SBA Communications (SBAC)

Business Overview

SBA is the smallest of the three American tower majors and the only one that treats six-times leverage as a permanent setting rather than a stage it is passing through. It owned 46,328 towers at 31 December 2025, of which 17,394 were domestic and 28,934 international. Read those two counts against the revenue split and the shape of the company appears: nearly two thirds of the towers sit outside the United States, and they earn 27% of site leasing revenue. Brazil alone is 13.6% of the total. Market capitalisation was roughly $21.8 billion on 10 June 2026, at $205.17 a share on 106.1 million shares.

One thing has to be said before any multiple is quoted. In early April 2026 Bloomberg and Reuters reported that SBA had received preliminary takeover interest from infrastructure funds and was working with advisers on its options. The shares rose sharply on the report. No transaction has been agreed, the company has not confirmed a process, and management spoke on the August call about buying its own stock back rather than about selling the company. But a price struck after that report is not a clean reading of what the market thinks the towers are worth.

Net debt to adjusted EBITDA was 6.4 times at 31 December 2025: net debt of $12,521 million against $1,944 million of annualised fourth-quarter adjusted EBITDA. In November 2025 SBA formally adopted a target of 6.0 to 7.0 times, so 6.4 sits in the middle of its own range rather than at a ceiling. American Tower runs 4.9 times. Crown Castle reported 6.3 times at June 2026 once the fibre sale proceeds were applied, inside a 6.0 to 6.5 target. So SBA and Crown Castle now sit close, and the difference is that SBA arrived there deliberately. The rating agencies have accepted it: S&P raised SBA to BBB during 2026, and the company issued its first $3.5 billion of unsecured investment-grade notes. Leverage here is a rate-sensitivity question, not a solvency one.

How Tower Economics Work

The escalator is a gross number. American tenant leases run five to ten years with renewal options at the tenant's election and step up by a fixed annual percentage written into the contract. SBA does not publish a single blended rate, but its own 2026 bridge does the work: domestic escalations of $51 to $52 million on a 2025 domestic base of $1,866 million, about 2.8%. That is what the contracts add before anybody leaves. The escalator guide works the mechanism through.

A fixed step is not inflation protection. Roughly 12.6% of SBA's international tenant leases carried fixed escalators at 31 December 2025; the rest track an inflation index or blend the two. Index-linked rent does hold its real value, which a fixed 3% does not once inflation runs above 3%. It also moves with something nobody can forecast, in Brazil, Chile, South Africa and elsewhere, and it is earned in currencies the shareholder does not spend. The 2026 outlook is struck at 5.10 Brazilian reais to the dollar; change that assumption and the international line changes with it.

Read both churn numbers, never one. SBA prints regular churn separately from churn caused by the T-Mobile and Sprint merger. Regular churn in FY2025 was 2.6% consolidated, 1.3% domestic and 7.1% international. Sprint churn was a further 2.3% consolidated and 3.0% domestic, on the same base. So the rent SBA actually lost was 4.9% at group level and 4.3% domestically, not 1.3%. Nothing is concealed and the split is a fair description of a normal year against a merger year, but a screen that ranks SBA's American book as tight on 1.3% has ranked it on a number the company does not claim as its whole experience. Crown Castle's 0.7% headline has the same construction. Both are annual figures, incidentally, where data centre operators quote churn per quarter. The churn guide lays the bases out side by side.

2026 is the year the block arrives. Carrier consolidation does not remove tenants smoothly; it removes them in lumps, years after the merger, as the acquirer switches off duplicate sites. SBA guides domestic churn of about $134 million for 2026: roughly $55 million of Sprint non-renewals, $56 million from EchoStar and about $22 million of regular churn. EchoStar, formerly DISH, told SBA in late 2025 that it was shutting its network business, defaulted in December and is disputing what it owes; SBA has taken all of its recurring revenue out of the outlook. Set that $134 million against $86 million of domestic escalations and new leasing on an $1,866 million base and the direction is plain. Domestic site leasing revenue is guided down to $1,813 to $1,827 million.

AFFO: read the multiple, never the level. FY2025 AFFO was $1,381.4 million, or $12.85 a share. That is more than American Tower's $10.76 and far more than Crown Castle's $4.36, and it means only that SBA has about 107 million shares against Crown Castle's 437 million. At $205.17 the multiple is roughly 16.0 times. Two cautions before ranking anyone on it. Nareit defines FFO; no standard-setter defines AFFO, so every filer draws its own line between maintenance capital, which is deducted, and growth capital, which is not. SBA's 2026 guidance puts non-discretionary cash capital expenditure at $65 to $75 million against $455 to $475 million of discretionary spend, so the great bulk of the capital budget sits outside the deduction by the company's own classification. And the level has been falling: $13.37 a share in 2024, $12.85 in 2025, guided to $11.95 to $12.40 for 2026. The AFFO for towers guide works through the deductions.

The escalator is booked before it is paid. Accounting spreads total contracted rent evenly across a lease, so a lease that steps up every year books more revenue early than the carrier has handed over. SBA strips that straight-line revenue back out on the way from FFO to AFFO, along with straight-line ground rent. Revenue growth is therefore the wrong line to judge a tower REIT on.

