Sba Communications Corp SBAC

Thesis

SBA Communications is a wireless infrastructure REIT that owns 46,328 towers with an average of 1.8 tenants per site. The economic shape is a real-estate toll road: long-term leases (5-10 year initial terms with multiple tenant-option renewals) to T-Mobile, AT&T and Verizon, contractual annual escalators, and incremental tenants added at near-zero marginal cost. Top-three U.S. customers were 31.1% (T-Mobile), 20.3% (AT&T) and 15.1% (Verizon) of 2025 site-leasing revenue. The company controls land under ~71% of towers for >20 years, owns long-lived assets, and benefits from a multi-decade structural tailwind in mobile data consumption and 5G/AI-driven network densification.

The scorecard composite is 82 (profitability 22, balance sheet 17, capital allocation 19, valuation 24). ROIC 10y avg is 8.7% and ROIIC 5y is 43.7% — incremental capital is being deployed at very high marginal returns even though headline ROIC is dragged down by the goodwill/intangibles on the balance sheet from past tower acquisitions. The defining number is the gap between price and value: at $218.58 versus IV-low $233.62, IV-base $384.43 and IV-high $415.68, the stock is below even the conservative IV; price/IV of 0.5686 implies a meaningful margin of safety. The reverse-DCF implied growth of 6.86% is below SBA's actual organic leasing growth profile, suggesting the market is paying for very modest expectations.

The single offsetting fact is leverage: net debt / EBITDA of 8.14x. That is high in absolute terms but normal for a contracted, recurring-cash-flow tower REIT, and SBA's debt is largely tower-securitization paper at the asset level. At today's price, the math is: pay $218.58 for an asset I conservatively value at $384, accept the leverage, collect escalator-driven cash flow.

Moat

SBA's moat is a layered cost-advantage and switching-cost moat anchored in real estate.

Cost advantages (primary). A communications tower is a piece of regulated, location-specific real estate. Once it is permitted, zoned, built and electrified, replicating it next door is uneconomic — and usually disallowed by local zoning that favors co-location over new construction. SBA owns 46,328 towers and controls the land under ~71% for >20 years; the average remaining ground-lease life including SBA-controlled renewal options is 35 years. Three players (American Tower, Crown Castle, SBA) own the bulk of the U.S. macro-tower stock. A new entrant cannot un-build that history. This matches Damodaran's framing that economies of scale and exclusive distribution provide durable cost advantages against smaller rivals [6]. The Munger 'fortified position' is the same shape that makes See's Candy's regional dominance hard to dislodge [4]: the moat was built once and then defended for decades.

Switching costs (very strong at the site level). A carrier's antenna on an SBA tower is engineered into that carrier's RF coverage map. Moving the antenna to a competing tower 200m away is not a like-for-like substitution: it requires re-zoning, re-permitting, RF re-engineering, fiber backhaul re-provisioning, truck rolls, and a coverage hole during the transition. Carriers therefore renew. This is the same dynamic Damodaran describes for Microsoft Office: technical lock-in plus the cost of retraining/converting [5][6]. The MLA structure described in the 10-K (master lease agreements with site-specific terms, multiple renewal options at tenant election, contractual escalators) is the legal expression of that switching cost. Once a carrier is on an SBA tower with an MLA, the easiest commercial decision each year is to keep paying the escalator.

Network effects (modest but real). A tower with two tenants is ~3x as profitable per dollar of cost as one with one tenant — the 1.8-tenant-per-site average is a key margin lever. As the U.S. tower stock has consolidated into three owners, each owner's tower becomes more attractive to each carrier because national MLAs let carriers manage thousands of sites under one contract. This is a softer effect than a true two-sided network, but it does compound.

Intangibles (regulatory/zoning). Local zoning typically requires new antennas to co-locate on existing structures rather than build new towers. That regulation is not a legal monopoly granted to SBA, but it has the effect of legally entrenching the incumbent tower stock. Damodaran cautions that legal monopolies regulated by government can have value capped by price regulation [2]; SBA is the better version — the regulation restricts supply but does not cap pricing.

