Report contents
Industry landscape
Where the company competes: a licensed three-carrier national market
AT&T sells connectivity. Since the April 2022 spin-off of WarnerMedia, its revenue comes mainly from wireless service, supplemented by fiber and copper broadband and by networking sold to business customers. That matters for defining the arena: AT&T is no longer a conglomerate whose pieces could be valued separately. It is a network operator whose economics depend on how many subscribers share the cost of one radio network and one fiber plant.
The core market is US retail wireless service. The second arena is US residential fiber broadband, which AT&T sells under the AT&T Internet name and bundles with wireless. Both are regulated by the FCC, which licenses the spectrum that carries the traffic (the radio frequencies a carrier needs, allocated by government auction rather than sold in an open market).
Upstream, downstream and who keeps the margin
Upstream, bargaining power is uneven by link. Spectrum is scarce and allocated by the regulator, so supply is set by auction design, not by demand. Radio equipment comes from a short list of vendors, with Ericsson, Nokia and Samsung the main suppliers of 5G radio gear; that side is concentrated, but AT&T's purchasing volume gives it countervailing weight. Tower sites are controlled by three large REITs, American Tower, Crown Castle and SBA Communications, which own or manage most US macro sites; long-term leases with contractual escalators hand some pricing power to the landlord.
Downstream, buyers are fragmented: tens of millions of households and a long tail of businesses, none with leverage over price. The exception is wholesale. Cable operators resell wireless capacity bought from the national carriers, which makes Comcast and Charter customers and competitors at once.
The network owner keeps the margin. Equipment vendors earn modest margins on hardware, tower landlords collect rent, and AT&T collects monthly service revenue and retains what survives depreciation on a very large asset base and interest on the debt that financed it. That is why subscriber counts matter more here than in most consumer businesses: one more household on an existing cell site or fiber strand is close to pure contribution, while the next million households in a new build require capital before they produce a dollar.
National radio, street-level fiber
Wireless competition is national. Licenses cover market areas, handsets work anywhere, and national advertising and retail distribution carry the brand. A carrier cannot be regional and match national pricing for long.
Fiber is the opposite. Competition is street by street, and the decisive threshold is homes passed per mile of new cable: below a certain density, the build does not pay back. AT&T's fiber footprint is therefore a set of local franchises facing one incumbent cable operator in most markets, which is a duopoly at the address level and an oligopoly at the national level.
| Company | Wireless model | Fixed broadband overlap |
|---|---|---|
| AT&T | Facilities-based national network | Fiber and copper |
| Verizon | Facilities-based national network | Fiber and fixed wireless |
| T-Mobile | Facilities-based national network | Fixed wireless only |
| Comcast | Wholesale access, no own network | Cable |
| Charter | Wholesale access, no own network | Cable |
Market structure: three networks, five brands, high walls
Three facilities-based national networks serve the US. The 2020 merger of T-Mobile and Sprint, approved after antitrust review, reduced the count from four to three, and regulators have since signaled little appetite for further consolidation among the big carriers. On the practical test that a market with more than a handful of meaningful competitors usually has weak entry barriers, wireless passes: the leaders fit on one hand.
Entry barriers are high, and they are capital and licenses, not customer captivity. Building a competitive national network takes years of site acquisition, spectrum at auction and tens of billions in spending before the first subscriber signs up. The barrier protects the network; it does not protect the customer relationship. Service plans are close substitutes, handset financing locks a customer in for only a couple of years, and churn is a live constraint on pricing every quarter.
Retail entry, however, has become cheap. An MVNO (a reseller that buys wholesale capacity and rebrands it) can rent a national network and undercut on price. Cable has done exactly that, turning itself into a wireless competitor without building a single cell site. So the moat sits at the network layer while competitive pressure arrives at the retail layer.
For AT&T, that structure cuts both ways. Three national networks have generally supported rational pricing and shared gains when demand grows; the same three have also competed away those gains in price wars. The open question is whether fiber and cable converge into stable local duopolies that lift broadband pricing, or whether fixed wireless from T-Mobile and Verizon, sold at a discount, caps what AT&T can charge for fiber. The evidence so far does not settle it, and the answer decides how much of the network's cost advantage reaches shareholders.
