Switch Has Superior Data Center Growth

Summary

• Switch has proprietary data center designs which make their developments more accretive than usual.
• This accretion is amplified by a large development pipeline relative to their existing footprint.
• Switch builds very high-end data centers capable of supplying more power than most current equipment uses.
• Looking for a portfolio of ideas like this one? Members of Portfolio Income Solutions get exclusive access to our model portfolio. Learn More »

Kwarkot/iStock via Getty Images

The Buy Thesis

Switch Inc. (SWCH) is well positioned with a massive development bank to capitalize on the continued incremental demand for data. In data centers the opportunity lies primarily in development and lease up rather than in ownership of stabilized properties. Therefore, I think SWCH is better positioned than peers with its land held for development dwarfing its existing footprint.

Switch also seems to be a superior developer with proprietary patents and designs as well as pooled purchasing which allow them to build higher tier data centers at a lower price than competitors.

As these developments come online SWCH will be able to dramatically increase its FFO/share and as the company converts to REIT status it will save on tax expense and be able to increase its multiple. Rising FFO combined with stable to rising multiple is a recipe for strong returns.

Development is better than stabilized assets

Data centers are an interesting real estate sector with a significantly different return profile than other sectors.

For standard REIT property types like office, apartments, industrial or retail there is a slight premium captured by developers in exchange for taking the risk. Development cap rates will generally be 1-2.5 percentage points higher than buying an already stabilized property. So you can buy at a 5% cap rate or take on the risk of development to get maybe a 6.5% cap rate.

Data centers are quite different. The going cap rates for stabilized properties are fairly low with each of the data center REITs trading at implied cap rates ranging from 4.87% to 6.51%.

Source: SNL Financial

In contrast, the yield for a successfully leased up development is often in the double digits.

That is a massive spread over stabilized and I think that makes development of data centers significantly more lucrative than buying stabilized assets.

Additionally, data center space is rather fungible within quality tiers. Tenants simply require standardized infrastructure that provides a certain level of redundancy in cooling and power. New data centers can be built to preset tier 1, 2, 3 or 4 specifications that provide higher and higher levels of redundancy which minimizes the downtime. Switch is introducing Tier 5, but more on that later.

Any given tenant will have a certain level of redundancy that they require, but there is minimal differentiation between different tier 3 data centers. This fungibility and lack of differentiation threatens lease rates of incumbent data centers with the newly built data centers able to compete on rate and threaten to pull tenants away unless existing centers cut price.

This shows up in weak same store net operating income growth for stabilized properties with Digital Realty (DLR) being one of the 10 weakest REITs in 3Q21 same store NOI due to their drop of 5.5%.

Source: S&P Global Market Intelligence, highlights from Author

Note that this drop is in an environment where REITs were broadly having tremendous fundamental success with the median REIT raising same store NOI 7.1%

Source: SNL Financial

Data centers have been labeled as a growth sector which is causing the market to pay high multiples for the data center REITs.

With challenged same store NOI growth as tenants have negotiating power on lease renewal, I think it is a mistake to pay a high multiple for data centers.

It pays to be more specific

Growth can mean many different things. In industrial, multifamily, manufactured housing and many other sectors growth is being captured by the incumbents because they can dramatically raise rent due to demand for their properties outstripping supply.

Growth in data center only refers to demand for data centers. It is not being captured by the existing data centers but instead captured by the developments. It is a dangerous thing to just buy “growth stocks” without knowing the source of that growth.

Most of the data center REITs have been able to grow their FFO/share quite nicely, but that has been the result of building new properties and leasing them up. It has not been from existing stabilized properties. The problem here is that as their portfolios grow ever more colossal, the drag of existing properties gets harder and harder to overcome and therefore the external growth of their developments will move the needle less and less.

Digital Realty is the worst positioned for this reason – terrible organic growth in a massive portfolio of stabilized assets.

Switch and Iron Mountain (IRM) fall on the other side of the spectrum with small existing data center assets dwarfed by massive development pipelines supported by well located land banks. I already wrote about Iron Mountain’s data center opportunity here so the focus of this article will be on Switch.

Development pipeline

Switch today has about 5.1 million square feet of data centers.

Source: SWCH

It has active developments of 1.31 million square feet anticipated to be delivered in 2022 and 2023 and another 1.9 million slated for 2024-2026.

Source: SWCH

Finally, Switch has another 7.87 million potential square feet held for development on land they already own. These developments are truly ground-up with an example photo of one such development below.

Source: SWCH

In total, the new developments are more than double the square footage of its existing data centers and Switch is not a small company.

To put the magnitude of this pipeline into perspective consider the following numbers:

• CoreSite (COR) has 2.8 million net rentable square feet – just sold for ~$10B
• CyrusOne (CONE) has 8.6 million gross square feet. – just sold for ~$15B

Thus, in terms of square footage, Switch is going to be building approximately a CONE plus a COR which the market just valued at a combined $25B.

Of course not all square footage is created equal, but in this case Switch is building at the highest specifications. Much of their development is going to be tier 4 with some of it even at Tier 5 or “Platinum” as Switch calls it.

