Data Centers vs Hosting Providers: Why the Distinction is Critical for Investors

by Fil Zucchi, Minyanville staff writer
Over the weekend, Barron’s ran a negative piece on the “data center” companies. It argued that valuing these stocks based on a multiple to EBITDA, rather than a multiple of earnings or Free Cash Flow (FCF), overstates the economic power of these companies. To that I say, "true and false."

As pointed out in a research note by Stifel Nicolaus, Barron’s makes no distinction between “data centers” and “hosting companies.” But understanding the difference is critical accurately valuing the stocks. The former lease “physical space” where customers can place their servers, and the data center supplies power, cooling systems, and storage facilities (see Equinix Inc EQIX “Business” section in its Form 10K). In the case of “hosting companies” such as Rackspace RAX, it may or may not own the real estate in which the servers reside, but in any event its primary business is to “lease” its servers to clients who want to host their websites without having to manage the IT infrastructure. (See RAX “Business” section in its Form 10K.)

From a valuation standpoint, the distinction is critical. The EQIX of the world spend the bulk of their capital expenditures on long-lived assets such as real estate and power and cooling infrastructure. EQIX builds a new data center when it needs more space for its clients’ gear, just as a straightforward office park developer would do for new tenants. On the other hand, the RAX of the world spend the bulk of their capex on fast depreciating servers and other IT infrastructure, which must be turned over rather frequently in order to serve up customer websites in a quality manner and/or to add new customers.

So it is perfectly sensible to view the Depreciation & Amortization (D&A) of EQIX’s assets just as office REITs treat the D&A of an office building, and the true value to shareholders stems from the distributable cash flow (aka Funds From Operations), while the real estate usually backstops the real estate construction and permanent financing. However, since “hosting providers” lease “server space” rather than “real estate space,” if one excludes the cost of constantly replacing/adding servers, the true funds available to shareholders are inevitably overstated. In reality, even though servers’ expenses may be capitalized, they are far closer to operating expenses.

So yes, valuing companies like RAX on a multiple of EBITDA distorts what RAX would be able to distribute should it be allowed to transform into a REIT (which it likely would not be). Whereas valuing EQIX or a Digital Realty Trust (DLR) on a multiple of adjusted EBITDA (aka Adjusted Funds From Operations once a company converts to REIT status), does reflect what the real, recurring distributable cash flowing power of the company’s operations.

One thought on differentiating between a DLR and an EQIX. EQIX serves “tenants” who care to be in proximity to specific other “tenants” so that each other's servers can work more efficiently. DLR simply houses tenants that need some room in which to shelve their gear and an adequate power and cooling system. In my humble opinion, EQIX’s clients ought to be valued at a premium because they require what is in essence “premium” space. So, at risk of “talking my book” paying less than 20x Adj. EBITDA for a company growing the same at a 20% plus rate, with premium clients and high barriers to entry seems eminently reasonable, particularly if one considers that some of EQIX tenants, such as Amazon AMZN and Salesforce.com CRM trade a much higher multiples, on far lower growth.

More From Minyanville

More Upside Likely for Risk Assets, but Key Resistance Levels Are Just Ahead for Bulls

Tech News: Samsung Galaxy S4 Will Turn Up the Heat on Apple

More Upside Likely for Risk Assets, but Key Resistance Levels Are Just Ahead for Bulls