I’ve researched some super high-quality businesses so far for this newsletter and on the podcast. But I have to say, American Tower (AMT) might be one of the widest-moat businesses I’ve ever seen. I think Shawn would agree with me on that, though you could argue that Verisign’s moat might be even wider.

AMT has one similarity to Verisign that as value investors, we don’t particularly like. Both trade at price-to-earnings ratios above 29x. But the legend investor, Chuck Akre, who has held on to three different 100-baggers including AMT (Berkshire and Markel being the other two), would argue that a high-quality business at a premium price can still make for an incredible investment. For context, Akre made over 280x returns investing in AMT(!)

What AMT offers its customers is a product they can’t live without: cellular towers. These are pieces of hardware often occupying farmland or the tops of skyscrapers, ensuring that you have access to wireless communication networks. You’re likely using an AMT service right now if you’re reading this on mobile. Despite its massive infrastructure footprint, the business actually has very compelling unit economics and operating leverage (more on that in a moment.)

Let’s see if such a critically important business makes for an attractive investment at current prices.

— Kyle

We’ve all heard the following quote from Warren Buffett: “The most important investment you can make is in yourself”, but how many of us take that idea seriously?

Value investing, by nature, is a pursuit of continuous learning, requiring us to turn over many stones to find opportunities. Knowing which stones to turn over, though, is the hard part. You could play this game alone, or, you could connect with the best and brightest minds of the value investing world, building relationships that make you a better investor.

Personally, I’ve chosen the latter, and I don’t regret it for a second. My life has been enriched beyond words, thanks to the wonderful folks I’ve met in our Intrinsic Value Mastermind Community, and the best thing is, yours can be too.

Take the next step in your investment journey today by clicking below(!)

At its core, American Tower operates a simple business: they lease space on cell towers via a real estate investment trust (REIT) structure, which requires them to pay out a large portion of their profits as dividends in return for more favorable tax treatment.

And boasting 150,000 towers globally, with 42,000 in North America in particular, complimented by 860 distributed antenna systems and 30 data centers, AMT’s infrastructure footprint is nearly incomprehensible.

And these towers are vital because every time you require wireless connectivity, whether you’re sending a text, watching a YouTube video (preferably of us pitching you a stock!), or having a video call, you’re using a service that AMT provides.

Customers of AMT include the three US wireless carriers AT&T, T-Mobile, and Verizon, and in Europe, there’s Telefónica S.A. And in Africa and the Asia Pacific region, you have Bharti Airtel and MTN Group. As you might imagine, these carriers are large businesses with massive customer bases, generating billions in revenues.

On the data center side of things, through an acquisition AMT made in CoreSite, they have all the big hyperscalers as customers, including Alibaba Cloud, Amazon Web Services, Microsoft Azure, and Lumen.

But let’s focus here on the tower business, because that’s the segment generating the lion’s share of AMT's revenue. And it’s the segment with incredible unit economics that I teased earlier.

AMT’s cell towers are special because they generate incremental revenue at ever-higher margins. When AMT builds a tower for, say, ~$275k, they then sign a tenant (one of the major cellular providers) that’ll pay them a rate of something like $20k+/year. But it only costs AMT ~$12k to operate the tower, so they’re making 40% margins to begin with.

But the real magic happens when they add more tenants (again, think cell providers) to the same tower. With three tenants on one tower, they might earn as much as $80k in total revenues, yet it only costs AMT about $2k to add those two new tenants to a tower.

Meaning, with three tenants, AMT earns gross margins north of 80%! There aren’t many businesses in the world that can add $60k of revenue with only $2k of incremental investment — that’s one heck of a return on capital.

Many of AMT’s towers are in the middle of farms and large fields. But they also have towers on top of office buildings, comprising what’s referred to as distributed antenna systems, which provide wireless access in indoor and outdoor areas not served by traditional cell towers.

Thanks to the unceasing demand for mobile internet access, AMT has grown nicely over the last two decades on both the top and bottom lines. But what is even more important than that growth are the moats around the business, allowing AMT to maintain robust margins despite a changing competitive landscape over time.

I see three distinct moats underlying this impressive growth.

