LINK), we recommended owning retail data centers, like EQIX, rather than wholesale data centers, like DLR, precisely because we believe retail data centers have pricing power and wholesale data centers do not.  Since most of the value of communications infrastructure companies lies in the terminal value, the multiple is very sensitive to assumptions for terminal growth.  Tower companies are well aware of this, which is one of the reasons why they never compromise on pricing escalators, even while giving customers other concessions for signing large, multi-year deals.  We assume current escalators at the data centers can be sustained into perpetuity; however, if we are wrong, every 50bps decline in the terminal growth rate would lower our price targets by 10-15%.

EQIX’s Multiple Has Collapsed More Than DLR’s

Data centers have underperformed both REITs and the broader market over the past three months.  During this time frame, EQIX’s AFFO multiple has fallen by more than DLR’s, and EQIX now trades in-line with DLR on 2019 AFFO per share and a discount on 2020 AFFO per share, despite having a much faster growth trajectory.  Moreover, EQIX has historically delivered an ROIC that is 300-400bps higher than DLR’s, and we expect the gap in ROIC to widen as DLR lowers its return thresholds for new builds.  Since EQIX offers faster growth and a higher ROIC, we think there is a strong case to be made that EQIX warrants a premium multiple to DLR and other wholesale data center providers.

A Recovery Could Drive at Significant Upside At EQIX

Over the last five years, on average EQIX has traded at a 2x premium to DLR on P / AFFO.  If EQIX were to regain their historic premium, the stock would trade at 17x 2020 AFFO would be worth $432 (+17%), based on consensus estimates.  If we factor in our above-consensus estimate, the stock would be worth $449 (+21%).  If EQIX’s multiple recovers to their historic five-year average multiple of 20x, the stock could be worth $525 (+41%), based on our estimates.  While sentiment across the data center sector may remain negative on the new bout of pricing concerns, we think this presents a great opportunity to buy EQIX. 

For the full weekly review and updated comp sheets, see HERE.