Learn
A little clarity for your next decision.
Loading your learning library…
Investor WorkspaceWelcome, there!Log Out
Investor WorkspaceA little clarity for your next decision.
By Jerry Baker
Cell tower REITs rent space and related infrastructure to wireless carriers and other communications customers. Their returns depend on customer payments, ground rights, added equipment, operating costs, debt, and the price paid for the shares. More mobile data can support demand, but it does not guarantee that every tower gains tenants or that dividends will grow.
Nareit includes towers, wireless infrastructure, and fiber within its telecommunications REIT category. That describes a broad group, not one standard business. A company can own some of these assets and none of the others. [1]
A tower supports equipment used to send and receive signals. The tower owner may own the land, lease it, or hold another right to use the site. Customers may own the antennas and radios. Read the ownership and service terms before deciding which party pays for each part.
A rooftop site, a large tower, and a small wireless node can have different costs and rights. Fiber routes have their own capacity, construction, and access issues. A data center adds another operating model. Do not treat the word infrastructure as proof that these businesses work alike.
Current facts matter. Crown Castle completed the sale of its fiber and small-cell businesses on May 1, 2026. Its second-quarter report treats the sold business separately from continuing operations. An old profile describing all three as current operating segments would misstate the company. [2]
I begin with a dated asset map and a revenue breakdown. Then I ask which assets produce today's cash and which still require spending.
One site can host several customers. Some costs may be shared across them, which can make an added customer valuable. But a new lease may still need equipment supports, stronger foundations, power work, or more ground space.
Consider an original example. A site collects $30,000 a year from its first tenant and costs $18,000 to operate. It produces $12,000 before company costs, debt, and capital work. A second tenant pays $24,000 and adds $4,000 of yearly costs. Site income becomes $32,000.
The added tenant contributes $20,000 before those other expenses. That is an attractive increase in this example, but it does not show the return on the money needed to secure the lease. If upgrades cost $100,000, the simple annual income-to-cost ratio is 20%. It ignores timing, taxes, financing, and future repairs.
Now suppose the upgrade costs $250,000. The same $20,000 equals 8%. The promise of another tenant is not enough; the construction estimate matters.
I also want to know whether the site has room for more equipment. Physical capacity, ground rights, customer demand, and permits can all limit growth. An empty position is not a signed lease.
A tenancy ratio usually describes the average number of tenants per site under the company's definition. Before comparing companies, check whether the count includes rooftop sites, inactive assets, partly owned sites, or different types of contracts.
Suppose 100 towers host 180 tenants. The average is 1.8. That could mean 80 towers have two tenants and 20 have one. It could also hide a smaller group of busy sites and many weak ones. The average does not show the distribution.
Ask for income by site and market. A site with one large customer may earn more than another with several small agreements. Counting tenants without rent or costs can miss the better asset.
Watch how purchases and sales change the ratio. Selling low-occupancy sites can raise the average even if no existing tower signs a new tenant. A new build may reduce the ratio at first while still having a sound long-term plan.
For each growth claim, I would ask whether it came from better use of existing assets, new investment, a changed reporting group, or a different definition.
A contract may raise base rent on a schedule. An amendment may add charges when a tenant installs more equipment. A new tenant can add another payment stream. These are different sources of growth.
American Tower's second-quarter 2026 release separates new tenant leases and equipment amendments, scheduled increases, cancellations, and new sites within its billing definitions. It also separates these items from accounting adjustments and certain pass-through revenue. Those definitions help explain why reported revenue and tenant billings can differ. [3]
Use a simple example: $1 million of annual base rent rises 3%, adding $30,000. New leases add $40,000. Lost leases remove $60,000. Before other changes, rent grows by $10,000, or 1%. Quoting the 3% escalator alone would leave out most of the story.
Read the actual increase formula. A fixed percentage, an inflation index, and a negotiated reset do not offer the same result. Caps, floors, timing, and exclusions can change the cash received.
I also check which entity owes the rent. A familiar carrier brand does not mean every agreement has the same parent guarantee. The signed lease and any guarantee define the claim.
Churn refers to lost revenue from cancellations, nonrenewals, or other lease changes under the company's definition. It can result from overlapping networks, changed plans, failed customers, or a tenant deciding that a location no longer meets its needs.
Crown Castle's second-quarter 2026 report showed $38 million of organic billing contribution before specified DISH and Sprint effects. Those effects were negative $49 million and $5 million. Including them changed the result to a $16 million decline. The excluded items matter to the owner's economics. [2]
A company may show an adjusted figure to explain ongoing leasing activity. That can be useful. I still want a clear bridge back to the actual result, with both numbers visible.
