Summary

  • China Tower's first-half 2026 operating revenue was RMB 48.69 billion: the tower business contributed RMB 35.26 billion, or 72.4 per cent; the distributed antenna system (DAS) business RMB 5.09 billion, or 10.5 per cent; and the Two Wings businesses RMB 7.92 billion, or 16.3 per cent.
  • The same record contains no segment cost, no segment result, no capital expenditure split and no pricing term in any agreement with the carrier customers, so it cannot show whether the Two Wings lines earn their cost of capital. The mix is disclosed; the return is not.
  • What the record does constrain is the cash statement: EBITDA of RMB 30.25 billion became operating cash flow of RMB 7.14 billion, a conversion of 23.6 per cent, while capital expenditure of RMB 11.65 billion ran at 1.63 times operating cash flow in the same half in which an interim dividend of RMB 0.19122 per share before tax was declared.

The half-year, as reported

The figures come from a retained transcript of China Tower's results call for the first half of 2026, which records operating revenue of RMB 48.69 billion and then splits it twice: once into a carrier-facing business and a non-carrier-facing group, and once again inside each of those groups. The carrier business, which the company calls TSP, reached RMB 40.36 billion. The tower business inside it amounted to RMB 35.26 billion and the DAS business to RMB 5.09 billion. The group the company calls Two Wings was RMB 7.92 billion, comprising Smart Tower at RMB 5.33 billion and energy at RMB 2.59 billion, with the energy line itself reported as RMB 1.6 billion of battery exchange and RMB 1.0 billion of power backup. The same record reports EBITDA of RMB 30.25 billion, operating cash flow of RMB 7.14 billion, capital expenditure of RMB 11.65 billion, 2.172 million sites, a tower tenancy ratio of 1.78 tenants per site, and a board decision to distribute an interim dividend of RMB 0.19122 per share before tax https://seekingalpha.com/article/4943666-china-tower-corporation-limited-ctowy-q2-2026-earnings-call-transcript.

Two retrievals of that transcript were made and they returned overlapping but not identical passages. The first carried the revenue splits, the cash figures and the dividend; the second carried the site count, the tenancy ratio and the EBITDA figure, and added the battery-exchange and power-backup split of the energy line. Neither retrieval contained segment cost, segment profit, a capital expenditure breakdown, a share count, net profit, borrowings, or any term of any service agreement.

That provenance matters for how much weight the numbers will bear: this is one publication's rendering of a company call, read in two instalments, not a reading of the filing.

An industry round-up supplies an independent second report of the same headline numbers. It gives revenue of RMB 48.693 billion and tower revenue of RMB 35.263 billion, which agree with the transcript to the third decimal place, and it reports 2.172 million sites and a tenancy ratio of 1.78x identically. It adds one figure the transcript did not carry, a tenant count of 3.871 million, and it states that China Mobile, China Telecom and China Unicom remain the company's major telecom customers https://telecomlead.com/telecom-equipment/top-telecom-tower-companies-in-2026-towers-revenue-investment-and-5g-infrastructure-127722. Two publishers reconciling on the same numbers is stronger than one; it is still two renderings of a company statement rather than an inspection of the statement.

What the record cannot say about growth

Neither retained source carries a prior-year or full-year comparative. That is not a small omission. It means the disclosed figures describe one half-year in isolation, and no growth rate, no year-on-year change and no trend can be asserted from them. A reader who wants to know whether the Two Wings lines are expanding faster than the tower base will not learn it here, and this article does not claim it. What can be tested is narrower and still useful: the composition of the reported revenue, the arithmetic consistency of the parts, and what the cash statement says about the capital being consumed in the same period.

That distinction is worth stating plainly because it is where most commentary on a tower company goes wrong. A revenue mix is a snapshot of where the money came from. A return is a statement about what was spent to get it. The first is in the record. The second requires a cost line and a capital line, and neither is.

