Bharti Airtel: the toll bridge made of radio waves
If you owned the whole company
Suppose you bought every share of Bharti Airtel today. The bill would come to about ₹11,72,987 crore — the market capitalisation, which is simply the price of one share multiplied by every share that exists. What do you get for that? Last year the business earned ₹33,823 crore of net profit — the money left over after every bill, every tax, every interest payment.
Divide one by the other and you get the earnings yield: roughly ₹2.9 of profit each year for every ₹100 you paid. Think of it as a lemonade stand that costs ₹100 and hands you ₹2.90 in the first summer. The stock's P/E — price divided by earnings, or how many years of current profit you're paying for — sits at 40.8. You are paying up front for a lot of future summers. The market is betting those summers get hotter. As we'll see, there's a reason it might be right — and a regulator who could rain it out.
What does this business actually do?
Airtel sells connection. Every phone call, every WhatsApp message, every cricket stream on a train platform travels as invisible radio waves between your phone and a tower Airtel built. You pay a monthly recharge; Airtel keeps the network alive. It does this for hundreds of millions of people across India, and in Sri Lanka and fourteen African countries besides.
The beautiful part: once the towers stand and the spectrum is bought, the extra cost of carrying your next gigabyte is nearly nothing. The terrible part: building and forever upgrading that network devours cash the way a teenage boy devours parathas. This one tension — cheap to run, brutal to build — explains everything about the industry.
The science underneath
Charlie here. Let's start with the physics, because the physics writes the industry's rules.
Radio spectrum is a slice of the electromagnetic dial — the same family of waves as light, just longer. Two properties matter. First, spectrum is finite and exclusive: two operators cannot shout on the same frequency in the same place any more than two batsmen can occupy the same crease. So governments auction slices, and a licence to use a frequency band is a legal monopoly on a piece of nature itself.
Second, a physical trade-off: low-frequency waves travel far and slip through walls but carry little data; high-frequency waves carry torrents of data but die within a kilometre or two and sulk at every concrete wall. So the more data your customers demand, the higher the frequencies you must use, and the more towers you must plant — cell density rises with data appetite. Each tower is steel, fibre, electronics, electricity, rent.
Now follow the money. Finite spectrum plus ever-denser towers means the price of admission runs into lakhs of crores before you earn your first rupee. That is a capex wall — a wall built of capital expenditure. Only a giant can climb it, and India learned this the hard way: a dozen operators entered, a price war followed, and the survivors are exactly three private players plus a state-run also-ran. Physics dealt the cards; economics folded the weak hands.
The chain a 14-year-old can repeat: spectrum is scarce and auctioned → carrying more data needs more towers → building costs lakhs of crores → almost nobody can afford entry → the three left standing can finally charge a fair toll.
The moat test
Hand a capable rival ₹11.7 lakh crore in cash and ten years of patience. Can they take the castle?
They'd face an interesting problem: much of the castle isn't for sale at any price. Spectrum comes up only when the government auctions it. Tower sites, fibre routes, and rights-of-way took twenty-five years to assemble. And here's the sobering part for our challenger — someone already ran this experiment. Reliance Jio arrived in 2016 with pockets as deep as any in Indian business history and gave data away nearly free for years. Look at Airtel's books for the scar: profits of ₹4,241 crore in Mar 2017 collapsed into a ₹30,664 crore loss by Mar 2020. Airtel survived; almost everyone else died or merged. The war ended not with a winner but with an oligopoly — a market of three sellers — and tariffs have marched upward since. Operating margin, the share of each sales rupee left after running costs, tells the story: 32% in Mar 2019, 57% by Mar 2026.
So which moat sources are real? Low-cost scale and the capex wall — genuinely real. Regulation — a moat and a menace at once, since the same government that limits entry also sets spectrum prices and once presented the industry with crushing retrospective dues. Switching costs — modest; number portability means a customer can defect in a day, though family plans and bundled apps add a little glue. Brand — real but thin; nobody weeps for a telecom logo.
Verdict on the moat: real, but it's a moat shared with two other castles, and the drawbridge is operated by New Delhi. That's narrow, not wide.