Pricing power. Tenant leases contain contractual annual escalators. SBA does not negotiate spot prices each year against carriers; the escalator is in the MLA. Pricing power is moderate-to-strong: not Coca-Cola brand pricing power [2], but a real, contracted, inflation-tracking lift.

Competitor stress test ($10B and 5 years). If a well-funded entrant tried to build a competing nationwide footprint, they would face: zoning denials, ground-lease scarcity (SBA already controls the best sites), carriers' reluctance to migrate equipment off existing towers without compensation, and the fact that carriers themselves have largely chosen to outsource tower ownership rather than build. $10B buys roughly 30-40k sites at replacement cost — but you cannot buy the locations or the tenant relationships at that price. Buffett's 1989 framing fits: the moat is not in the asset, it is in the position [Munger canon].

Erosion risk. The plausible erosion vectors are (a) carrier consolidation (Sprint/T-Mobile already happened; further consolidation removes a tenant) and (b) technology shift away from the macro-tower form factor toward small cells, satellite (Starlink direct-to-cell) or fiber-based densification. Neither erodes the tower base in a 5-year window, but both deserve underwriting weight.

Moat verdict: WIDE.

Management & Capital Allocation

Capital allocation at SBA is the single most interesting question in the analysis, because the headline ROIC of 8.7% and net-debt/EBITDA of 8.14x are both consequences of the chosen strategy.

Reinvestment. Management's stated primary use of cash is acquiring and building towers. The 10-K explicitly says they pursue acquisitions and new builds 'at prices that we believe will be accretive to our shareholders both in the short and long term and which allow us to maintain our long-term target leverage ratios.' During 2025 they purchased over 7,000 sites from Millicom in Central America and obtained a seven-year exclusivity right to build up to 2,500 build-to-suit sites with 15-year initial leases. This is the right kind of M&A: bolt-on, contracted-revenue assets in an existing line of business. Crucially, the scorecard's 5y ROIIC of 43.7% says recent reinvestment has been highly accretive. The dragging force on 10y ROIC (8.7%) is the cumulative goodwill on the balance sheet from a decade of tower acquisitions — the marginal dollar earns far more than the average historical dollar. That is the right pattern for a compounder: high incremental returns even if reported book ROIC is mediocre.

Acquisitions. They simultaneously exited the Philippines, Colombia and substantially all of Canada in 2025. That is unsentimental portfolio pruning — admitting that those markets do not meet the investment hurdle and redeploying capital into Central America. Buffett's framing applies: 'no rule that you have to invest money where you've earned it' [Buffett 2007 letter, canon]. Grade-positive.

Debt. Net debt / EBITDA of 8.14x is high. The defense is that tower cash flow is some of the most contracted, recurring revenue in the public markets, and SBA's debt is largely tower-securitization paper backed by specific tower assets at low spreads. Buffett's regulated/capital-intensive framing is relevant: BNSF and MidAmerican carry large debt 'not guaranteed by Berkshire' because the underlying earnings reliably cover interest [Buffett 2010, 2012 letters, canon]. SBA's balance sheet is in that family. That said, the scorecard does not give us interest coverage (null), so we cannot independently confirm coverage, and the 'balance sheet 17' subscore reflects real risk, not noise. In a higher-for-longer rate regime, refinancing tower-backed paper at materially wider spreads is the live risk.

Buybacks. Share count change over 10 years is -1.62%, i.e. a small net reduction. SBA has been a modest, opportunistic repurchaser rather than a heavy buyback story. Given the stock is now at price/IV-base of 0.5686, a Buffett-style allocator should be aggressive here. We do not have direct evidence of repurchase price discipline (avg P/IV when buying) in the excerpts. Mild concern: management has not used the current price gap to buy back stock at scale.