For a portfolio example, suppose one carrier supplies 25% of rent and stops using sites that represent 20% of its payments. The exposure is 5% of total rent before replacements, settlements, or other changes. Losing 5% of revenue can reduce profit by more than 5% when costs do not fall at the same pace.
Do not assume a long agreement removes all risk. Check termination rights, payment disputes, legal remedies, and the customer's ability to pay. A valid claim is not the same as collected cash.
Owning a tower does not always mean owning the soil beneath it. Review the ground lease, access rights, utility rights, renewal choices, and any purchase option. Losing access can threaten more than the rent paid to the landowner.
American Tower's June 2026 filing states that its schedule of future rental receipts assumes continued access to the sites and ground space. Its footnotes also distinguish contractual amounts from expected collectibility. A large backlog is not a risk-free bank balance. [4]
I would place the ground agreement and customer agreement on the same timeline. If the tenant can stay for 15 years but the ground right ends in eight, explain how the gap is resolved. A renewal option must be enforceable and affordable to provide useful protection.
Consider a site collecting $50,000 a year with $10,000 of ground rent and $5,000 of other costs. Income before debt and capital is $35,000. If ground rent resets to $18,000, income falls to $27,000, a decline of about 23%.
Buying the land can remove some renewal uncertainty, but the price still matters. Compare the purchase cost with avoided rent, future obligations, access protections, and other uses of that cash.
A useful tower needs more than a point on a map. Crews need lawful access, safe working conditions, and rights to bring power or other services to the site. Ask whether those rights survive a sale of neighboring land.
Structural capacity is a separate question. New equipment can change weight and wind loads. I would want an engineer's review and a funded work plan before counting a proposed addition as easy income.
Historic-preservation rules also have specific conditions. The FCC's collocation agreement provides defined exclusions from Section 106 review for certain antenna placements, with exceptions involving size changes, prior review, and historic effects. It is not a blanket exemption from every approval. [5]
Ask who obtains each permit and who bears delay costs. A projected lease start should line up with the work and approval schedule. A signed customer agreement can still depend on conditions.
For an original timing example, an addition expected to earn $3,000 a month is delayed four months. That shifts $12,000 of expected rent. Interest and site costs may continue during that period. The final effect depends on the contract and whether the lost months are recovered later.
Mobile data growth is a broad demand story. A tower owner earns money only when a customer needs its location and agrees to pay on terms that cover the costs.
Ask how the site fits into the customer's network. Is the need coverage across an area, capacity in a busy place, or a specific route? More demand in one location does not prove demand for a tower elsewhere.
Review whether an equipment upgrade requires an additional payment. The carrier's spending can be substantial without producing the same increase in rent for the owner. The agreement may already allow certain changes.
I would also test alternatives without pretending to predict the winning technology. Could a customer use another nearby site, add smaller nodes, share equipment, or change its network design? What would moving cost, and when can it do so?
The right question is how a technology change affects this contract and this asset. A presentation filled with network-generation labels does not replace a lease schedule, an engineering report, and evidence of customer demand.
A large number of towers can still depend on a small number of customers. FINRA's concentration guidance explains why holdings that share risks may provide less diversification than their count suggests. [6]
Measure revenue by carrier, parent company, country, and major agreement. Then check how much expires or resets in each year. Separate a single customer with many sites from many independent customers.
For example, owning 1,000 sites across ten states may reduce exposure to one local storm. It does not reduce a 40% dependence on one carrier's payments. Geography and customer credit answer different questions.
Also examine shared cost exposures. Many sites may use one maintenance provider, depend on similar backup equipment, or face the same insurance renewal. These links can make problems arrive together.
I would compare the holding with the rest of your portfolio. An investor who already owns a large telecom position may be adding a related business risk through the landlord. A different ticker symbol does not always create a different source of income.
Foreign operations can expand the customer base. They also add questions about currencies, taxes, land rights, local enforcement, and the ability to move cash to the parent company.
Use an original translation example. A site earns 1 million units of local currency. At ten units per dollar, that equals $100,000. At eleven units per dollar, the same local income equals about $90,909. Local results can stay flat while dollar results fall.
If local income rises 5% to 1.05 million units while the rate moves to eleven, the translated amount is about $95,455. The local increase does not fully offset the currency move. This example excludes taxes and hedges.
Match revenue, expenses, and debt by currency. Local borrowing can offset some exposure but creates its own interest and refinancing duties. A hedge may cover a limited amount and period rather than every future payment.
American Tower's second-quarter 2026 release distinguishes foreign-currency effects from operating results and identifies different regional businesses. I would use that kind of breakdown instead of treating consolidated dollar growth as a pure measure of new demand. [3]
A tower can need inspections, repairs, reinforcement, and replacement parts even when its tenant count is unchanged. A new site may need land rights, construction, power, and a long wait before rent starts.
