Newly publishedNew
Fathom.

Telecom

ARPU tells you if monetisation is actually working.

ExamplesBHARTIARTLIDEAINDUSTOWER
How this business works

Telecom builds a hugely expensive network of towers and spectrum, then earns a small amount from each subscriber every month. Once the network is built, every extra rupee of revenue is almost pure profit, so the whole game is filling the pipe and charging more per user. Revenue per user is the honest scoreboard: adding cheap subscribers while that number stalls is just competitive bleeding. Spectrum auctions are brutal cash outflows that can starve a player of the money it needs to keep the network good.

The whole industry compressed into five boxes
Spectrum
Network
Subscribers
Monthly ARPU
Reinvest, then cash

Airwaves are bought at auction, towers turn them into a network, the network fills with subscribers, and each subscriber pays a small amount every month. Then comes the catch: most of that money goes straight back into the next round of spectrum and the next generation of network before the owner sees any of it. The whole investment case is whether anything escapes the loop.

The telecom checklist

Six questions to run against any telco

Rising ARPU, self-funded capex, a survivable balance sheet and quality subscribers. In telecom that short list separates the one or two investable companies from the rest of the industry.

The six questions that explain this business

Answer these six and you have understood the industry. The rest of the page is the detail behind them.

Where does demand come from?
From more or less everyone, and it does not go away. A mobile connection and data have become as basic as electricity: people keep paying the monthly bill through good times and bad, so demand is both recurring and structural rather than a fashion that fades. On top of that, each user tends to consume more data every year, so even a flat subscriber base can grow in value. This is one of the most dependable sources of demand in any industry.
Who controls the price?
It depends on how many players are left standing. When a market is crowded with rivals fighting for share, prices get driven to the floor in an endless war and nobody has pricing power. Once the industry consolidates into a handful of players, as it has in India, the survivors can finally raise tariffs and push users onto pricier data plans without fear of being undercut. So pricing power in telecom is earned through consolidation, not through the product itself.
What's the hardest thing to get?
Spectrum, the finite slices of airwaves that carry every call and every byte of data. There is only so much of it, the government owns it, and it is handed out through brutally expensive auctions, so no amount of ambition lets a company simply conjure more. A telco also needs a vast network of towers, but the true bottleneck is spectrum: without it, the towers have nothing to transmit.
Where does the money disappear?
Into the network, endlessly. Spectrum auctions swallow tens of thousands of crores, and each new generation of technology, 3G to 4G to 5G, forces another huge round of spending on towers and equipment before customers pay a rupee more for it. Because this capex is usually funded with debt, the money keeps leaving through interest payments long after the network is built. The revenue may be dependable, but the cash is forever being ploughed back into the pipes.
What usually breaks first?
A price war meeting a mountain of debt. A telco carries heavy borrowings from spectrum and network spending, and that debt does not shrink when a rival slashes prices and drags ARPU down. Caught between debt it must service and capex it cannot stop without losing customers to the better network, a weak player spirals. That is exactly how Vodafone Idea nearly collapsed: not a sudden loss of customers, but years of ARPU too low to cover its debts.
Why can't rivals just copy it?
Spectrum and the network built on it, plus the sheer scale needed to afford both. Only a few balance sheets can fund the tens of thousands of crores a nationwide network demands, which is why the industry naturally shrinks to a handful of giants rather than staying fragmented. Once a player owns prime spectrum, dense towers and millions of subscribers, a newcomer cannot cheaply replicate any of it. The moat is the wall of capital and the finite airwaves that money alone cannot buy more of.
The question beginners always ask
If everyone sells the same data, why is one operator so much more profitable?
Because the difference is not the price of data, it is the cost of serving each customer. An operator with more customers spreads the same enormous fixed cost of spectrum and towers over more people, so each one is cheaper to serve. Better spectrum and a denser network also mean fewer dropped calls and less money spent fixing complaints and winning back angry customers. And the operator with the stronger balance sheet pays far less interest on the debt that funded all of it. Same data plan, but the bigger, better-run, less-indebted operator keeps much more of every rupee.

First, what is a telecom business really?

Strip away the jargon and a telco is a landlord of invisible pipes. It spends enormous sums up front to build a network, then rents access to it, one subscriber at a time, for a monthly fee.

01

Fixed cost up front, near-free after that

Building towers and buying spectrum costs tens of thousands of crores before a single customer connects. Once that network exists, serving one more subscriber costs almost nothing extra, so the business has huge operating leverage. That is why market share swings hit profit so hard in both directions, a subscriber gained or lost falls almost straight to the bottom line. This is also why the industry consolidates into a handful of players rather than staying fragmented like a grocery store chain.

