Summary

  • Telefónica Brasil reported 32.0 million fibre homes passed against 8.2 million connected fibre customers, so roughly one home in four of the network it can already reach was a paying fibre line at the reported date.
  • Connected lines grew 11.3% while homes passed grew 6.4%, and FTTH revenue grew 10.7% — slower than the customer count — so growth came from filling existing plant at a broadly flat revenue per line.
  • Capital expenditure grew 6.1% to R$2.59 billion, equal to 16.4% of revenue, while revenue grew 7.6% to R$15.8 billion: the company is now converting faster than it builds, which changes what shareholders should watch.

The first number to look at in Telefónica Brasil's second quarter of 2026 is not a number at all. It is a direction. Revenue grew faster than capital expenditure.

For most of the past decade, Brazilian fibre ran the other way round. Operators built ahead of demand because the household you did not reach this quarter was a household a rival would reach the next. Capital expenditure led, revenue followed, and the gap between them was the price of a land grab. Telefónica Brasil — the listed company Brazilians know as Vivo — reported revenue of R$15.8 billion for the quarter, up 7.6%, against capital expenditure of R$2.59 billion, up 6.1%, or 16.4% of revenue, with first-half capital expenditure of R$4.64 billion, up 7.6% (https://telecomlead.com/telecom-services/telefonica-brasil-q2-2026-revenue-rises-7-6-to-r15-8-bn-as-vivo-expands-mobile-fiber-digital-services-and-5g-126969). The same report put 5G coverage in 978 municipalities, roughly 73% of the population, with investment focused on 5G and fibre.

That sequence — revenue outpacing capex — is what a completed build looks like from the outside. It is not proof of anything about profits. The retained record used here carries no EBITDA, operating margin, net income, ARPU or net-debt figure for the company, so what follows rests on volume, capital intensity and the relationship between two growth rates, not on a margin story. That limit does not weaken the mechanism; it simply confines it to the part of the accounts that were actually read.

The plant is built; the lines are not

The two figures that matter most are 32.0 million and 8.2 million.

The first is homes passed: premises where the company's fibre is in the building and an order could be fulfilled without digging another metre of street. It rose 6.4% in the quarter. The second is connected fibre customers: households that actually pay. It rose 11.3%, to 8.2 million.

Divide one by the other and the shape of the business becomes obvious. Eight point two million paying lines against 32.0 million homes passed means roughly one connected customer for every four premises the network already touches. Three-quarters of the built footprint was not yet a paying fibre line at the reported date. That ratio is arithmetic on the company's own reported figures rather than a metric the company published, and the difference matters: the company reports the two totals, not the yield between them.

The vocabulary of this industry hides how large that gap is. "Homes passed" sounds like progress, and it is — but it is progress measured in construction. A home passed is an addressable premise. A connected customer is a billing relationship. Between the two sit marketing, installation, credit checks, the incumbent cable operator's counter-offer and the simple inertia of a household that has not yet got round to switching. Every one of those steps is cheaper than trenching. None of them is free.

The strategic consequence is easy to state and hard to execute. The addressable market no longer sits across the street in a district the company has not reached. It sits inside the plant the company already owns. Growth from here does not require new construction, new spectrum or a new licence. It requires conversion. An operator in that position has traded a capital problem for a commercial one, and commercial problems are settled one household at a time.

Four growth rates that describe one mechanism

Four reported growth rates, set side by side, describe a single transition rather than four unrelated facts.

Homes passed grew 6.4%. Connected fibre customers grew 11.3%. FTTH revenue grew 10.7%. Capital expenditure grew 6.1%.

Read them in that order and the mechanism assembles itself.

First, connections outgrew the footprint by 4.9 percentage points. The marginal fibre customer therefore came disproportionately from plant that already existed rather than from new build. This is the arithmetic of fill-out: when the paying base expands nearly twice as fast as the network reaching new premises, the company is selling into its own map, not extending it.

Second, FTTH revenue grew 10.7% while connected customers grew 11.3%. Revenue grew more slowly than the customer count. Revenue per connected fibre line therefore did not rise over the period; on the reported growth rates it edged down. The money grew because the base grew. That is a specific and unsentimental fact about how this expansion is being paid for: volume, not price.

Third, capital expenditure grew 6.1% while revenue grew 7.6%, with capex standing at 16.4% of revenue. Capital intensity eased even as the paying base expanded faster than the footprint. In a build-out story capex accelerates and connections lag; here the reverse happened.

Each of those readings has a counterfactual that would overturn it, and the counterfactuals are worth stating plainly because they are what a sceptical reader should test. If FTTH revenue had grown faster than connected lines, the correct reading would be revenue expansion per customer — pricing power or a richer mix — and the fill-out framing would be wrong. If homes passed had grown faster than connections, the company would still be in build-out and this would be an ordinary capex story.

If capex had grown faster than revenue while connections lagged, the reading would be over-building into a soft market — the classic way a fibre bet goes wrong. None of those three conditions is what the reported figures show.

The combined effect is a company whose capital programme has changed character. The same real of capex buys less new footprint than it bought during the land-grab years, because a growing share of it goes into connecting premises rather than reaching them. That is not a retreat from fibre. It is fibre entering the phase where returns depend on commercial execution rather than construction crews.

