Power Grid: the toll bridge every electron crosses
If you owned the whole company
All of Power Grid Corporation of India can be bought today for ₹2,65,439 crore. In the year ended March 2026 it earned ₹15,928 crore of net profit on sales of ₹46,733 crore. Per ₹100 you pay, the business earns about ₹6 a year; the P/E is 16.7 — ₹16.70 for each ₹1 of annual profit — and the dividend yield is 3.14%, with the company paying out 79% of its FY26 profit in cash to owners.
Now the number that should make you sit up straight. Out of every ₹100 of sales, this company keeps roughly ₹85 as operating profit — the operating margin has run between 85% and 88% for most of a decade. Reliance keeps about ₹17; a grocer keeps ₹4. What kind of business keeps ₹85? Only one kind: a toll bridge. You built it once, at colossal expense, and now travellers pay merely to cross. I have said for fifty years that the ideal business is a toll bridge on the only road to town. Here is one, with wires instead of planks, and every electron in India's interstate power trade must cross it.
What does this business actually do?
Power Grid owns and operates India's electricity highways. Power stations make electricity in one place — a coal plant in Chhattisgarh, a solar park in Rajasthan — and people use it somewhere else. Someone must carry it across the country on tall towers and thick wires. That someone is Power Grid: India's largest transmission company, running the Inter-State Transmission System, moving big blocks of power from surplus regions to hungry cities.
It does not make electricity, and it does not sell it to your house. It is the courier in the middle, paid a regulated fee for the use of its wires — a fee set by the central regulator on a cost-plus basis, much like NTPC's: recover your costs, earn a fixed return on the equity invested in each line. Crucially, the fee is paid largely for availability — keeping the wires up and humming — not for the volume of power that flows. Rain or shine, the toll accrues.
The Government of India holds 51.34% as promoter. As always with state companies, your senior partner has other constituents besides you.
The science underneath
Charlie's section, and for once the physics is the moat, directly and beautifully.
When electricity flows through a wire, the wire resists, and the resistance turns precious energy into useless heat. The heat lost follows a simple law: it grows with the square of the current. Double the current, and you lose four times as much energy; triple it, nine times. Engineers write it as I²R — current squared, times resistance — and it is the tax nature levies on every electron delivered.
But there is a loophole, and the entire transmission industry lives inside it. The power delivered equals voltage times current. So to push the same power through a wire, you may choose: high current at low voltage, or low current at high voltage. Since losses punish current — and punish it squared — you crank the voltage as high as engineering allows and let the current fall. Raise the voltage tenfold and the same power flows with one-tenth the current and one-hundredth the heating loss. That is why Power Grid's lines run at extreme voltages — hundreds of thousands of volts on towers taller than buildings — and why, for the longest hauls, it uses high-voltage direct current, which sheds even more of the losses on cross-country runs.
Now follow the physics to the economics, in three steps a 14-year-old can repeat. One: low losses demand extreme voltage. Two: extreme voltage demands colossal, specialised infrastructure — towers, conductors, transformer stations — costing crores per kilometre, built across rivers, forests and a million land negotiations. Three: once one such network exists, building a second, parallel set of towers along the same route would be pure waste — society needs exactly one national grid, the same way a town needs exactly one road to the bridge. Economists call this a natural monopoly: the cheapest number of competitors is zero. The physics of I²R didn't just shape this business; it decreed that there would be only one.
The moat test
Hand a rival ₹2,65,439 crore and ten years of patience. Can they take the castle?
No. And this time we mean flatly, physically, legally no. They cannot duplicate the national grid — the land, the rights-of-way, the transformer stations, the operational nerve centre — and even if they somehow could, the regulator would never sanction a redundant parallel network, because every rupee of it would be waste passed on to consumers. The existing wires are irreplaceable, and the tolls on them are contracted for decades. This is as close to an unassailable castle as the stock market offers. Wide moat.
But hold the applause and invert. Two honest cracks in the ramparts. First — as with NTPC — the treasure inside the castle is rationed: the regulator grants a fixed return on equity, and return on equity has duly sat around 16-18% for a decade (16.5% last year). Guaranteed, and capped. Second, new transmission lines are now largely awarded through competitive bidding, where Power Grid must fight developers on price with no assured return. The old kingdom is safe; the new territories must be won at auction, at thinner margins. The moat protects what exists more than what comes next — and what exists has stopped growing: sales have crawled at 1% a year over three years.