For exampleIn India this has settled into a 3 player oligopoly, Jio, Airtel and Vi (Vodafone Idea), because only a few balance sheets can afford the network build in the first place.
02

You are renting a monthly relationship, not selling a product

A telco does not sell you a phone call, it sells you a subscription that renews every month whether you think about it or not. That means the real product is retention: keeping a customer paying long after the network cost is sunk. A customer who stays 5 years is worth many times a customer who leaves after 5 months, because the acquisition cost (SIM, marketing, dealer commission) is paid once but recovered slowly.

For exampleIf it costs a telco ₹1,500 to acquire a subscriber and that subscriber pays ₹200 ARPU a month, the company needs roughly 7-8 months just to break even on acquiring them.
03

The demand is wonderful. Hold that thought.

Here is the paradox worth sitting with before anything else. Telecom sells one of the most dependable products in any economy: a mobile connection is now as basic as electricity, people pay through good times and bad, and each user consumes more data every year. Wonderful demand, terrible history. Indian telecom has destroyed enormous amounts of shareholder money despite everyone using its product all day, every day. The rest of this page explains that paradox, and the explanation transfers to every industry: dependable demand is worth nothing to the owner if pricing is broken and the cash never stops leaving. Demand pays the customer's bill. Structure decides who keeps the money.

For exampleBetween 2010 and 2020 Indian data usage grew perhaps a hundredfold. Over the same stretch, most telecom shareholders earned roughly nothing, and several operators ceased to exist. Usage is not the same as profit.

ARPU: the one number that is the whole scoreboard

Every industry on this site has one number that tells you most of the truth. In telecom it is ARPU, and it deserves its own section because everything else hangs off it.

01

What ARPU is and why it cannot lie

Average Revenue Per User is simply monthly revenue divided by subscriber count: how much money the company collects per customer per month. It is the cleanest read on whether a telco is winning quality customers at fair prices or buying cheap subscribers with discounts. Rising ARPU on a stable base means real pricing power at work. Huge subscriber growth with flat ARPU means the company is purchasing market share and calling it growth. Subscriber counts flatter easily; the average revenue each one pays does not.

For exampleBharti Airtel's ARPU crossing ₹200 while Vodafone Idea sat near ₹150 in the same market, selling the same data, tells you plainly which one has pricing power and which one is fighting to survive.
02

Falling ARPU is the plainest warning in the sector

Because the product renews monthly, ARPU is reported every single quarter, which makes it a running commentary on the company's power to price. When it falls for quarters on end, the message is unambiguous: customers are being kept only by charging them less, and the business is shrinking per head even as the subscriber line grows. This is the single most important thing to watch, and it is handed to you in every results presentation. No modelling required.

For exampleAirtel's ARPU fell from about ₹202 in 2014 toward about ₹104 after the price war began. Three full years of quarterly warnings, printed in the company's own presentations, preceded the record loss of November 2019. Nobody who tracked that one number was surprised.
03

How ARPU actually rises

A telco lifts ARPU three ways. Tariff hikes, which only stick when competitors follow, which is why they only happen once an industry has consolidated. Mix, moving customers from prepaid to postpaid and from basic packs to bigger data bundles, postpaid customers pay more and leave less. And attach, layering extras onto the same connection: streaming bundles, cloud, enterprise lines, home broadband. The first depends on the industry structure. The second and third are the company's own work, and they are how a good operator grows revenue even between tariff hikes.

For exampleThe same customer who paid ₹150 for a basic pack can, over a few years, become a ₹400 customer: a postpaid plan, a bigger data bundle, a streaming subscription attached. No new customer was won. The existing one was deepened.

The treadmill: spectrum and the network that is never finished

Now the cost side, and the reason this industry eats cash. A telecom network is not built once. It is built forever.

01

Spectrum: renting the air from the government

Spectrum, the slices of airwaves that carry every call and byte, is the industry's scarce resource. The government owns it and sells it through auctions, and because no telco can operate without it, auctions become forced purchases at fearsome prices: tens of thousands of crores, recurring every few years, usually funded with debt. Think of it as rent for the air, paid in advance, in lump sums large enough to bend a balance sheet. When you hear that a telco is cash-rich going into an auction year, wait for the auction before believing it.

For exampleIndian spectrum auctions have raised lakhs of crores for the government over the years. Every rupee of it came off telecom balance sheets, mostly borrowed, and had to be earned back ₹200 a month at a time.
02

The generation upgrade trap

Every technology generation, 3G to 4G to 5G, forces telcos to rebuild the network at enormous cost, and here is the trap: customers do not pay more just because the technology changed. The new capacity earns its keep only years later, once usage grows into it and tariffs catch up. So a telco's reported profits right after a big capex cycle understate the business, and its profits late in a cycle, when the network is sweated and paid for, overstate what the next cycle will allow. Investors who buy right after the spending, before the payback, are paying for a network that has not yet earned anything.

