Telecom looks like a consumer business. You recharge a prepaid plan the way you buy toothpaste, the advertising is about films and cricket, and the product arrives on a phone in your pocket. Underneath that costume it is one of the most capital-hungry infrastructure businesses in the country — towers, fibre, power, and government-licensed airwaves bought at auction years before a single customer uses them.
That gap between how it looks and what it is explains almost everything about the sector's numbers: why a handful of operators survive where dozens once competed, why one tariff decision can reprice the whole industry, and why the same balance sheet that looks crushing in a build year can look like a cash machine three years later.
An infrastructure business wearing a consumer costume
Before an operator earns its first rupee from a new network, it has to buy spectrum at a government auction, acquire tower sites, install radio equipment, lay or lease fibre to connect those towers back to the core, and pay for the electricity and diesel that keep it all running. That spending happens years ahead of the revenue.
Once it is built, the network costs roughly the same to run whether it is half loaded or nearly full. Rent on a tower does not fall because fewer people made calls this month. Spectrum charges do not fall because usage dipped. This is what economists call a high fixed-cost base, and it is the defining fact of the sector.
Everything else follows. Scale is not a nice-to-have here; it is the business model. An operator without enough subscribers to cover the fixed cost does not have a thin-margin business — it has a loss-making one, and the losses compound while the network keeps running.
Average revenue per user — the number the sector lives on
Average revenue per user (ARPU) is exactly what it sounds like: the service revenue earned in a period divided by the average number of subscribers over that period. It is usually quoted per month, in rupees.
ARPU is the sector's single most-watched metric because it captures pricing, plan mix and usage in one figure. It rises when tariffs go up, when subscribers move from prepaid to postpaid, when they upgrade to higher-data plans, or when low-value inactive users are cleaned out of the base. It falls when operators discount to win share, when the mix shifts toward cheaper plans, or when a price war starts.
The reason it matters more than subscriber count is that subscribers can be bought. Cutting price adds users and lowers ARPU at the same time, and the two effects can leave revenue flat while margin gets worse. Reading subscriber additions without ARPU beside them is how people misread a price war as growth.
ARPU = service revenue ÷ average subscribers for the period- Service revenue excludes equipment and one-off sales
- Quoted monthly, in rupees, in the quarterly results release
- Read alongside net subscriber additions — never on its own
Note — Check the definition before comparing two operators. Some report ARPU across the whole base, some split prepaid and postpaid, and some restate it after removing long-inactive subscriber identity module (SIM) connections. A definitional change can look like a business change.
Operating leverage cuts both ways
Because the cost base barely moves, almost every additional rupee of revenue above the break-even point falls through to profit. Add a subscriber to an existing tower and there is essentially no extra cost to serve them. Raise tariffs across a large base and the increase is close to pure margin, because you have not added a single item of equipment.
That is operating leverage, and it is why telecom margins can expand very fast once scale is reached. It is also why the sector is punishing on the way down. If ARPU falls or subscribers leave, the towers, spectrum charges, power bills and interest do not shrink with them. Revenue falls faster than cost, and the same leverage that produced spectacular margin expansion produces spectacular losses.
This is the reason telecom rarely has a comfortable middle. Operators tend to be either well above scale and generating strong cash, or below it and burning it — and moving from the second state to the first usually requires either a tariff repair across the industry or a large capital infusion.
Four different businesses live inside 'telecom'
Treating the sector as one thing is the most common analytical mistake. Consumer mobility — prepaid and postpaid phone connections — is the visible business, and the most volatile, because it is exposed directly to tariffs and to churn. Enterprise and data services sell connectivity, cloud links and data-centre capacity to businesses on contracts, which makes revenue steadier and less tariff-sensitive.
The two infrastructure businesses behave differently again. A tower company owns the physical site and rents space on it to operators, typically on long leases with escalation clauses. Its revenue rises when a second or third operator adds equipment to a tower that is already built, so its economics turn on tenancy — how many tenants per tower. Indus Towers is a listed example of exactly this model, and its customers are the operators themselves.
Fibre is the same idea in a different physical form: enormous front-loaded cost to lay a route, then long-dated leasing of capacity to operators and enterprises. Both are annuity businesses rather than consumer ones, which is why they are often held in separate vehicles — including infrastructure investment trusts (InvITs), which are pass-through structures designed to hold operating infrastructure and distribute its cash flow.
| Sub-model | How it earns | Capital intensity | Cash-flow character |
|---|---|---|---|
| Consumer mobility | Monthly recharges from retail subscribers | Very high, in bursts | Volatile; tracks tariffs and ARPU |
| Enterprise & data | Contracted links, cloud and data centres | High but staged over time | Steadier, contract-backed |
| Towers | Rent per tenant, on long leases | High upfront, light afterwards | Annuity-like and highly visible |
| Fibre | Long-dated capacity leases to operators | Very high and front-loaded | Slow to fill, annuity once let |
Spectrum, dues and levies: the balance-sheet reality
Spectrum is the airwaves an operator is licensed to use. In India it is sold by the government at auction, for a fixed term, in defined bands. It is bought before it earns anything, it is expensive, and much of the price is typically paid in instalments over years — so an auction commitment made once keeps sitting on the balance sheet as deferred payment obligations long afterwards.