I would separate work needed to keep current income from work intended to create new income. Then ask whether the company's chosen cash-flow measure deducts each category.
FFO reverses specified real estate accounting charges and makes other adjustments. AFFO definitions vary. Neither label should end the review of actual cash needs. Read the reconciliation and the capital budget together. [7] [8]
Suppose a portfolio produces $20 million after site operating costs. It pays $5 million of interest, $2 million of company expenses, and $3 million for needed repairs and ground-right payments. That leaves $10 million before taxes, principal, new construction, and other obligations.
If management wants to build $12 million of new sites, it needs another source of funds or must use less cash elsewhere. Growth may be attractive, but it still has to be financed.
Look at cash per share after that financing. A larger tower count can benefit existing investors, leave them unchanged, or dilute their results.
Long customer contracts do not make debt harmless. A lender can require repayment before those contracts end. Higher rates can reduce cash even if rent arrives as expected.
The OCC's refinancing guidance highlights the interaction of borrowing costs, property values, and repayment gaps. For an investor, the useful question is how much new money a stressed refinancing would require. [9]
Imagine $50 million of debt moving from 4% to 6%. Annual interest rises from $2 million to $3 million. That $1 million increase must be covered before any added cash reaches shareholders.
Now suppose a lender will advance 55% of an $80 million value. It offers $44 million, leaving a $6 million gap against the existing balance, before fees. A higher interest bill and a paydown requirement can occur together.
Valuation also matters without debt. If sustainable annual cash earnings are $4 per share, a $100 share price is 25 times earnings. A $120 price is 30 times the same amount. Strong assets can still disappoint an investor who pays too much.
I would review a lower-growth case and a higher-cost case together. Testing only one favorable assumption at a time can hide how the business behaves under pressure.
Even a site with strong contracts can have an unexpected repair. I would separate the cash needed immediately from insurance proceeds that might arrive later. Deductibles, exclusions, and claim timing belong in that estimate.
Suppose a storm creates $400,000 of repair and access costs. Management expects a valid policy claim to reimburse $250,000 after review. The expected final cost is $150,000, but the initial funding need can still be the full $400,000. A reserve sized only for the final cost could leave a gap.
Then check the customer contract. Is rent still due during the interruption? Are credits available? Who must restore power, the access road, and the equipment? Each duty needs an owner and a timeline.
This is also a useful test of management reporting. A clear response explains both the physical recovery plan and the cash plan, without treating expected reimbursements as money already in the bank.
The file should connect physical assets with contracts and cash. I want current ownership records, ground-right terms, customer concentration, lease expirations, and the cost to support proposed additions.
Then I want a bridge from prior-period rent to current-period rent. Show increases, amendments, new leases, cancellations, acquired assets, and accounting changes separately. State which figures exclude difficult items.
The next page should explain funding: debt dates, interest exposure, planned construction, repair needs, and dividends. Avoid counting undrawn borrowing capacity as cash already earned.
Finally, match the investment structure to the investor. Listed shares can be sold in a market but can fall sharply. Private and nontraded REITs may restrict sales or repurchases. The SEC warns that structure, fees, and liquidity need their own review. [10]
For 1031 planning, ordinary REIT shares are not direct replacement real property. Owning shares in a company that owns towers does not change that rule. Any different proposed structure needs separate tax review before exchange funds move. [11]
Generally, no. It provides real estate or infrastructure used by customers. The carrier's business and the landlord's business are separate. Check which equipment, land rights, and services the REIT actually owns or provides.
Some site costs may already be covered, so another tenant can add income without a matching rise in yearly costs. But upgrades and other spending can reduce the return. Review the total cost of securing the new lease.
No. A long contract can support predictable billing, but customer credit, disputes, costs, debt, and share prices still create risk. A REIT dividend is not a guaranteed payment backed by the government.
It is lost lease revenue under the company's definition, often from cancellations or nonrenewals. Check whether a reported growth figure includes it. An adjusted number can help explain leasing but should be reconciled to the full result.
A tower may stand on land owned by someone else. The right to remain, reach the site, and renew on workable terms affects its value. Compare the ground agreement's dates with the customer agreements.
No. Demand must lead to customer spending that benefits the particular sites and contracts. Changes in technology, network design, and carrier budgets can affect that link. Review actual leasing and cash results.
Ordinary REIT shares are not direct Section 1031 replacement real property. Ask your qualified intermediary and tax advisor to review the exact proposed investment before you commit exchange money. [11]
Educational information, not an offer or a personal tax, legal, or investment recommendation. Examples are hypothetical and omit stated adjustments. Tax treatment depends on your facts and current law. Review your transaction with your CPA, attorney, and qualified intermediary. Real estate investments can lose value and may be illiquid.
Enter your approved email to explore investments, save your shortlist, and keep learning.