For exampleIndian telcos spent through 2022 to 2024 on 5G spectrum and towers with minimal extra revenue to show for it by 2024, since most 5G usage sat inside 4G-era pricing. The spending was real and immediate. The payback was a promise.
03

The mouth that never closes

Put the auctions and the upgrades together and you get the defining test for any telco: put cash generated from operations next to purchase of fixed assets plus spectrum payments, side by side, for five years running. In a healthy business the first line comfortably exceeds the second and the surplus reaches the owner. In most of telecom's history, the second line kept swallowing the first, so however fast revenue grew, nothing was left over. This is arithmetic, not opinion, and it sits in every annual report's cash-flow statement waiting to be read.

For exampleAirtel's cash-flow statements from 2010 onward showed exactly this: operating cash pouring straight back into towers and spectrum, year after year. Revenue grew for a decade while owners waited. The Airtel case study calls it the mouth that never closes.

The price war: how India got to three players

In 2016 India ran the most brutal natural experiment in telecom economics anywhere in the world. Understanding it explains the entire industry you see today.

01

What happens when a fixed-cost industry gets a rich new entrant

In a high fixed-cost business, a new entrant with deep pockets can price below everyone's cost, because its own cost per user falls as users pile in, and its losses are an investment in market share. Jio launched in 2016 with free voice and data cheaper than anyone could match. Every incumbent had the same choice: match the prices and bleed, or hold prices and watch subscribers walk. Everyone matched. Industry revenue per user roughly halved, and a dozen operators became three. Fixed-cost industries do not stay crowded; a price war is how they thin out, and the entrant with the strongest balance sheet writes the rules.

For exampleBetween 2016 and 2019, operators merged, sold out or shut down until three remained: Jio, Airtel, and the merged Vodafone Idea. The war ended not when anyone won the argument, but when there was nobody left to undercut.
02

Consolidation is where pricing power comes from

Here is the transferable lesson: in telecom, pricing power does not come from the product, which is identical across operators, it comes from how many players are left. A crowded market prices at the floor. Once three players remain and the weakest can no longer afford to fight, the survivors can finally raise tariffs together, and every hike falls almost straight to profit because the network is already paid for. This is why the investment case for Indian telecom flipped after 2019: the same companies, the same towers, but a structure that finally allowed prices to rise.

For exampleThe tariff hikes from late 2019 onward, the first meaningful ones in years, happened only after the market had shrunk to three players. ARPU began climbing back, and Airtel's stock finally escaped its lost decade. Structure changed; that was the whole difference.
03

The weakest survivor is not really a survivor

A three-player market with one player on life support behaves like a two-and-a-half player market. The weak operator cannot invest in its network, so it loses its best subscribers each quarter; it cannot let go of price, so it cannot rebuild ARPU; and its debt keeps compounding regardless. It survives only as long as lenders and the government prefer delay to burial. For an investor, this is the sharpest sorting question in the sector: are you buying the consolidation winners, or the company the consolidation happened to?

For exampleVodafone Idea has lost subscribers to both rivals every quarter for years while carrying obligations far beyond its earning power. Its continued existence is a policy choice, not a business outcome, and its equity is a bet on rescue terms, not on telecom economics.

The rule-maker: when the government changes the bill

A telco's biggest costs and its right to operate both come from the government. That makes regulation not a background risk but a first-order one, and one judgement in 2019 proved it.

01

The AGR shock

Indian operators pay the government a share of their adjusted gross revenue, AGR, as licence and spectrum fees. For over a decade, the operators and the government disputed what counted as revenue: only telecom income, or everything including interest and asset sales. In October 2019 the Supreme Court ruled for the widest definition, retrospectively, with interest and penalties compounding on 14 years of dues. The bills landed in the tens of thousands of crores, for years the companies had already closed. Airtel booked a provision of about ₹28,450 crore and reported a record quarterly loss. Nothing about the network, the customers or the tariffs had changed. The rules had.

For exampleA definition argument over the word revenue produced one of the largest corporate losses in Indian history. That is what regulatory risk means in a licensed industry: the past itself can be repriced.
02

How to hold this risk in your head

The lesson is not that telecom is uninvestable; it is that in any industry where the government sets the biggest costs, the balance sheet must be judged against the regulator's power, not just the market's. A strong operator absorbs a rule change, raises capital, and carries on. A leveraged one is pushed to the edge by the same ruling. Before buying any telco, ask what an adverse ruling or a surprise levy would do to this specific balance sheet, because the sector's history says such things arrive roughly once a decade. The same test applies to power, banking, and every other licensed industry on this site.

For exampleThe same AGR judgement hit Airtel and Vodafone Idea. Airtel raised capital within months and moved on. Vodafone Idea never recovered. One ruling, two balance sheets, two completely different outcomes.

Where telecom breaks

The failures in this industry follow one script closely enough that you can learn it once and recognise it anywhere.