On top of that, operators pay recurring levies to the government calculated as a percentage of adjusted gross revenue (AGR): a licence fee and a spectrum usage charge. What exactly counts inside 'adjusted gross revenue' was disputed for many years. The dispute was ultimately settled against the operators, which left them with large retrospective dues, and a later government relief package offered a moratorium on payments together with an option to convert some dues into government equity.
The investing point is not the history. It is that in telecom, a meaningful part of the liability side is owed to the government under terms the government can revisit — so policy is not an external risk factor here, it is inside the balance sheet.
- Spectrum
- Government-licensed radio frequency, bought at auction for a fixed term. The raw input a mobile network runs on.
- Adjusted gross revenue (AGR)
- The revenue base on which government levies are calculated. Its definition has been the sector's single biggest legal issue.
- Licence fee & spectrum usage charge
- Recurring percentage levies on AGR paid to the Department of Telecommunications (DoT). They fall due whether or not the operator is profitable.
- Deferred spectrum liability
- Auction dues payable in instalments over years. Real debt in economic terms, even when it is presented separately.
How many operators there are decides how much they can charge
Indian telecom went from a crowded market of more than a dozen operators to one dominated by a small number of private players alongside a state-owned operator. Entry by a new, very well-capitalised operator triggered a price war; the operators that could not fund losses at those tariffs either merged, sold their spectrum or exited.
Consolidation changed the sector's economics more than any technology upgrade did. When many operators chase the same subscriber, price is the only weapon and nobody earns their cost of capital. When few remain, tariff increases become possible — and because of operating leverage, an industry-wide tariff repair flows through to margin very quickly.
This is why the market watches tariff announcements the way it watches an earnings release. The relevant question is never just 'did one operator raise prices'. It is whether the others follow, and whether the increase survives without subscribers churning down to cheaper plans.
Watch out — Market-share gains are not automatically good news. Share won by discounting, funded with debt, on a network that still needs capital spending, is a worse outcome than flat share with rising ARPU. Always ask what the gain cost.
The capex cycle around each network generation
Every network generation runs the same loop. First an auction, where operators commit large sums for spectrum. Then a build phase, where capital spending spikes: radios on towers, fibre to connect them, upgrades to the core network. Then a coverage phase, where the network exists but the tariff for using it is not yet higher. Only after that comes monetisation, when usage and plan mix finally lift ARPU.
Cash flow through that loop is unrecognisable from one stage to the next. During the auction and build years, free cash flow can be deeply negative while the business is perfectly healthy strategically. In the monetisation years, with capital spending falling and revenue rising against a built-out network, cash generation can look extraordinary — until the next generation's auction resets the cycle.
The practical consequence is that a single year of telecom cash flow tells you almost nothing on its own. You have to know which part of the loop the company is standing in, and management's own capital-spending guidance in the earnings call is the most direct evidence of that.
What moves these stocks, and when
Telecom has an unusually clear information calendar. Subscriber data is published monthly by the regulator, financial detail arrives quarterly, and the two genuinely large events — tariff decisions and spectrum auctions — are irregular but well telegraphed. Knowing the cadence is most of the work.
| What to watch | Where it comes from | How often |
|---|---|---|
| Subscriber additions | Telecom Regulatory Authority of India (TRAI) reports | Monthly |
| ARPU and subscriber mix | Quarterly results and investor presentation | Quarterly |
| Tariff changes | Operator announcements and plan pages | Irregular, sector-wide |
| Spectrum auctions | Department of Telecommunications notices | Every few years |
| Dues and levy rulings | Court and government orders on AGR | Irregular |
| Capex guidance | Earnings calls and annual reports | Quarterly and annual |
What retail investors get wrong here
- Reading subscriber additions without ARPU. Users added by discounting can raise the count while lowering the quality of the revenue.
- Treating deferred spectrum payments as something other than debt because they are disclosed separately from borrowings.
- Judging a telecom company on one year of free cash flow, without asking whether that year was a build year or a monetisation year.
- Assuming a tower or fibre company will behave like the operator that spun it off — annuity infrastructure and consumer mobility are different businesses with different risk.
- Expecting a single operator's tariff increase to stick when competitors have not matched it.
- Underestimating how quickly operating leverage works in reverse when ARPU slips, because the cost base gives nothing back.
How to actually track it
- 1
Once a month, read the regulator's subscriber report
TRAI publishes operator-wise subscriber numbers and additions. It is the highest-frequency, most objective read on who is winning connections.