A tier 5 data center has all the redundancy bells and whistles of a tier 4 data center allowing it to have 99.995% uptime as well as what Switch describes as:

“More than 30 additional key elements, such as long-term power system capabilities, the number of available carriers, zero roof penetrations, the location of cooling system lines in or above the data center, physical and network security and 100-percent use of renewable energy.”

Some of that is probably marketing mumbo jumbo, but the idea here is that these are top of the line buildings as good or better than everything else that is available today.

So Switch has a massive pipeline of top quality developments but neither big nor high-end are inherently good for shareholders. We know from countless REITs growing for the sake of growth that it doesn’t always translate into value creation for shareholders. Trophy properties are nice, but since one usually has to pay trophy price for them it doesn’t always perform as well as one would hope.

In this case, however, it looks exceedingly accretive.

Accretive Growth

Through the use of patents and proprietary data center designs, Switch is developing in a cost-effective manner. They also have power supply agreements and pooled purchasing which affords getting the specialized equipment at bulk pricing. Overall, they claim to be able to build a tier 5 data center at the cost of a tier 3 data center as described by Gabe Nacht (CFO of SWCH) at a recent Barclays conference.

These developmental advantages bear out in the numbers as well with a cash flow yield on invested capital of 14.8%.

Source: SWCH

Funding this development

Switch has a fairly low leverage balance sheet with net debt of $1.6 billion against a market cap of about $6.4 billion.

Source: SNL Financial

Thus, they are well below normal REIT leverage and have announced a willingness to move up to normal REIT leverage.

I think roughly the first $1B of developments can be funded just by levering up to a normal range and after that it would be a mix of debt and equity in typical REIT fashion.

ROIC >>>> WACC

Switch has one of the most impressive spreads between cost of capital and return on invested capital that I have seen. A table in the 10-Q reveals long dated senior notes at 3.75% and 4.125%.

Source: 10-Q

Going forward I think SWCH will have access to 3% cost of debt, but for the sake of conservancy let's use the 4% figure.

Switch is trading at 26X forward AFFO so that is an equity cost of capital of roughly 4% as well making for a blended cost of capital of about 4%.

Given the 14.8% return on invested capital of their developments, the spread is absolutely enormous.

Maybe inflation will make construction a bit more expensive or perhaps lease-up will take slightly longer than expected but this spread is so strong that the developments look wildly accretive even if things go a bit wrong.

This 10%+ spread on billions of dollars of developments is a recipe for substantial growth. The analyst community seems to agree with estimates on AFFO going up to $1.42 by 2024.

Source: SNL Financial, highlights from author

2024 only covers the first couple million square feet of developments with another 7.8 million square feet to be delivered after that.

This, in my opinion, is a decades long runway of double digit annual AFFO/share growth.

Valuation

With $1.00/share of AFFO estimated for 2021, SWCH is trading at just over 26X current AFFO.

That is roughly in line with other data centers, but I think SWCH has substantially more growth ahead.

Specifically, three aspects of Switch make its growth likely higher than peers:

1. Superior development capabilities make return on invested capital higher
2. Product quality protects lease up and lease rates
3. Magnitude of pipeline relative to current assets makes its developments move the needle more than those of peers.

Similar current valuation with much higher growth is exactly where we like to be at Portfolio Income Solutions.

REIT conversion catalyst

Switch is not currently a REIT but given the precedence of large cap data center REITs they are able to convert to REIT status without the need for a PLR (private letter ruling) from the IRS.

With announced plans to be a REIT as of 1/1/23, 2022 will be the prep year to get all accounting and tax structure in line.

I see the conversion as beneficial in earnings and a nice catalyst for its valuation. Since REITs do not have to pay corporate taxes, conversion legitimately saves shareholders money because Uncle Sam will not be taking a bite before it gets to shareholders. Of course he will claim his share from shareholders, but that already happens. At least upon conversion to a REIT the earnings will no longer be taxed twice.

While broader market companies tend to be valued on GAAP earnings, REITs are valued on AFFO and FFO which add back depreciation. Since depreciation is a non-cash expense and the properties are holding their value or appreciating, Switch is already quite profitable, but it doesn’t look all that profitable in GAAP numbers.

I believe REIT conversion will help the market see Switch’s true profitability and this could have strong implications for its valuation.

Anyone looking at Switch can see that it has high growth potential, but there is a big difference between a rapid growth company that might become profitable someday and one that is already profitable. Switch is the latter and upon REIT conversion the market will be more likely to value it as such.

Risks to thesis

Switch has a somewhat unusual share structure with both class A and class B common shares. The two securities are not strictly pari passu, but given that the B are in the process of converting to A I am valuing it as if they are already A and simply using the combined diluted share count.

There is some murkiness that comes with dual-class common shares and SWCH may trade at a discount until the capital structure looks a bit cleaner.

Lumpiness – in addition to the development pipeline being enormous, it will cashflow in a rather lumpy fashion. As properties complete and open up they will lease in stages resulting in some quarters having well above run rate growth and others almost no growth. Thus, we should anticipate more of a stair step than a ramp.

Disclosure: I/we have a beneficial long position in the shares of SWCH, IRM either through stock ownership, options, or other derivatives. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.