Two groups, four lines, one residual

The segment architecture has to be stated exactly, because it is easy to misdescribe and misdescription changes the argument. TSP at RMB 40.36 billion is not a fifth business; it is the sum of the tower business and DAS. Two Wings at RMB 7.92 billion is the sum of Smart Tower and energy. DAS therefore sits inside the carrier-facing group alongside towers, not inside Two Wings. An account that files distributed antenna systems under the newer businesses would be describing a different company from the one in the retained source.

The arithmetic over those parts produces three results of different weight. First, the two TSP components sum to RMB 40.35 billion against a reported TSP figure of RMB 40.36 billion, a difference of RMB 0.01 billion, which is consistent with rounding and carries no further meaning. Second, the two Two Wings components sum to RMB 7.92 billion exactly, with no residual at all. Third, and this is the one to hold carefully, the two top-level groups sum to RMB 48.28 billion against reported operating revenue of RMB 48.69 billion, leaving RMB 0.41 billion, or 0.8 per cent of reported revenue, that the retained passages do not explain.

That residual is recorded here as a disclosure-precision observation and nothing more. It is not evidence of an error, a restatement or a misstatement, and the retained material gives no basis for saying which line it belongs to or whether it belongs to any of them. Its only analytical use is as a reminder of the resolution at which this record operates: the parts reconcile to within 0.8 per cent of the total, which is close enough to trust the shape of the mix and far enough from exact to discourage pretending the disclosure is complete.

Where the mix stops being free

Revenue composition on its own is a neutral fact. It becomes a question about returns only when it is set against the capital consumed to produce it, and that is what the cash statement in the same half does. EBITDA of RMB 30.25 billion is 62.1 per cent of the reported RMB 48.69 billion of revenue. Operating cash flow of RMB 7.14 billion is 23.6 per cent of that EBITDA. Capital expenditure of RMB 11.65 billion is 1.63 times operating cash flow, exceeding it by RMB 4.51 billion within the half, and equal to 23.9 per cent of revenue.

The gap between EBITDA and operating cash flow is the single most consequential number in the record and also the least explained. Nothing retained says how much of it is interest, how much is tax, how much is working capital and how much is operating cost sitting below the EBITDA line. The record contains no net profit, no attributable profit, no interest expense, no tax charge, no borrowing balance, no lease liability and no working-capital detail. So the gap can be measured and cannot be attributed. Anyone who tells a reader that the difference is the cost of the new businesses is guessing, and this article does not guess.

What can be said without guessing is the direction of the capital flow. In the half, the company spent RMB 11.65 billion of capital and generated RMB 7.14 billion of operating cash, and it declared an interim distribution of RMB 0.19122 per share before tax in the same period. The per-share amount is known; the total cash cost of the distribution is not, because no share count was obtained.

That is a real limit on the analysis and it is stated rather than papered over: the article can say that the distribution and the investment programme were not both covered by the operating cash reported for the half, and it cannot say by how much the distribution exceeded or fell short of any particular measure of free cash.

The three decisions that actually set the return

A tower company's return on an installed site is set by three separable decisions, and separating them is the whole of the analysis. The first is how many tenants share a site, and that is decided by the carrier customers, not by the tower company. The second is the price charged per tenancy, and that is set by an agreement with those same customers. The third is the return earned on incremental capital, and that one is management's.

In the retained record, the first appears only as an outcome: 2.172 million sites carrying 3.871 million tenants for a tenancy ratio of 1.78. Dividing the tenant count by the site count gives 1.7822, which reproduces the reported ratio and confirms the two figures are internally consistent rather than independently sourced. The second appears not at all: no pricing basis, no price formula, no contract term, no renewal mechanism, no volume commitment and no tenancy commitment in any agreement between the company and its carrier customers was obtained. The third is not visible either, because the cost side and the capital split are absent.