The first is in what’s known as cornered resources. What this means is that, even if building new towers is technically easy with a couple hundred grand, the problem lies in convincing a property owner or neighborhood to allow a 200-foot behemoth on their property. NIMBY, as they say.

On top of having a bureaucratic advantage over new competitors, AMT also has some of the best real estate in the country for cell towers. And network topology is vital. You can’t close your eyes and put a point on a map and say this is where you’re going to put a tower — it has to make sense to put it there in relation to where your other towers have coverage.

In my past life, I actually built networks inside customers’ homes. I’d lay out the network, then have my colleague physically install the router, wire it to all areas of the house, and test the coverage to make sure there were no dead zones. And even inside a house, I would have to get him to move the hardware to make sure the signal worked throughout.

And with cell towers, you cannot simply pick up a tower and move it somewhere else. Once it’s installed, it needs to work seamlessly.

Since AMT has such strong tower economics, they naturally leverage economies of scale, hence why gross margins climbed from 68% to 74% over the last decade, driven by their ability to add tenants at lower incremental costs. Again, for a dense physical-asset business, those are some objectively breathtaking margins.

We don’t actually know how many of their towers have one, two, or three tenants. They used to disclose this, but stopped in 2017. But the numbers support the view that AMT is experiencing significant operating leverage due to economies of scale.

Lastly, we have switching costs. For a business like AT&T, Verizon, or T-Mobile, paying AMT for a tower is a rounding error at $20k-$30k. Additionally, AMT has been around for decades, so its customers know it and trust it, which is an important consideration given how critical these towers are to their own businesses.

If customers did want to switch anyways, they’d have to wait until the contract is up, then decommission their equipment from the tower, move the equipment to a different tower, and set it up again. You may be thinking, “Well, waiting for the contract to end can’t be that big of a deal.” But I assure you, it kind of is.

AMT has $54 billion in non-cancellable leases. $54 billion! And these leases are very long-term, with terms of 5 to 10 years. And the churn rate supports the view that AMT has switching-cost moats, with churn averaging 2% historically.

One of the best aspects of these leases is their structure. The land that AMT leases to its customers is generally leased from someone else. AMT will then pay that person a monthly fee, or a one-time perpetual lease payment.

But what’s most interesting is how the leases are structured for AMT’s customers. In the US, for instance, where the currency is among the most stable in the world, they use an annual fee escalator of about 3%. This gives them pricing power in perpetuity, enabling revenue growth even without tower expansion.

And their pricing power is even stronger in international markets. In fact, in all of AMT’s markets, they tie price escalations to local inflation benchmarks, so even if a country is experiencing higher inflation, AMT will enjoy offsetting increases in tenant’s lease payments.

AMT has a few competitors worldwide, but to give you a simpler view, we’ll focus on the opposition in its primary market, the US. This refers to, namely, Crown Castle and SBA Communications. Because these competitors are also deeply entrenched, their competitive picture most closely resembles an oligopoly, where there’s a degree of tacit collusion (meaning that, by implicitly agreeing not to compete too directly with each other, they can all enjoy excess profits.)

New entrants, though, will continue to have a hard time entering the market for the reasons I’ve outlined. And I think we may continue to see some market share donors, not necessarily from losing business to the oligopoly, but from industry consolidation.

To really compete with a business like AMT, you’d need billions of dollars to build new towers, and decades of time to gradually add them to your portfolio, get the zoning and bylaw consent, and establish relationships with the largest cellular providers. It just doesn’t seem like an industry that’s particularly attractive to enter, despite the rewards provided to existing players.

When I began analyzing AMT’s debt, I had to admit I was pretty put off by it. But the more I dug in, the more my mind was somewhat put at ease. I think AMT has hummed along well, growing revenue 11% annually over the last two decades, but during this time, the business has slowly crept up in its net debt-to-EBITDA ratio (meaning more debt relative to operating income.)

Generally, the big spikes in this ratio occur when AMT makes an acquisition. For instance, in 2021, they spent about $20 billion dollars buying a European-based tower company called Telxius and a US Data Centre biz called CoreSite. But even if you exclude these acquisitions, you can see that the general trend in greater debt reliance points upward.