Dividends. SBA pays a dividend (REIT structure requires distribution of taxable income). The dividend is real but is constrained by the leverage and reinvestment-first strategy.

Communication. The 10-K is unusually plain-spoken for a REIT — they explain the unit economics (1.8 tenants/site, escalators, ground-lease tenor), they openly admit which countries they are exiting, and they tie strategy to specific numerics (15-year initial leases on Millicom build-to-suits, 71% of land controlled >20 years). That earns trust.

Net assessment. The capital allocator is running the playbook of a contracted-cash-flow infrastructure operator: lever up against secure cash flows, reinvest at 40%+ incremental ROIIC, prune underperforming geographies, return modest cash via dividend. The two demerits are leverage and the lack of aggressive buybacks at today's discount. The merits are clean reinvestment math, willingness to exit, and honest disclosure.

Capital allocator: B.

Industry Structure

Porter's Five Forces on the U.S. macro-tower industry:

Threat of new entrants — LOW. The supply of buildable tower locations is constrained by zoning, NIMBY opposition, ground-lease scarcity and carriers' preference for co-location on existing structures. A greenfield entrant cannot replicate SBA's 46,328-tower footprint or its long-dated ground-lease control (~71% controlled >20 years). Even capital-rich entrants face local-permitting friction that does not yield to balance-sheet size. The credible threat is not 'a fourth tower co' but 'carriers building their own again,' and the industry has spent two decades training carriers out of that habit.

Bargaining power of buyers — MEDIUM-HIGH and concentrated. The Big 3 (T-Mobile 31.1%, AT&T 20.3%, Verizon 15.1%) account for ~66% of SBA's domestic site-leasing revenue. That is real concentration. The offsetting facts are (a) carriers cannot easily migrate equipment off an installed tower without coverage holes, re-permitting and RF re-engineering, (b) MLAs are multi-year and contain escalators, and (c) the same Big 3 are the customers of all three tower cos, so SBA's competitors do not undercut via price wars (the structure resembles a stable oligopoly facing a stable oligopsony). Carrier consolidation (Sprint/T-Mobile) was the single biggest churn event in tower history and was absorbed.

Bargaining power of suppliers — LOW. Suppliers are landowners (ground lessors) and contractors. Landowners are fragmented; SBA buys perpetual easements and long-dated leases when available, which extinguishes the supplier over time. Contractors are commoditized.

Threat of substitutes — MEDIUM and rising. Substitutes include (a) small cells and DAS for dense-urban capacity, (b) fiber/wireline for backhaul and some fixed-wireless use cases, (c) low-earth-orbit satellite (Starlink direct-to-cell) for rural coverage, and (d) Wi-Fi offload. None of these dispatches the macro-tower in a 5-year horizon for the bulk of mobile data carriage, but each shaves the edges. Starlink direct-to-cell in particular is a non-zero tail risk for the rural tower stock. The Munger-honest answer is: macro-towers carry the bulk of mobile data today and will for the next decade, but their share of the wireless capex pie is likely to decline at the margin.

Rivalry — LOW. Three rational owners, similar cost structures, similar customers, no incentive to price-war because each tower's tenants are mostly captive. Competition is for new acquisitions and build-to-suit awards, not for poaching tenants off each other's towers.

Value pool location and trajectory. Within the wireless value chain, value has migrated from carriers (where ARPU is flat and capex is heavy) to passive infrastructure owners (towers, fiber, data centers) over the last 20 years. That trajectory continues so long as data demand keeps outrunning spectrum efficiency gains. The Ericsson data referenced in the 10-K and the 5G/AI/edge narrative all point to continued high carrier capex on densification, of which SBA captures a contracted share via escalators and amendments.

Industry Verdict: Good.

Inversion (Bear Case)

I am playing a short-seller. The bull case is the consensus; here is what kills it.