That leaves the article in a specific and defensible position. It can describe how many tenants share how many sites, and it can say that this ratio, not the revenue mix, is the variable that determines whether an installed tower earns more than it costs. It cannot say what the tenants pay, and it cannot say what the newer businesses earn, because the record it has does not contain either answer. The revenue mix is therefore consistent with a rising, flat or falling return on capital, and nothing in the retained material distinguishes among those three.

The agreement that sets the price is not in the record

The three carriers appear in the record in two roles, and the distinction matters. They are the company's major telecom customers, and they are also the operators that agreed to transfer tower assets into the tower joint venture. A retained news excerpt records that China Mobile, China Unicom and China Telecom announced a final agreement on the structure of their planned towers joint venture and that each was to transfer its tower assets into it https://www.telecompaper.com/news/china-mobile-operators-get-state-support-for-towers-venture--1107794.

That excerpt is truncated by the provider at the point shown and carries no publication date in the retained text. It supports three propositions and no more: that the three named operators reached a final agreement on the structure of a planned tower joint venture, that each was to transfer tower assets into it, and that the three are therefore connected to the tower company as asset contributors. It does not support any statement about pricing, contract term, renewal, volume commitments, shareholdings or the date of the agreement, and none is made here.

The consequence for the thesis is direct. A price set by agreement with the customers who are also the asset contributors is a different constraint from a price set in a competitive market, and a reader cannot judge which constraint binds without the agreement's terms. Those terms were not obtained. So the article treats the pricing of the carrier business as an open question with a named missing document, not as a finding.

The access boundary, stated as a dated fact

The primary documents behind every figure in this article could not be read as text. The company's investor-relations landing page returned an access-failure notice stating that the relevant pages were not accessible through the search methods used https://www.chinatowercom.cn/en/investor/index.html, and a second retrieval against the 2026 results page returned a further notice that the requested first-half figures could not be retrieved and that the official portals remained inaccessible to automated browsing https://www.chinatowercom.cn/en/investor/results/2026/index.html. Neither page yielded document content, and neither is cited here as a statement about what the company does or does not disclose. They are retained as evidence of a boundary.

This article is therefore reported-only. Every figure in it rests on the two retained secondary reports and on arithmetic performed over them, not on a reading of an interim report, an annual report, an exchange announcement or a regulator decision. That is a weaker evidentiary position than a filing-based analysis, and it has one practical consequence for the reader: the numbers should be treated as an accurate rendering of what the company said, and the interpretation of those numbers should stop where the rendering stops.

What would change the answer

Three observations would settle the question this article raises, and none of them is in the record. A segment disclosure showing the Two Wings businesses earning a margin and a return on capital at or above the tower business would falsify the reading that the non-tower return is untested. A full-year statement in which operating cash flow exceeded capital expenditure would falsify the reading that capital intensity currently runs ahead of internal cash generation. A disclosure showing the tower business margin falling while the Two Wings margin rose would falsify the reading that the mature tower base carries the return.

Until one of those appears, the honest description of the first half of 2026 is this: a company that reported RMB 48.69 billion of revenue, of which 16.3 per cent came from businesses whose cost and capital intensity were not disclosed, and that spent 1.63 times its operating cash flow on capital in the same six months. That is a revenue-composition fact set against a cash fact. It is not yet a statement about returns, and treating it as one is the mistake this article exists to name.

Sources and method

Two secondary reports carry the figures: a transcript of the company's first-half 2026 results call, read in two retrievals, and an independent industry round-up that reconciles with it on revenue, site count and tenancy ratio and adds a tenant count. One historical news excerpt, truncated and undated, records the three operators' agreement on the structure of their planned towers joint venture and their intention to transfer tower assets into it. Two company investor-relations URLs are retained solely as evidence that the primary documents could not be read as text. All arithmetic in this article is performed on those retained figures and is stated so that a reader can repeat it. The company's directory entry is at https://btw.media/en/directory/china-tower-corporation-limited-cn.