Cell tower REITs own the steel-and-concrete structures on which wireless carriers (T-Mobile, AT&T, Verizon) mount antennas and radios. Each tower can host multiple tenants — each additional carrier pays rent but costs the tower owner almost nothing in incremental expense. The US has roughly 158,500 purpose-built cellular towers and about 254,850 total macrocell sites as of end-2025, per the Wireless Infrastructure Association's 2025 annual report compiled by iGR. American Tower (AMT), Crown Castle (CCI), and SBA Communications (SBAC) collectively hold just under 70% of privately owned US sites. Uniti Group (UNIT) is a fiber infrastructure company that merged with Windstream in August 2025 and now operates ~240,000 route miles of fiber plus a consumer/business telecom operation; it is included here because some investors group it with telecom infrastructure REITs, but it is not a tower company.
The first tenant on a tower covers costs; each additional tenant drops 70–80% to profit (per Mordor Intelligence). est. Leases run 10–25 years with 3–4% annual escalators. est. The near-term headwind is DISH/EchoStar churn — 5–7% of the big three's US rental revenue — and moderating carrier capex after the initial 5G buildout. est. The structural tailwind is mobile data growth, network densification, and the question of whether AI edge inference at tower sites creates a new demand layer.
The product is vertical real estate. A cell tower is a structure — typically a steel monopole or lattice tower 100–300 feet tall — on which wireless carriers install antennas, radios, cables, and ground-level equipment shelters. The tower company owns the structure and the land (or holds a long-term ground lease from the landowner). Carriers pay monthly rent for the right to mount equipment at a specific height on the tower.
How money flows in: A carrier signs a master lease agreement (MLA) covering hundreds or thousands of towers, typically for 10–25 years with annual escalators of 3–4%. est. Each tenant on a tower pays roughly $1,500–$3,000+ per month depending on the market (Tower Genius; actual rents vary widely and are not publicly disclosed per-tower). est. AMT earned $2.74B in total revenue in Q1 2026, CCI earned $1.01B, and SBAC earned $703M — overwhelmingly from site leasing (93%+ of revenue for SBAC, per its Q1 2026 filing).
How money flows out: The main costs are (1) ground lease payments to landowners beneath the tower, (2) property taxes and insurance, (3) tower maintenance and power, and (4) corporate overhead. These costs are mostly fixed per tower — adding a second or third tenant requires only marginal structural reinforcement and permitting. Tower cash flow margins run 79–84%: SBAC's domestic margin was 84.3% in Q1 2026 (SBAC Q1 2026 filing); AMT's US domestic segment operating margin was roughly 84%. est.
Colocation economics: The first tenant on a tower covers costs. Each additional tenant generates 70–80% gross margins with minimal additional capital (Mordor Intelligence). est. The tenancy ratio — average tenants per tower — is the key operating metric. As it rises from 2 to 3 to 4 tenants, revenue roughly doubles or triples while costs barely move.
Uniti Group (UNIT) is different. UNIT is a fiber infrastructure company. After merging with Windstream (closed August 2025), it operates ~240,000 route miles of fiber, serves over a million consumer and business subscribers, and earns revenue from leasing dark fiber to carriers, enterprises, and data centers, plus consumer broadband service. Its economics are capex-heavy (Q1 2026 capex: $349M on $988M revenue) rather than the tower model's low-incremental-cost colocation.
Sources: AMT, CCI, SBAC Q1 2026 earnings releases; UNIT Q1 2026 earnings release; Mordor Intelligence US Telecom Towers report; Tower Genius for lease rate ranges.
Carrier MLAs lock in years of revenue. AT&T's MLA with Crown Castle alone covers 40,000 towers. These contracts contain built-in escalators of 3–4% per year, providing a floor of organic revenue growth independent of new leasing activity. est. SBAC's 2026 guidance includes $71–74M from escalations alone.
New lease/amendment activity is positive but moderate. SBAC reported ~$10M per quarter in incremental US new lease/amendment billings YoY in Q1 2026 (SBAC Q1 earnings call). CCI's Q1 2026 organic growth was 3.1% (3.6% excluding DISH and Sprint churn). AMT's Q1 2026 US property revenue is guided to decline about 3% at midpoint, with tenant billings growth of only 0.5%.
Densification is replacing greenfield builds. The WIA/iGR report notes carriers are "shifting from coverage-first 5G deployment toward densification and colocation rather than new greenfield builds." Manhattan alone recorded over 700 rooftop amendments in 2025. Demand shows up as additional tenants and equipment upgrades on existing towers, not as orders for new tower builds.
US tower market growth rate: Mordor Intelligence sizes the US telecom towers market at $7.34B in 2025, growing to $9.01B by 2031 — a CAGR of 3.38%. est.
Mobile data traffic continues doubling every 2–3 years, per industry reports, which eventually forces carriers to add radio equipment to towers. Each technology generation (4G → 5G → 5G Advanced → 6G) requires new antenna installations, often at different frequencies and tower heights, generating lease amendment revenue. est.