The ground under the tower is the cost SBA does not set. Tenant rent escalates on a contract SBA wrote. Ground rent escalates on a contract the landowner wrote, and comes up for renewal on the landowner's clock. About 71% of SBA's towers stand on land it owns, holds under a perpetual easement, or controls for more than 20 years; the average remaining ground lease life is 35 years, and 14.5% of towers have land interests maturing within ten. The company spent $48.9 million buying land and easements in 2025. That programme is not housekeeping. It is the sector's answer to the one significant cost in the model that the tower owner does not control.

What to Watch in the Financials

Domestic organic growth, built from its parts. Escalations plus new leases and amendments, less all churn. On the 2026 guide that arithmetic is negative domestically, and the group only grows because of acquisitions, currency and the international book. Do not read the consolidated revenue line and assume the American towers are growing.

Whether international churn rolls down. At 7.1% it runs more than five times the domestic regular rate, driven by carrier consolidation and bankruptcies across Latin America and Africa. Management expects it to moderate as the remaining customer negotiations close. That is the single line with the most improvement available in it.

Leverage against the band, and what it costs. 6.4 times inside 6.0 to 7.0. Above 7 times is where the primer's screen asks for an explanation. The more useful question is not covenant room but interest: net cash interest expense is guided at $490 to $498 million for 2026, roughly a quarter of adjusted EBITDA, and every refinancing resets part of it.

Where the cash goes. SBA spent $497.8 million on buybacks and $479.0 million on dividends in 2025, against $1,009.9 million of acquisitions. The $5.00 annual dividend takes about 41% of guided 2026 AFFO, low for a REIT, which leaves the buyback as the swing item. Management said in August it expects to be active in the second half.

Valuation Framework

On FY2025 AFFO of $12.85, SBA traded at about 16.0 times at 10 June 2026, the bottom of the tower band of roughly 15 to 22 times. On the 2026 guidance midpoint of about $12.18 the multiple is nearer 16.9 times, because AFFO per share is guided down. Part of the price is a takeover report rather than a tower.

What the discount is paying for is visible enough: the churn ahead, an international book that churns at 7.1% in currencies that have to survive the trip back to dollars, and 6.4 times leverage on cash flows that are long, contracted and slow to change. Long cash flows put most of the value a long way out, which is where a change in the discount rate does the most damage, and leverage amplifies it. Hold AFFO at the guided 2026 midpoint and move the multiple to the 15 times band floor and the share is worth about $183, some 11% below the June price, with nothing having happened to a single lease. That is a duration illustration, not a forecast.

Key Risks

Three tenants, and consolidation removes them in blocks. T-Mobile, AT&T and Verizon paid 88% of domestic site leasing revenue in the fourth quarter of 2025. Sprint churn is still arriving six years after that merger closed, with roughly $75 million of cash site leasing revenue still to go, and EchoStar has demonstrated the other way a tenant disappears. With three national carriers left there is one fewer merger available to do this again, and correspondingly more damage if one happens.

The EchoStar dispute is unresolved. $56 million of 2026 revenue has been removed from the outlook, the counterparty defaulted in December 2025 and disagrees with SBA about the size of the obligation, and recovery runs through litigation rather than through a lease.

International concentration. Nearly two thirds of the towers and 27% of site leasing revenue sit outside the United States, with Brazil the largest single market. That book carries the currency, the inflation indexation, the sovereign exposure and the 7.1% churn all at once.

Rate sensitivity at 6.4 times. The exposure arrives twice: through the interest bill as maturities roll, and, larger, through the multiple applied to a long-dated rent stream.

The takeover report cuts both ways. If a deal appears, the shares reprice to it. If interest lapses, the part of the price it supports goes with it, and what is left is a company whose domestic revenue is guided lower.

What the Screening Shows

Against the Infrastructure & Digital REITs Primer thresholds:

• AFFO: $12.85 per share for FY2025, guided $11.95 to $12.40 for 2026. The level is a share-count artefact; the direction is not.
• P/AFFO: about 16.0x trailing, 16.9x on 2026 guidance. Bottom of the tower band of roughly 15x to 22x.
• Leverage: net debt to adjusted EBITDA 6.4x against a 6.0x to 7.0x target. Highest of the three tower names; above 7x is where the screen asks for an explanation.
• Churn: 4.9% total for FY2025, of which 2.6% regular. Domestically, 4.3% total on 1.3% regular. Internationally, 7.1%.
• Scale: ~$21.8 billion market cap on 46,328 towers. The smallest of the three, and the most internationally weighted by tower count.

The multiple is at the bottom of the band and the reasons are on the page rather than hidden in it: a domestic revenue line guided down, a churn headline that needs its second number, and leverage that the rating agencies have accepted but the discount rate has not stopped charging for. The case for SBA is that international churn moderates and the Sprint and EchoStar rent finishes leaving while the escalators keep running. The case against is that the same sentence has been available for several years.