The single event that kills this. A sustained higher-for-longer rate environment combined with a credible carrier defection. Concretely: imagine 5-year and 10-year Treasury rates at 5%+ for three more years, and T-Mobile (31.1% of U.S. site-leasing revenue) decides to dramatically slow new amendments and let some leases run to expiry as it shifts capacity to small cells and Starlink direct-to-cell. SBA's tower-securitization paper rolls at 200-300bps wider, the dividend gets squeezed by interest expense, AFFO per share goes flat-to-down for 2-3 years, and the equity re-rates from a 24x AFFO multiple to 14-16x. That is a -40% to -50% equity drawdown without any operating disaster — purely a multiple-and-cash-flow squeeze.

Why the moat is narrower than bulls think. Macro-towers are a moat against new tower entrants. They are not a moat against the carriers themselves choosing a different topology. 5G densification was supposed to require many more macro sites; in practice, much of the densification has gone to small cells and indoor DAS where towers are not the host. Starlink direct-to-cell is now in commercial trials with T-Mobile and is an existential challenge for the rural macro-tower stock — perhaps 20-30% of SBA's U.S. towers serve geographies where satellite-direct could substantively reduce carrier need. The bulls answer 'satellite cannot serve dense urban'; the bears answer 'satellite does not need to — it just needs to remove the rural floor under tower demand and force tower amendments to reprice on the next cycle.' Moat verdict bear-case: NARROW, not WIDE.

Why management is worse than it appears. Three demerits. (1) Net debt / EBITDA at 8.14x in a normalizing-rate environment is aggressive; the right move two years ago was to deleverage, and they did not. (2) Share count is down only 1.62% over a decade despite the stock now trading at 0.57 of base IV — a true Buffett allocator with the company's balance-sheet flexibility would be repurchasing aggressively at this level, and they are not. (3) The Millicom Central America acquisition (7,000 sites + 2,500 build-to-suit commitment) is heavy capital deployment into emerging-markets currency exposure at exactly the wrong time, when domestic stock buybacks would have higher risk-adjusted returns at the current price. Capital allocator real grade: B-/C+, not A.

What bulls are extrapolating that won't hold. Bulls extrapolate (a) that contractual escalators (typically ~3% domestically) will persist, (b) that the 1.8 tenants/site number trends to 2.0+, and (c) that the 5G amendment cycle has years to run. Bear-case pushback: (a) escalators are contractual on existing leases, but renewals at expiry can and will reprice if carriers have substitutes, (b) tenants/site has been roughly flat for years and Sprint/T-Mobile consolidation actually reduced it on overlapping sites, and (c) the 5G amendment cycle peaked in 2022-2023 and is now decelerating. Domestic organic leasing growth has been mid-single-digits, not the high-single-digits the IV-base implicitly assumes.

Valuation trap (multiple compression / regime change). SBA traded at a 10-year average P/E of 563.96 — the math says GAAP earnings have been compressed by depreciation on the tower stock, which is why P/E is meaningless here and you have to value on AFFO/owner-earnings. EV/FCF of 32.76x is the more honest multiple. The historical regime that justified 30x+ EV/FCF was zero-rate, infinite-duration-asset enthusiasm. In a 5% Treasury world, the right multiple for a contracted, levered, low-organic-growth infrastructure asset is closer to 18-22x EV/FCF. Apply 20x EV/FCF to SBA's owner earnings and the equity value is materially below today's price. The 'price/IV 0.57' framing assumes the IV model has the right discount rate and the right terminal multiple — change either, and the discount disappears.

Bear-case price target. If I am right — rates stay high, carrier capex shifts to small cells and satellite, leverage forces a refinancing repricing, multiple compresses to 18-20x EV/FCF — SBA owner earnings of ~$0.81B grow to ~$1.0B in three years, applied at 18x EV/FCF gives EV of ~$18B, less ~$13B net debt, gives equity of ~$5B, which is roughly $46 per share on ~108m shares. More charitably, applying 22x to $1.05B gives EV of $23B, equity of $10B, or $93 per share. If I am right, the stock could be worth $90 within 3 years — a roughly 60% drawdown from $218.58.