AI / edge compute at tower sites is an emerging demand driver with no contracted revenue yet. AMT CEO Steve Vondran cited the company's positioning as "its strongest in over a decade" driven by AI workloads. Crown Castle management referenced "edge compute opportunities" at their 40,000 sites. The WIA report notes tower sites are being considered as part of the "AI/edge computing stack," requiring power upgrades, fiber terminations, and equipment shelter modifications. No contracted AI-edge revenue at tower sites has been publicly reported by any tower REIT. est.
Fixed wireless access (FWA) has connected over 14 million US households (WIA/iGR, end-2025), providing incremental demand for tower capacity from T-Mobile and Verizon.
DISH Wireless (part of EchoStar) defaulted on its tower lease obligations in late 2025 after selling its spectrum to AT&T and SpaceX for $42.6B. DISH leases span roughly 25,000 macro sites and represent 5–7% of annual rental revenue for the big three tower companies. est. AMT's exposure is ~$210M/year (~4% of US property revenue). CCI terminated its DISH agreement in January 2026 and is seeking $3.5B+ in remaining lease payments. SBAC has also filed suit. The Brattle Group estimated $9B+ in total contracted tower revenue at risk. Tower companies would need rent increases of 5.7–10.7% above standard escalators to recover the losses, per the March 2026 Brattle/WIA study.
Sprint network consolidation is the other churn source. T-Mobile's shutdown of the legacy Sprint network removes a tenant from towers where Sprint and T-Mobile were both present. SBAC's 2026 guidance includes $55–56M in Sprint consolidation churn. This is a known, finite event that will largely work through the system by 2027.
Sources: SBAC Q1 2026 press release (ir.sbasite.com); CCI Q1 2026 earnings (MarketBeat); AMT Q1 2026 earnings (stockti.com); WIA/iGR 2025 report (wirelessestimator.com); Mordor Intelligence; Brattle Group/WIA DISH impact study (wirelessestimator.com).
The US has 158,500 purpose-built towers and 254,850 total macrocell sites (including rooftops and alternative structures) as of end-2025. Net new tower builds have slowed: only 3,700 purpose-built towers were added in 2025 (up from 154,800 in 2024). SBAC built 80 new towers in Q1 2026 in the US; internationally it built ~60 in Central America. New tower permitting takes 12–24 months. est.
What is scarce is specific locations, not towers in general. In dense urban corridors, carriers compete for positions on towers that provide coverage at the right spot and height. In these markets, a fourth or fifth tenant per structure is being sought (Mordor Intelligence). est. In suburban and rural areas, many towers have available capacity.
The binding constraint on new tower supply is permitting, zoning, and local opposition, not materials or construction. Since demand is primarily for colocation on existing towers, this bottleneck is less relevant than in markets like data center power. The workforce is shrinking: 342,350 workers in 2025, down 26,400 from 2024 (WIA/iGR). Skilled field workers (tower climbers, riggers, RF technicians) are aging out, and compressed contractor margins are accelerating attrition.
UNIT's fiber supply: Uniti operates ~240,000 route miles post-Windstream merger, targeting 3.5 million homes passed with fiber by end-2029. Q1 2026 capex was $349M on $988M revenue.
Sources: WIA/iGR 2025 report (wirelessestimator.com); AMT Wikipedia/Q1 2026 earnings; CCI fiber sale (wirelessestimator.com); SBAC Q1 2026 press release; UNIT Q1 2026 press release.
| Factor | Demand side | Supply side |
|---|---|---|
| Direction | Slowly rising — data traffic grows, but carrier capex is flat to down | Stable — very few new towers being built; demand met via colocation on existing structures |
| Pace | 3–4% organic growth from escalators + modest new leasing est. | ~2.4% net tower stock growth (3,700 added in 2025 on base of 154,800) |
| Hard signal | Tenancy ratio rose from 2.4 to 2.6 even as carrier capex hit 6-yr low in 2024 est. | 158,500 purpose-built towers with available colocation capacity in most markets |
| Key churn | DISH: 5–7% of US revenue at risk est.; Sprint: finite, mostly done by 2027 | N/A |
| Pricing power | Escalators of 3–4% are contractually locked in. est. New lease/amendment pricing described as stable. Switching costs are high (a carrier cannot easily move its antenna to a competing tower in the exact same coverage footprint). | |
| Net balance | Neither acutely short nor oversupplied. Revenue growth in the low-to-mid single digits, with DISH churn as a temporary drag and AI edge as a speculative upside. | |
Where UNIT is different: Fiber markets are more competitive — overbuild competition from cable, FWA, and satellite is increasing. UNIT's fiber revenue grew 15% YoY in Q1 2026 (pro forma), driven by consumer fiber subscriber additions (39,000 gross adds, a record). But capex is heavy, and the company guided a net loss of $400–450M for full year 2026.