The numbers Warren would check
| What we check | What it means | Power Grid |
|---|---|---|
| Sales growth, 10 yr / 5 yr / 3 yr | Bigger business? | 9% / 3% / 1% a year |
| Profit growth, 10 yr | Better business? | 10% a year |
| Operating margin | Kept per ₹100 of sales | ~85% for a decade (75% in FY26) |
| Return on equity | Profit per ₹100 of owners' money | 16.5% |
| Return on capital employed | Profit per ₹100 of all money used | 9.74% |
| Borrowings | The mortgage | ₹1,48,071 crore |
| Cash from operations, FY26 | Cash actually collected | ₹40,931 crore |
| Dividend payout, FY26 | Profit share mailed to owners | 79% |
| P/E | Price per rupee of profit | 16.7 |
Cash from operations of ₹40,931 crore against reported profit of ₹15,928 crore is the toll-bridge signature: enormous depreciation on the towers flows back as cash every year. And because the network is largely built, that cash increasingly goes to owners rather than into new steel — payout has climbed from 21% in 2015 to 79% in FY26. Borrowings of ₹1,48,071 crore dwarf owners' funds of about ₹1,00,494 crore, which is why ROCE is only 9.74% while ROE is 16.5% — the regulated-utility trick of earning a protected return on a thin equity slice atop a mountain of cheap debt. Normal for the species; know that you own the thin slice.
Equity capital rose from ₹5,232 crore to ₹6,975 crore in 2022 and to ₹9,301 crore in 2024 — bonus issues, the cake cut into more slices for the same owners, not dilution for cash.
The growth line is the sobering one. Sales: ₹45,581 crore in March 2023, ₹46,733 crore in March 2026. Profit: ₹15,417 crore to ₹15,928 crore over the same stretch. The bridge is magnificent; traffic revenue has plateaued.
What could go wrong
- Growth has stalled, and the price hasn't noticed enough. Sales growth of about 3.35% over five years, per Screener's own flag, and the stock trades at 2.66 times book value — a premium price for regulated equity whose return is capped near 16%. Pay 2.66 times book for a 16.5% ROE and your own earnings yield on purchase price is far humbler.
- Competitive bidding erodes the future. New lines won at auction carry no assured return. The wide moat guards yesterday's network; tomorrow's is a knife-fight.
- Regulatory pen-stroke risk. The toll rate exists in a rulebook the government can rewrite. A shaved return on equity flows straight to the bottom line.
- The majority owner's conflicts. The state sets the tolls its citizens ultimately pay, can nudge the company into strategic projects on nomination, and — per Screener — the accounts show a low tax rate and possible capitalising of interest, both of which flatter reported profit.
- Payment risk downstream. The tolls are billed to a power sector containing chronically stressed state utilities.
- Technology's slow flank. Cheap batteries and local rooftop generation, decades out, could reduce how much power needs hauling across the country. Not a five-year worry; a thirty-year one.
What management must do to keep the castle
The moat is wide, so the memo is short:
- Bid for new lines with discipline — walk away from auctions priced below sensible returns; monopolists make poor knife-fighters.
- Keep paying out. With the core network built, the 66-79% payout habit is exactly right; hoarded cash in a mature monopoly breeds adventures.
- Guard reliability fanatically. The regulated toll is earned by availability; the political licence is earned by never being the reason the lights went out.
- Report conservatively — expense interest where honesty suggests, and let the accounts be as boring as the business.
The verdict
Moat: wide. The dinner-table version: Power Grid owns India's electricity highways — a network physics itself declares should exist only once, which makes it the purest toll bridge in the Indian market, keeping ₹85 of every ₹100 of sales and mailing owners 79% of profits. The castle is genuinely impregnable for the next decade. But the government that grants the toll also caps it near a 16% return on equity, new bridges must now be won at auction, and traffic revenue has barely grown in three years. At 16.7 times earnings and 2.66 times book, you are paying a full price for safety rather than a bargain price for growth. A wonderful moat around a deliberately modest treasure — own it for the tolls, not for adventure.
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.