As of now, they have $37 billion in debt with $6.8 billion in trailing 12-month EBITDA. While the debt number is high, what matters is examining the debt structure in detail.

The first detail is that AMT’s debt matures over a long period, with a few notes extending as far as 2051. The next consideration is the average interest on this debt. It’s quite low compared to many businesses I’ve looked at, with an average interest rate of just 3.5%.

Part of the reason that AMT needs so much debt is simply that, as I wrote earlier, they have to pay about 90% of their taxable income to their shareholders. As a result, they cannot reinvest much profit back into the business. And for a business like AMT, which has grown its tower count substantially over the years through M&A, growth has necessarily been funded with debt.

AMT has historically created value because its ROIC exceeds its cost of capital. ROIC has been as dependable as the sunrise, with the number never dropping below 8% or rising about 11% over the last decade. One crucial area to evaluate in AMT is its capital allocation strategy.

Since much of its growth comes from buying up competitors, we should evaluate how they are doing in that regard. Instead of looking at some of the tower businesses they’ve acquired, though, I actually want to look at a business that they bought that was out of their wheelhouse in CoreSite, which comprises the data center wing of AMT.

The CoreSite deal is telling because it shows that, if AMT wants to continue growing in the high double digits, they need to expand beyond towers alone. But let’s get to the deal. In 2021, AMT paid $10.4 billion dollars for 24 data centers with an EBITDA multiple on the deal of 27x. While the business does have recurring revenue, and serves many of the AI hyperscalers, I wasn’t impressed with the multiple paid in the acquisition.

Since they’ve bought CoreSite, the unit’s revenue has compounded at 8% annually. In my view, that’s a low number, but on the plus side, as the business has compounded, CoreSite’s operating margins have surged from 46% to 53%, which is commendable.

To further evaluate its performance, I should mention that CoreSite has two decent public comps: Equinix and Digital Realty Trust. CoreSite is improving margins faster, but they’re all producing profits, highlighting that it is possible to make generate returns from some of these AI data centers, which are attracting billions of dollars in investment.

As the business has steadily grown its net income, management has provided AMT shareholders with a growing dividend. See below:

One potential area of concern, however, is that key metric for REITs, annual funds from operations (AFFO) has decreased by 12% since 2021. Yet the dividend is continuing to go up. So the debt they’re raising is clearly helping pay the dividend, which I’m not crazy about. At the same time, if dividends are cut, that poses a risk to the valuation multiple investors are willing to pay to acquire shares. Having to cut the dividend, then, with the company’s P/E multiple contracting correspondingly, would be a realistic bear case to consider here.

When evaluating AMT’s management, I feel as though they have a pretty good incentive structure, plus a solid track record of creating shareholder value. But one area I found disappointing was insider ownership, with insiders owning just 0.7% of shares. (I generally like to see insider ownership above 5%.)

Still, the incentive structure is an improvement. The long-term incentive plan accounts for total shareholder return with a 30% weighting (great to see!), while the other two weighted KPIs are based on average returns on capital (30% weighting) and cumulative annual funds from operations (AFFO) per share (40% weighting), which I think are excellent incentives, since ultimately, these KPIs will drive shareholder value over the long-term.

I like this a lot because it’s rare to see a business with capital efficiency and per-share metrics that create shareholder value in its incentive plan.

While AMT does have a wide moat, it doesn’t come without risks. The first risk I’d highlight is the concentration risk. AMT currently gets the lion’s share of its revenue from a few big telecom players worldwide. 59% of revenue comes from only four customers…

I’m not as worried about their customers jumping ship, though. It hasn’t happened very often, but what I find more worrisome is the consolidation of its customer base. Let’s imagine that two of those large customers were to consolidate into a single entity. If together they contributed 30% of AMT’s revenues, the merged companies might decide to remove as many as half their towers for redundancy purposes, cutting this chunk of revenue in half for AMT.

And this has happened before: when T-Mobile absorbed Sprint’s network, churn was significantly elevated.