Competitive advantage
Does the company have a moat? Yes, but it protects the network, not the returns
Moat rating: Moderate. AT&T owns one of three facilities-based national wireless networks in the United States and is stringing fiber past millions of homes. Building a fourth national network is close to unaffordable: spectrum is allocated by government auction, and a competitive build runs at something like $20 billion of capex a year for a decade. That barrier is real, and it sits at the asset layer. It does not stop the retail layer above it from giving away most of the profit. AT&T's returns show the split. Estimated after-tax return on invested capital was 9.2% in FY2025 and 6.2% in FY2024, above the 6%-8% band that usually signals no advantage at all and well below the 15%-25% that marks a durable one.
Start with the ratio that flatters. AT&T reported a 20.4% return on equity in FY2025, which reads like a protected franchise. Equity is 26% of assets ($110.5 billion against $420.2 billion), so that number measures a leveraged balance sheet as much as a business. ROIC puts debt in the denominator alongside equity, which is the right test when borrowings total $136.1 billion. Interest expense of $6.8 billion in FY2025 works out to about 5% on that debt, a running cost the network has to cover before equity earns anything.
A three-carrier market ought to behave like an oligopoly. High fixed costs, a handful of owners and a product most households buy every month generally support pricing discipline, and that structure has held since the 2020 T-Mobile/Sprint merger. What breaks the discipline is interchangeability. Coverage is broadly comparable across the three networks, number portability makes switching free, and a customer can walk into a rival's store and leave with a subsidized handset the same afternoon. Revenue went from $120.74 billion in FY2022 to $125.65 billion in FY2025, about 1.3% a year. Operating income has not trended: $24.77 billion, $19.05 billion and $24.16 billion in FY2023, FY2024 and FY2025. One arithmetic caution on the last of those: EBIT of $33.81 billion exceeded operating income by $9.65 billion, which points to earnings booked outside the operating line, so the return figures below rest on operating income.
Market position: one of three national network owners
Market position: Fairly Strong. Three facilities-based networks serve the US market, and AT&T is one of them. Scale shows in two places: the revenue base described above and $154.2 billion of net property, plant and equipment, the physical plant a rival would have to duplicate. The three are not ranked here by subscribers. The rating rests on structure: a top-three slot in a market where the number of credible national owners is fixed by licenses and capital. That is protection against new entry, not the protection a 40%-share leader with no close rival enjoys.
Market position stability: held, but unproven
Stability: Moderate. Share stability over five years cannot be confirmed from a dated series. What the revenue line shows is a position held inside a narrow band: $120.74 billion in FY2022, $122.43 billion in FY2023, $122.34 billion in FY2024 and $125.65 billion in FY2025, a span of about 4% across four years. A company holding share in a fast-growing market and a company ceding share slowly produce the same picture. The sub-rating stays at Moderate for that reason: no evidence of erosion, no evidence of gains.
Returns on capital: close to the cost of capital
Returns on capital: Moderate. The financial statements run from FY2021 through the first nine months of FY2026. The estimate below uses the three most recent full years, because the FY2021 and FY2022 balance sheets still carry the WarnerMedia assets spun off in April 2022 and are not comparable.
| USD bn | FY2023 | FY2024 | FY2025 |
|---|---|---|---|
| Operating income | 24.77 | 19.05 | 24.16 |
| Effective tax rate (%) | 21.3 | 26.6 | 13.4 |
| NOPAT (est.) | 19.49 | 13.98 | 20.93 |
| Invested capital (est.) | 233.9 | 224.6 | 228.4 |
| After-tax ROIC (est., %) | 8.3 | 6.2 | 9.2 |
NOPAT is operating income times one minus the tax rate; invested capital is equity plus total debt less cash. The three-year average of 7.9% sits at the top of the range where an industry typically shows no protected returns. It is not zero, and the barriers behind it are demonstrably high, which is why this dimension rates Moderate instead of Weak.
Margins: high at the EBITDA line, ordinary after the capital bill
Margins: Moderate. Gross margin was 59.6% in FY2025 and 59.8% in FY2024; the 43.7% shown for FY2023 reflects a different split between cost of revenue and operating expenses, so it is not comparable. An EBITDA margin of 43.5% ($54.70 billion on $125.65 billion of revenue) looks like pricing power. It is mostly a capital-intensity artifact: depreciation on the network is a real cost that EBITDA excludes, and the capex that replaces the plant is larger still. Free cash flow was $19.44 billion in FY2025, 15.5% of revenue, which is the margin that survives the capital bill.
Overall rating: Moderate, with profitability setting the ceiling
Returns on capital averaging 7.9% over FY2023-FY2025 keep the profitability dimension at Moderate, and the overall rating follows it down even though market position is Fairly Strong. The barrier is genuine and expensive to cross; the rent it collects is thin.
