← all companies · 2026-07-07 · moat: narrow · raw .md

NTPC: Heat, Steam, Torque, Current

The loudest room in India

There is a room in Uttar Pradesh roughly the size of a football field, and inside it a single steel shaft, thick as a tree trunk and longer than a cricket pitch, spins fifty times every second. It has been spinning, with occasional pauses for maintenance, for decades. The room is a turbine hall; the shaft connects a steam turbine to a generator; and the number fifty is not an engineering accident — it is the frequency, in cycles per second, of every alternating current socket in India. That shaft, and a few hundred like it, are the fifty hertz. When your ceiling fan turns, it turns because somewhere a shaft like this one is turning, and the grid's wires have married their motions together.

The company that owns more of these halls than anyone else in India is NTPC — founded as the National Thermal Power Corporation — which generates roughly a quarter of the country's electricity. But here is the odd thing you must hold onto for this whole chapter: NTPC's fortunes are governed less by the magnificent machine in that hall than by a paragraph of regulation — a formula, published by a tariff commission, that decides in advance almost exactly how much money the machine is allowed to make. The machine is physics; the paragraph is economics; NTPC lives where they meet. Take them in that order.

Boiling water, the hard way

Strip away the scale and a coal power station does something a kitchen does: it boils water. Everything else is refinement, and the refinements are where the money and the physics live.

The sequence is a loop, and engineers call it the Rankine cycle. First, pump: water is squeezed to enormous pressure — two hundred atmospheres and more — while still liquid, which takes surprisingly little energy, because liquids are nearly incompressible. Second, boil: the high-pressure water runs through kilometres of tubing lining the walls of a furnace where pulverised coal — ground finer than talcum powder so it burns like a gas — roars at over a thousand degrees. The water flashes to steam and is then superheated to around five hundred and fifty or six hundred degrees Celsius: an invisible, furious gas storing energy in both its temperature and its pressure. Third, expand: the steam is released through the blades of the turbine, stage after stage, each row of blades extracting a push the way a sailboat extracts push from wind, spinning the shaft — heat has become torque. The generator on the same shaft is the final trick: spin a magnet inside coils of copper and, by Faraday's law of induction, the moving magnetic field drags electrons back and forth in the wire. Torque has become current. Fourth, condense: the exhausted steam is cooled back to water — this is what those giant waisted cooling towers do — and returns to the pump, closing the loop.

Why condense at all? Why throw heat away on purpose? Because of the deepest rule in thermodynamics: a heat engine only produces work while heat flows downhill, from hot to cold, and the fraction of heat you can convert to work is capped by how far apart the hot and cold ends are. Carnot proved the ceiling two centuries ago: no engine, however clever, can beat the limit set by its temperatures. A modern plant's furnace end runs as hot as its steel can bear and its cold end as cool as river water allows, and the arithmetic lands where it lands: the best coal stations convert a little over forty per cent of the coal's energy into electricity. The rest — the majority — leaves as warm air and warm water. That is not sloppiness; it is the tax the second law of thermodynamics levies on every fire, and it is why each percentage point of efficiency is a prize worth billions: it means burning visibly less coal for the same fan turning in your room.

The paragraph that tames the fire

Now the second machine — the regulatory one — and here NTPC becomes genuinely unusual among the companies in this book.

Electricity in bulk is the purest commodity imaginable: an electron is an electron. A generator has no brand, no loyalty, no pricing power. Left to raw markets, power generation is a brutal business — enormous fixed costs, a product that cannot be stored, prices that spike and crash hourly. India chose, for its central generating stations, a different arrangement, and you should understand it the way you would understand a bond.

NTPC sells almost all its power through long-term contracts under a tariff set by a national regulator, and the tariff has two parts. The energy charge simply passes the fuel bill through to the buyer: if coal costs more, the customer pays more, and NTPC neither gains nor loses. The capacity charge is the interesting part: it repays, on a schedule, the money that built the plant — depreciation, interest on debt, operating costs — plus a return on equity fixed in advance by the regulator, currently of the order of fifteen and a half per cent, post-tax, on the equity deemed invested in the plant. And the capacity charge is earned not for producing power but for being available to produce it: keep the plant ready around eighty-five per cent of the time and the return flows whether or not the buyer draws the power.