Second place in a market that counts paying lines

The competitive frame comes from a different source and a different series. Anatel's fixed-broadband access data, as reported for the June 2026 close, put Claro first with 10.8 million active connections after 21 thousand net additions in the month, and Vivo second with 8.4 million accesses after 57 thousand net additions (https://tecnoblog.net/noticias/banda-larga-vivo-lidera-starlink-avanca-e-nio-cresce-pelo-terceiro-mes/).

The distance is 2.4 million connections. The monthly flow points the other way: in that month Vivo added 36 thousand more net connections than the leader.

A single month of net additions is a thin basis for a trend claim, and it should be treated as what it is — one data point in a contest that will be decided over many of them. But it is enough to locate the lever. In a market where rank is counted in paying lines, the leader's advantage is the size of its paying base, not the exclusivity of its reach. Rivals can build glass to the same buildings; what they cannot instantly duplicate is the billing relationship inside those buildings. Telefónica Brasil's position is therefore not determined by how many premises its network touches.

It is determined by how quickly it converts the ones it touches, and by how much revenue each conversion carries.

One caution belongs here rather than at the end. The 8.4 million figure is an Anatel access count reported by Tecnoblog; the 8.2 million figure is the company's own connected fibre customer count. They come from different sources and may not use the same counting rules, and they must not be added, netted or treated as one series. What can be said is directional and modest: the company's own fibre base is approaching, but is not established to equal, its measured fixed-broadband access base.

What an exhausted address pool does to a fill-out story

The layer beneath the fibre has its own arithmetic. LACNIC states that IPv4 addresses have reached the exhaustion phase in its service region, that new allocations from the exhausted pool are restricted, and that remaining space is subject to specific policies (https://www.lacnic.net/en/web/lacnic/ipv4-exhaustion).

That is the registry's own institutional statement about its own pool, and it should be read as exactly that. It establishes nothing about Telefónica Brasil's holdings, registrations, autonomous systems or block sizes; repeated attempts to read those registry records in the preparation of this article returned provider summaries that recorded the registry response was never obtained, so no holding claim is made here at all. What the statement does establish is the environment in which a conversion story plays out.

Every connected household at the margin is not only a fibre line but a bundle of digital identity: an address, a route, a way of being reached. The physical half of that bundle is already paid for, sunk into streets the company has already dug. The identity half is not reproducible at will in a region whose registry says its free pool is exhausted. An operator growing by converting installed plant therefore depends on number resources it already holds rather than on resources it can procure on demand.

Scarce complements change behaviour: they raise the value of what is already on the books, they make the wholesale relationship between an infrastructure owner and the services riding on it more consequential, and they quietly penalise any growth strategy that assumed the addressing layer would always be available at the marginal cost of a form.

How much that constraint actually binds Telefónica Brasil is not something the retained record can answer. That is a gap, stated as a gap, and it is the most important thing this article does not know. What the record does support is the narrower claim: the company's growth path now runs through assets it already owns, on both the physical and the addressing side, which is a different kind of business from the one that spent a decade building.

Three distinct actors appear in this account and none of them is the others. Telefónica Brasil S.A. is the Brazilian listed telecom that trades as Vivo and reports the capex, footprint and connection figures — its directory record is here (https://btw.media/en/directory/telefonica-brasil-s-a-brazil Only for Strategic Circle). Claro is the competing operator that Anatel's access data places first in fixed broadband. LACNIC is the regional internet registry whose exhaustion statement is reported above as its own claim. No figure from one is attributed to another.

Tests that would break this reading

A reading that cannot be falsified is a mood, not an analysis. Four tests would settle this one, and all four are observable in the next disclosures.

If FTTH revenue grows faster than connected fibre customers in the next reporting period, the fill-out framing weakens materially. That combination would mean revenue per line is rising, and the story would shift from volume to pricing power — a better story for the company, a different one for this argument.

If homes passed grows faster than connected customers, the company has returned to extending the footprint ahead of demand, and this becomes a capex-cycle piece again rather than a conversion piece.

If capital expenditure grows faster than revenue while connections lag, the correct diagnosis is over-building into a market that is not absorbing it, and the eased capital intensity reported here was a single-quarter artefact rather than a policy shift.

The fourth test is not a ratio but a presence: whether a wholesale or open-access revenue line appears at all. The retained record contains none, which means how far Telefónica Brasil already monetises its plant through third parties cannot be measured from what was read. An infrastructure owner sitting on three-quarters unconnected capacity has an obvious alternative to retail conversion — renting the capacity to someone else who converts it — and the accounts as read here neither confirm nor exclude that business.

Q2 2026 is the most recent completed disclosure in the record used for this article. That is a limit of access, not a statement that the company has not reported since. When the next period lands, these four tests can be run in an afternoon, and the answer will be visible in four numbers rather than in a narrative.

What the current numbers say is that Telefónica Brasil has stopped being a company whose constraint is concrete and become one whose constraint is persuasion. The plant is there, the capex curve has bent, and the paying base is growing faster than the footprint that carries it. Everything now depends on how many of those 32 million premises can be turned into bills before a competitor reaches them first — at a price the company will be able to defend afterwards.