Arguably, the most interesting risk, though, is how satellite-based internet will disrupt AMT’s business model. As of today, cell towers remain the better solution. Speeds range from 200 megabits per second (Mbps) to 1 gigabit per second (Gbps). Satellite speeds, however, range between 50 and 250 Mbps. Cellular towers also have much lower latency, simply because the signal has a lesser distance to travel.

So, cellular towers currently have a massive advantage over satellites, except for in areas with lower population density. If you live in a rural area area, you may have no choice but to accept the slower speeds of satellite-based internet, but it’s still better than nothing.

If companies like SpaceX and ASTS can continue improving speed and reducing latency through technological advances, perhaps it will become a real threat, but I don’t know how far they can take this technology, given the distance between a satellite and Earth.

As of now, the biggest risk I see is industry consolidation. Luckily for AMT, telecom businesses and governments tend to avoid creating additional monopolies. So I think the chances of the big three in the US consolidating, for instance, aren’t very high, and even if a deal were announced, there’s a real chance regulators would not allow it to proceed.

It’s evident that AMT is a wide-moat business that is very likely to exist in 10, 20, or more years. It’s deeply entrenched in its customers’ products and services, doesn’t have much customer churn, and there isn’t a viable competing product that offers the same performance.

When evaluating the value of AMT, a few key points helped me. The first is in revenue growth. AMT is a mature business that will grow revenue in the mid-single digits, thanks to annual fee escalators as well as modestly increasing the tenant count on their existing towers. In the base case, I conservatively assume a revenue growth of about 5%.

I also assume the business slightly expands its EBITDA margin from 65.4% today to 66% over the next five years. I figure they’ll continue getting some operating leverage as they increase their tenant count per tower. Lastly, I apply an EV/EBITDA multiple of 21 times.

The multiple is actually quite close to their 10-year low. But given the slowing growth and increasing leverage, I have a hard time believing the market will give it as high a multiple as it has in the past when it was growing faster, had less leverage, and didn’t have the specter of satellite-based internet as a threat.

In my bear case, I assume revenue growth continues to decelerate to about 3% (in line with their inflation-based price escalators for tenants.)

As for margins, I assume the margins come down to 62%. I imagine that satellite-based internet continues to become more affordable, eroding some of AMT’s pricing power. And so, I apply an 18x EV/EBITDA multiple in the bear case. Even though the business isn’t growing much, it would still likely have very good margins and robust recurring revenues, justifying a relatively high valuation for a bear case.

If you asked me to estimate the probabilities of each of these scenarios, I’d assign the base, bear, and bull cases probabilities of 60%, 25%, and 15%, respectively (but this is arbitrary!) AMT is one of these businesses where you probably won’t lose much money, even in the bear case, thanks to the strength of its business model.

But it also doesn’t carry much upside. I can’t see a scenario in which this business ever returns to its previous growth rates, and therefore, a double-digit top-line growth seems next to impossible. As such, I’ve recommended to Shawn and Daniel that we pass on AMT at current prices, but you can see my model below to determine potentially attractive entry prices in different scenarios:

And for more on AMT, you can listen to my podcast with Shawn about the company here.

More updates on our Intrinsic Value Portfolio below 👇

"A moat is an intrinsic characteristic that gives the business a durable competitive advantage."

— Charlie Munger

📺 WATCH: The Trillion Dollar Gap, Aswath Domodaran’s thoughts on SpaceX, AI, and the market in general

🎧 LISTEN: Acquired’s podcast on Vanguard, and the rise of index funds

📖 READ: Guy Spier and I, Vitaliy Katsenelson’s beautiful blog article on Guy Spier

You can also read our archive of past Intrinsic Value breakdowns, in case you’ve missed any, here — we’ve covered companies ranging from Alphabet to FICO, Transdigm, Lululemon, PayPal, DoorDash, Crocs, LVMH, Uber, and more!

Enjoy reading this newsletter? Forward it to a friend.

Was this newsletter forwarded to you? Sign up here.

Apply to join our Intrinsic Value Community of investors.

Kyle, Shawn & Daniel use Fiscal.ai for every company they research — use their referral link to get started with a 15% discount!

Read our full archive of Intrinsic Value Breakdowns here.

Keep an eye on your inbox for our newsletters on Sundays. If you have any feedback for us, simply respond to this email or message [email protected].