01

The standard collapse: debt meets a price war

Spectrum and networks are bought with debt, and debt does not shrink when a rival cuts prices. A telco caught with heavy borrowings in a price war faces a closing trap: it cannot cut capex without losing customers to the better network, and it cannot keep spending without drowning in interest. ARPU falls, the debt stays, and the gap between what the business earns and what it owes widens every quarter until lenders decide the story is over. Note what the killer is: not losing customers, but carrying obligations sized for prices that no longer exist.

For exampleVodafone Idea's near collapse was not a sudden exodus of customers. It was years of ARPU sitting below the level its spectrum and bank debt required. The customers stayed longer than the balance sheet did.
02

The Airtel mechanism: surviving your own mistake

Airtel made the classic error too. In 2010 it paid $10.7 billion, borrowed, for Zain's African operations, and net debt swelled toward ₹60,000 crore. From that day, the first claim on every rupee belonged to lenders, and the stock went nowhere for a decade. So why did Airtel survive what killed Jet Airways, the same debt-funded acquisition mistake? Because of what the debt sat on. Jet's acquisition added losses to a business with no pricing power and evaporating revenue. Airtel's sat on subscriptions that renewed every month through the entire ordeal; the revenue never fell away, so the debt could be serviced, however painfully, until consolidation restored pricing. The transferable lesson: the same mistake is survivable on recurring revenue and fatal on cyclical revenue. Judge the error by what carries it.

For exampleThe Airtel case study traces the whole lost decade, tell by tell: the borrowings note of 2010, the ARPU slide from 2016, the cash-flow statement that never left anything over. Growth for ten years, nothing for owners, and then, after consolidation, everything at once.

How to value a telco

Telecom valuation has two rules that beginners skip, and skipping either one leads to buying the wrong company at the wrong time.

01

Value the enterprise, not the equity

Because telcos carry enormous and unequal debt, market capitalisation alone is meaningless for comparison. Always work with enterprise value, market cap plus net debt, because that is the real price of the business including the obligations you inherit. The standard yardstick is EV against EBITDA, the operating profit the network throws off. Two telcos with the same market cap can be wildly different purchases once the debt is counted; one of them, you are mostly buying the lenders' problem.

For exampleA telco with ₹50,000 crore of market cap and ₹2,00,000 crore of net debt is a ₹2,50,000 crore purchase. Its equity is a thin slice on top of a mountain of obligations, which is why such stocks swing violently on any news about the debt rather than news about the business.
02

Free cash flow after spectrum is the only honest profit

EBITDA in telecom flatters, because it ignores precisely the two things that eat this industry alive: capex and spectrum. The number that matters is what remains of operating cash after both, free cash flow after spectrum, because that is the money that can actually reach you. A telco can report fat EBITDA margins for a decade while this number stays at or below zero, and its shareholders will have owned a growing network and earned nothing. Pay for telecom cash flows only when the structure, consolidation, rational pricing, a funded network, lets them actually exist.

For exampleThrough Airtel's lost decade, EBITDA margins looked respectable almost throughout. Free cash after capex and spectrum told the true story: nothing left over. The investors who waited for that number to turn positive, after consolidation, caught the entire re-rating. The ones who bought the EBITDA story in 2010 waited ten years.
The five numbers that decide the story

Every sector reduces to a demand metric, a pricing metric, an efficiency metric, a capital metric, and a risk metric. For telecom, these are the ones that matter.

Demand
Subscriber growth
Pricing
ARPU
Efficiency
EBITDA margin
Capital
FCF after spectrum
Risk
Churn / debt
The full metric set

What each number tells you, and how to read it.

MetricWhy it matters
ARPU (Avg Revenue Per User)Pricing power per subscriber. Rising ARPU means monetisation works; flat ARPU in a growing base means competitive pressure.
Subscriber GrowthScale, which brings fixed-cost leverage. Quality (postpaid versus prepaid) matters as much as quantity.
ChurnCustomer loyalty. Low churn signals a sticky network effect and better coverage or brand.
Data Consumption / UserMonetisation potential. Higher data usage is the path to higher ARPU tiers and add-on services.
Spectrum CostsCapital intensity. Auctions are massive cash outflows; heavy spectrum debt constrains network capex.
EBITDA MarginEfficiency on fixed infrastructure. Telcos should post 40-50% EBITDA margin at scale.
Net Debt / EBITDASurvival gauge. Spectrum is bought with debt; this ratio says whether the next auction or ruling is absorbable or fatal.
FCF after SpectrumThe only honest profit in telecom: operating cash minus capex minus spectrum. Until this is positive, owners are funding the network, not the reverse.
One sentence to remember

Telecom is a fixed-cost machine that rewards its owners only once consolidation gives the survivors the power to raise prices.

Take these ideas further

Fixed-Cost MachinePricing PowerCapital IntensityConsolidation