- 2
Every quarter, put ARPU next to net additions
Both rising is genuine strength. Additions rising while ARPU falls is share bought with price. ARPU rising while the base shrinks is a clean-up of inactive users, not necessarily growth.
- 3
Track capital spending against the network cycle
Note guided capex and where in the build-to-monetise loop the company sits. Falling capex with rising ARPU is the combination that turns telecom cash-generative.
- 4
Read the liability side properly
Add borrowings, deferred spectrum obligations and any government dues together. That combined figure, set against operating cash flow, is the real solvency picture.
- 5
Watch tariff news across all operators, not one
A price increase only holds if the others follow. Note who moved, who matched, and whether entry-level plans changed, because that is where churn happens first.
What sector analysis will not tell you
The mechanics above are true of the industry. They do not tell you how any one company is placed inside it, and telecom has an unusually wide spread between its participants — one operator can be generating strong cash from a built network while another is still funding a turnaround, in the same quarter, under the same tariffs.
Use the sector view to know what to ask: how much spectrum, bought when and owed how; where in the capex loop; ARPU direction and what produced it; how much of the liability side is government-linked. The answers are company-specific, they are all disclosed, and this page is not a view on any particular name.
Key points
ARPU (average revenue per user) = service revenue for the period ÷ average subscribers during the period
Example — Consider two operators with similar networks. One has enough subscribers to cover its fixed cost with room to spare; the other is just below that level. An industry-wide tariff increase of the same size lands very differently on them: for the first it is mostly incremental margin on a base that already pays for itself, while for the second it is the difference between funding the network and not. Same tariff, same technology, opposite consequences — that asymmetry is operating leverage, and it is why sector-level conclusions travel badly in telecom.
Pro tip — Keep one simple table per operator with three columns — net subscriber additions, ARPU, and guided capital spending — updated every quarter. Almost every important question in telecom, from pricing power to cash generation to balance-sheet stress, becomes visible in how those three move relative to each other over a year.
Warning — Deferred spectrum payments and government dues are real obligations even when they are presented separately from borrowings. A telecom balance sheet read on reported debt alone can look far more comfortable than it is. Add spectrum liabilities and any outstanding levy dues before you judge leverage in this sector.
Frequently asked questions
What is ARPU in telecom and how is it calculated?
ARPU stands for average revenue per user. It is the service revenue earned in a period divided by the average number of subscribers over that period, usually expressed per month in rupees. Operators disclose it in their quarterly results and investor presentations. Because definitions vary — some split prepaid from postpaid, some exclude long-inactive connections — check the basis before comparing two companies.
Why does telecom need so much capital before it earns anything?
A mobile network is unusable until it is essentially complete across a service area. Spectrum has to be bought at auction, tower sites acquired and equipped, fibre laid or leased to connect them, and the core network upgraded — all before subscribers can be served at scale. That spending is front-loaded by years, which is why free cash flow in a build phase can be deeply negative even when the business is executing well.
What are AGR dues and why do they matter so much?
Adjusted gross revenue (AGR) is the revenue base on which operators pay a licence fee and a spectrum usage charge to the government. What counts inside that base was disputed for years and was ultimately settled against the operators, leaving large retrospective dues. A later relief package allowed a payment moratorium and an option to convert some dues into government equity. The lasting lesson is that a significant part of a telecom balance sheet is owed to the government on terms policy can change.
How is a tower company different from a telecom operator?
A tower company owns physical sites and rents space on them to operators under long leases, so its revenue is contracted, its customers are the operators themselves, and its key metric is tenancy — how many tenants share each tower. An operator carries the consumer relationship, the tariff risk, the churn and the spectrum obligations. The tower business is an annuity; the operator business is a leveraged consumer business sitting on top of it.
Why do tariff hikes matter more in telecom than in most sectors?
Because the cost base is largely fixed. When a company with a mostly variable cost structure raises prices, part of the benefit is offset by higher costs on the extra volume. In telecom the network is already built and already paid for on a monthly basis, so an increase applied across a very large subscriber base flows through to margin with very little offsetting cost. The catch is that it only holds if competitors match it.
Where do I find monthly subscriber numbers for Indian telecom operators?
The Telecom Regulatory Authority of India (TRAI) publishes a regular subscription report with operator-wise wireless and wireline subscriber figures, additions and market shares, on its official website. It is the highest-frequency public data in the sector and it arrives well ahead of quarterly results, which is why subscriber trends are often known before the financials confirm them.
What makes one telecom balance sheet stronger than another?
Three disclosed things. First, total obligations properly counted — borrowings plus deferred spectrum liabilities plus any outstanding government dues, not reported debt alone. Second, whether operating cash flow comfortably covers interest and committed spectrum instalments without new funding. Third, where the company sits in the capital-spending cycle, because obligations that look heavy at the start of a build look very different once the network is monetising.