The numbers Warren would check
| What we check | What it means | Airtel |
|---|---|---|
| Sales growth, 10 yr | Did the top line compound? | 8% a year |
| Sales growth, 5 yr | The post-war recovery | 16% a year |
| Profit growth, 5 yr | Owner earnings trend | 24% a year |
| ROE (last year) | Profit per ₹100 owners left in | 21.9% |
| ROCE | Return on all capital, debt included | 18.5% |
| OPM, Mar 2026 | Paise kept per rupee of sales | 57% |
| Borrowings, Mar 2026 | The mortgage on the castle | ₹1,95,412 crore |
| Cash from operations, Mar 2026 | Real cash the till rang up | ₹1,22,230 crore |
| Stock P/E | Years of profit you pay today | 40.8 |
Read the story in order. Ten-year sales growth of just 8% hides a war: sales actually shrank from ₹96,101 crore in Mar 2015 to ₹80,780 crore in Mar 2019, then doubled to ₹2,10,973 crore by Mar 2026. Cash from operations — the actual cash the business collects, harder to fudge than reported profit — grew from ₹28,059 crore to ₹1,22,230 crore over the same stretch, and comfortably exceeds net profit. That's the sign of a genuine cash machine, not an accounting one.
Now the debits. Borrowings of ₹1,95,412 crore remain enormous, though they're finally falling from the ₹2,26,020 crore peak of Mar 2023. Equity capital rose from ₹1,999 crore to ₹3,047 crore — the company printed new shares during the crisis years, diluting old owners; every slice of the pizza got a bit smaller. Promoters hold 50.07%, down from 54.75% in late 2023. Dividend payout last year was 55% of profit, and Screener notes a healthy 48% habit. Screener's cons: the stock trades at 7.84 times book value, and promoter holding fell 4.95% over three years.
What could go wrong
Invert, always invert. What kills this business?
The regulator. Spectrum prices, licence fees, retrospective demands — the state giveth the oligopoly and the state can tax it to death. It nearly did once; the Mar 2020 loss of ₹30,664 crore carries the fingerprints.
The rival with infinite patience. Jio has started price wars before. If it values market share over profit again, that lovely 57% margin is the first casualty.
Technology leapfrog. Satellite constellations beaming broadband from orbit could, over a decade, hop over the tower network entirely. We don't know how fast or how cheap. We genuinely don't know.
The treadmill. 4G, then 5G, then 6G — each generation demands another mountain of capex. A toll bridge you must rebuild every seven years is a poorer bridge than the one over the Missouri.
The price. At a P/E of 40.8 and 7.84 times book, the market has already RSVP'd for a decade of good news. Pay a rich price and the business must be right and the crowd's expectations must be met. That's two hurdles, not one.
What management must do to keep the castle
- Keep cutting debt. ₹1.95 lakh crore of borrowings is a mortgage that must shrink every single year the sun shines.
- Never start a price war; never be afraid to follow a price rise. In a three-player market, discipline is the dividend.
- Milk the premium customer — data, enterprise, home broadband — where switching is stickiest.
- Spend 5G capex like it's your own money, because it is the owners'.
- Keep the regulator relationship boring. No surprises, no arrears, no headlines.
- No more equity dilution. The share count grew in the lean years; it should not grow in the fat ones.
The verdict
Narrow moat, and we mean that as a compliment hard-earned. The physics of spectrum and tower density built a capex wall that reduced a dozen contenders to three, and Airtel stands among the survivors with margins at 57% and cash from operations of ₹1.22 lakh crore rolling in. Over five to ten years the oligopoly should hold, because nobody sane will fund a fourth network. But this castle has two landlords — a regulator with a history of surprise rent hikes, and a rival with the deepest pockets in India — and the entry ticket today costs 40.8 years of current earnings. Wonderful escape from a terrible industry; just remember that the market has already priced the escape. Quality of business is not an invitation to pay any price for it.
Written in the style of Buffett & Munger for the Omaha Investments book project. Educational material, not investment advice. Numbers from Screener.in and live NSE data via Angel One as of the date above.