| Ticker | Company | Sites / assets | Q1 2026 rev | Mkt cap | Total debt | EV/EBITDA | Key fact |
|---|---|---|---|---|---|---|---|
| AMT | American Tower | ~149,700 towers (42,200 US/Canada) | $2.74B | $84.9B | $45.1B | 19.3x | Most geographically diversified; sold India ($2.5B, 2024); Q1 AFFO $1.32B; 2026 AFFO/sh guide ~$10.90–11.07; div yield 3.8% |
| CCI | Crown Castle | ~40,000 US towers (pure-play post May 2026 fiber/small-cell sale) | $1.01B | $38.7B | $29.9B | 25.1x | US-only; sold fiber + small cells for $8.5B (paid $17.5B to build); ~$7B to repay debt, $1B buyback; 2026 AFFO guide ~$2.1B post-divestiture; div yield 4.8%; negative book value |
| SBAC | SBA Communications | ~46,400 towers (17,400 US, 29,000 int'l) | $703M | $20.8B | $15.4B | 19.7x | Highest int'l growth (32.6% YoY ex-FX in Q1); 6.6x leverage; AFFO/sh $3.03 (Q1), 2026 guide $11.93–12.38; div yield 2.5%; 80% tower cash flow margin; acquired 4,300 Millicom sites |
| UNIT | Uniti Group | ~240,000 fiber route miles (NOT a tower co.) | $988M | $2.7B | $11.1B | 9.8x | Merged with Windstream Aug 2025; no longer a REIT; no dividend; net loss guided ($400–450M) for 2026; 2026 rev guide $3.6–3.7B, adj. EBITDA $1.43–1.48B; fiber rev +15% YoY; $785M interest expense/yr |
Balance sheet context: AMT, CCI, and SBAC use debt to finance tower acquisitions backed by predictable, escalating cash flows — their leverage is a feature of the REIT structure (required to distribute 90%+ of taxable income as dividends, so growth is financed with debt). UNIT's $11.1B debt on a $2.7B market cap reflects the Windstream merger's legacy telecom obligations and heavy fiber-build capex, with $785M/year in interest expense on $1.4B EBITDA.
Cash return mechanics: AMT, CCI, and SBAC distribute cash as REIT dividends (yields of 2.5–4.8%). CCI's payout ratio exceeds 100% of earnings (funded from AFFO, which adds back depreciation). SBAC's payout ratio is ~41% of AFFO. UNIT suspended its dividend pre-merger and does not pay one.
Sources: yfinance live data (June 2, 2026); AMT Q1 2026 earnings (stockti.com); CCI Q1 2026 (MarketBeat); SBAC Q1 2026 (ir.sbasite.com); UNIT Q1 2026 (markets.financialcontent.com); CCI fiber sale (wirelessestimator.com).
What you pay per dollar of cash flow: AMT trades at ~$182/share on 2026 guided AFFO of ~$11.00/share — roughly 16.5x AFFO. CCI trades at ~$89/share; post-divestiture annualized AFFO of ~$2.1B on 436M shares = ~$4.82/share, so ~18.4x AFFO. SBAC trades at ~$197/share on guided AFFO of ~$12.15/share midpoint — roughly 16.2x AFFO. EV/EBITDA runs 19–25x across the group.
What you get for that price: A stream of contracted, escalating lease payments with 80%+ margins, backed by physical infrastructure with high switching costs. Revenue grows in the low-to-mid single digits organically. Dividends yield 2.5–4.8%. AFFO generation: AMT ~$4.6B/year, CCI ~$2.1B, SBAC ~$1.3B (per recent filings). Total debt: $45B (AMT), $30B (CCI), $15B (SBAC).
The arithmetic: If organic revenue growth is 3–4%/year from escalators plus modest new leasing, and DISH churn subtracts 5–7% of US revenue as a one-time hit spread over 2025–2027, then net revenue growth is roughly flat to slightly positive near-term, accelerating once churn rolls off. At 16–18x AFFO, the price embeds low-to-mid single-digit compounding.
What you pay: At ~$11.23/share with 243M shares, market cap is $2.7B. Enterprise value is $12.9B (including $11.1B debt). EV/EBITDA is ~9.8x. 2026 guidance is a net loss of $400–450M. Interest expense of $785M/year consumes over half of adjusted EBITDA ($1.43–1.48B guided). No dividend.
What you get: 240,000 route miles of fiber, a consumer broadband operation growing at 26% YoY in fiber revenue, and a target of 3.5M homes passed by 2029. The equity sits beneath $11.1B of debt. If the fiber buildout succeeds and EBITDA grows enough to cover interest and capex, the equity accretes value. If not, the debt stack leaves little room for error.
Sources: yfinance live quotes (June 2, 2026); AMT, SBAC, CCI 2026 AFFO guidance from Q1 2026 earnings releases; UNIT Q1 2026 earnings release.
Grounded in filings and primary sources:
Approximate / directional (flagged est.):