Where the advantage comes from and how long it can last: spectrum and scale at the network, thin protection at the counter
For a facilities-based carrier the durable advantages are physical and regulatory. Brand counts for something in wireless, but a brand does not keep a competitor out; a license and a cell grid do. Two sources carry AT&T's position, economies of scale and spectrum licenses, and a third, customer captivity, is largely missing. The metric that tests all three at once is free cash flow against the capital tied up: $19.44 billion of free cash flow on roughly $228.4 billion of invested capital is 8.5%, which is what the moat is worth in FY2025.
Economies of scale: Fairly Strong, because the cost base is national and annual
The radio grid, the spectrum, the engineering staff and the national advertising are fixed costs that do not rise with an added subscriber. Until a cell site runs out of capacity, one more household on it is nearly pure contribution; after that, the carrier adds small cells and spectrum, which is why traffic growth eventually pulls capital spending up again. Minimum efficient scale is effectively the whole country, because a network that covers part of the country cannot sell a national plan to a family that travels or a business with offices in several states. Capex shows the size of the entry ticket: about $20.26 billion in FY2024, estimated as operating cash flow of $38.77 billion less free cash flow of $18.51 billion, equal to 16.6% of revenue. An entrant must spend at that rate for years before collecting a dollar. The reason rivals cannot win customers away by outspending AT&T is that coverage is already comparable, so extra capital buys capacity rather than advantage, and the contest moves to price and handset subsidy, where scale helps much less. This protection is structural rather than temporary, but it stops short of the customer. It breaks if wholesale access becomes cheap enough that a reseller undercuts on price while the network owner carries the capex: the plant then becomes a commodity input and the scale advantage accrues to the reseller's brand.
Spectrum licenses: Fairly Strong, and the barrier stops at the network edge
The FCC licenses the radio frequencies a carrier needs, allocating them by auction rather than selling them in an open market. Nobody assembles a national network without them, and the supply is set by auction design and by how much incumbents are willing to pay to defend their grids. The 2020 T-Mobile/Sprint merger cut the number of national owners from four to three, and regulators have since shown little appetite for further consolidation among the large carriers. That is a structural cap on the number of competitors, and it does not depend on management skill. What the license does not do is set a price. A rival's discount requires no new license, so the protection covers the right to operate and leaves the monthly bill exposed. It breaks if policy lowers the cost of entry, whether through shared spectrum, or through a cable operator assembling its own plant on rented capacity and owned small cells.
Customer captivity: Weak, since a handset plan binds for two years at most
Handset installment plans tie a customer to the network for 24 to 36 months, after which switching is a phone call: number portability is free and takes minutes, and no long-term contract stands in the way. Customers stay out of inertia or coverage quality, not because leaving costs them money. The numbers read like a business without pricing power. In the first nine months of FY2026 revenue rose 3.5% to $95.39 billion, but free cash flow fell 10.7% to $13.31 billion, and implied capex climbed to $15.64 billion from $14.06 billion. Growth is arriving through new plant, not through higher prices on the existing base. This is the weakest of the three sources and the reason the overall rating stops at Moderate. It breaks further if cable's wireless offers get good enough that a household takes mobile and broadband together from the cable operator, which resells capacity without carrying network capex.
Substitutes: fiber against fixed wireless decides the next decade
The substitute to watch is not another wire. T-Mobile and Verizon sell home broadband over the same cellular spectrum, which reaches a household without a fiber drop, and satellite covers addresses neither will build to. If fixed wireless is good enough for most households, the price of broadband gets capped by a cheaper alternative and AT&T's fiber build earns less than the capital committed to it. If fiber's speed and reliability hold the household, AT&T ends up in a local duopoly against the incumbent cable operator at each address, where a second wire is expensive to duplicate and local density protects the return. So far the evidence shows the spending, not the payoff: heavy capital going into fiber and free cash flow falling year over year. The durability of this moat therefore rests on a local economics question that the national figures cannot answer, and it will be settled address by address.
Stacked up, the sources rank the way the returns suggest. Captivity combined with scale is the strongest position a business can hold; scale alone is weaker because an entrant can rent the asset it would otherwise have to build. AT&T has the scale and lacks most of the captivity.
Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.
Continue reading the full report
- Comparison and summaryHow AT&T Inc. compares with its closest rivals, and our moat rating
Free to sign up · 20 research points · unlocking this report uses 10
20 points cover 2 full reports. No card needed.
Research reflects its generation date. Preview the first two modules.
More on AT&T Inc.
AT&T Inc.
Business, moat and key questions
Open company page