Lollapalooza Bias Check

Biases active in me right now as the analyst, in rough order of strength:

Anchoring (strong). The scorecard hands me an IV-base of $384.43 against a price of $218.58. That 0.5686 ratio is anchoring my recommendation toward 'Buy' before I have stress-tested the IV inputs. The honest move is to ask: would I underwrite SBA at $218 if the scorecard had given me an IV-base of $260? Probably 'Hold,' not 'Buy.' The discount is doing a lot of work in my conviction, and the discount is only as real as the IV model.

Authority / social proof (medium). Tower REITs are a beloved Buffett-adjacent compounder category — Crown Castle, American Tower and SBA have all featured in respected long-term-investor portfolios. I am tempted to give SBA the benefit of the doubt because 'this is the kind of business smart people own.' Munger's lesson is that this exact pattern is how authority bias bleeds into bad underwriting. I should evaluate SBA on its own numbers, not on the category's reputation.

Confirmation (medium). The 10-K is well-written and tells a coherent compounder story (escalators, multi-tenant economics, ground control). I noticed I read the bull paragraphs more carefully than the leverage and customer-concentration paragraphs. The inversion section was the antidote — I forced myself to take T-Mobile concentration and Starlink direct-to-cell seriously rather than waving them off.

Recency (medium). The 5G capex cycle has been a multi-year tailwind. I am at risk of extrapolating that tailwind into perpetuity rather than treating it as a finite cycle that has likely peaked.

Incentive bias on management (medium). I am evaluating capital allocation from public disclosures written by people whose comp depends on AFFO growth. The leverage choice is exactly what their incentive predicts. I should weight the balance-sheet score (17, the lowest of the four) more heavily than the headline composite of 82.

Commitment (low). I have no prior commitment to SBA in this analysis — this is a clean read.

Deprival super-reaction (low). Not strongly active here.

The lollapalooza interaction to watch is anchoring + authority + recency stacking together to push 'Buy with high conviction.' The discipline is to size the position for the leverage and the substitution-risk tail, not for the IV-base discount.

10-Year Outlook

Same fundamental business model in 10 years? Mostly yes. SBA will still own ~50,000 communications towers, lease them to wireless carriers under long-dated contracts with escalators, and clip recurring cash flow. The form of the antenna and the spectrum bands will change; the real-estate business of owning the location and the structure will not.

Customer base larger? Probably flat-to-slightly-larger by tenant count. The Big 3 U.S. carriers will likely still be the Big 3 (Sprint/T-Mobile already consolidated; further consolidation is plausible but not certain). Internationally, the Millicom transaction enlarges the Central American footprint. Tenants per site at 1.8 today could move to 1.9-2.1 with edge-computing, fixed-wireless and private-network tenants, but it could also stagnate.

Profit per customer higher? Yes, modestly. Contractual escalators (typically ~3% domestically) compound over 10 years to roughly 30-35% nominal revenue per existing lease, before any amendment revenue. The risk is that renewal repricing at expiry partially offsets the escalator math if substitutes have improved.

Moat wider? Roughly the same. The zoning and ground-lease moat is durable. The substitution risk from satellite direct-to-cell, small cells and (in some markets) carrier-owned infrastructure is meaningfully larger in 10 years than today.

Single biggest threat? Technology substitution at the topology level — specifically, a world where 30%+ of mobile traffic is carried by small cells, satellites or unlicensed-spectrum alternatives, such that the macro-tower's share of carrier capex is structurally lower. This is not extinction; it is multiple compression on the asset class.

Confidence call. The business model is highly predictable. The capital structure and the substitution path are the uncertainties. On balance I can underwrite a 10-year compounding hypothesis here with reasonable confidence — not the same confidence I would have for a Coca-Cola-style consumer monopoly, but materially higher than for a tech name where the unit of competition is changing.

CONFIDENCE: medium