Read that formula twice and its two faces appear. It caps the risk: fuel costs pass through, demand risk falls on the buyer, and the return arrives in recession and boom alike — this is why NTPC's profits march with an almost bond-like gait. And it caps the dream: no brilliance, no efficiency heroics, no bull market in power can ever earn the shareholder much more than the regulator's number. NTPC is that rare thing, a company whose best case and worst case were both written down by someone else. Growth, therefore, can come from exactly one source: building more regulated equity — more plants under the formula. The company is, financially speaking, a machine for converting construction into annuities.

Blackouts built this company

Why does such a company exist? Because in 1975 India was a country of blackouts. Generation was the job of state electricity boards, which were underfunded, politically squeezed, and chronically behind demand; the economy was rationing evenings. The central government created NTPC to do what the states could not: build very large, modern coal stations at the coalfields themselves and ship the power outward over high-voltage lines. The first unit came alive at Singrauli in 1982, and the company then did the one thing India's power sector had never managed — it executed, decade after decade, adding giant projects with a reliability that made it the natural home for every subsequent national generation programme: gas stations, hydro, solar parks, and now nuclear ambitions. Along the way it became the country's largest power producer and was entrusted with the "Maharatna" autonomy given to the most consequential state enterprises.

Notice what really happened there: the government did not just found a company; it founded a counterparty. The regulated-return formula only works if someone credible stands ready to build endlessly under it. NTPC is that someone.

Why no one races a regulated giant

The barriers here are real but subtle, because the moat is not technology — turbines are bought from a handful of global engineering firms, and anyone may purchase one. The barriers are:

Scale of capital on terms no private rival enjoys. A single large plant consumes tens of thousands of crores, and NTPC's borrowings stand at ₹2,71,005 crore — a number that would signal distress at a private firm and here signals the business model working as designed: the formula assumes seventy per cent debt, the consumer services it through the tariff, and the sovereign's shadow behind a 51.10 per cent state shareholding keeps the debt cheap. Sites, water, and coal linkages accumulated over fifty years — a thermal station needs a river, a railhead, and a mine's allegiance, and the good combinations are taken. An execution organisation that has actually delivered projects of this size on Indian ground, which is a rarer asset than any licence. And the long power purchase agreements themselves: state utilities are contracted to NTPC plants for decades ahead, leaving little open demand for an aspiring rival to serve.

This is a narrow moat — wide enough that no one has displaced the company in fifty years, but built of contracts and position rather than pricing power, and its returns are confiscated in advance by design.

Charlie reads the tariff order

Munger's method is to hunt for the incentive baked into the structure and then predict the behaviour it must produce. Apply it to the formula and something jumps out immediately: a company paid a fixed return on capital deployed is being paid to deploy capital. Not to economise on it — to deploy it. Every additional crore of regulated equity earns the regulated return; a crore saved earns nothing. Economists have a name for the resulting disease in regulated utilities worldwide — the tendency to gold-plate, to prefer the grander project, the earlier replacement, the larger balance sheet. The regulator's audits push back, but the gradient never changes. An owner of NTPC must simply accept this: the company's compounding is real but administered, and its capital discipline is enforced from outside, not chosen from within.

Second, the owner-conflict — and it wears a different face here than at the other state companies in this book. The government does not suppress NTPC's price; the formula protects it. The conflict arrives through assignment and through receivables. Assignment: the majority owner decides what NTPC builds and where — plants sited for regional politics, stressed private plants absorbed for system stability, coal stations kept running for energy security past their economic prime. NTPC's capital goes where the republic needs it, which is not always where returns are best. Receivables: NTPC's customers are state distribution companies — owned by the same political class — which historically pay late when their own finances buckle. The formula promises the return; collecting it is a permanent diplomatic mission. Cash from operations of ₹50,902 crore last year against profit of ₹27,546 crore shows collection ultimately works, but the float is forever hostage to the health of the weakest state utility.

Third, the ten-year to-do list, and for NTPC the energy transition reads differently than for any coal miner or oil driller: the formula does not care what makes the steam. NTPC's true asset is not its furnaces — it is its licence to build under the regulated-return regime, its sites, its grid connections, and its execution machine. All of it transfers. Solar farms, pumped-hydro storage that banks afternoon sun for the evening peak, and — the boldest line on the list — nuclear stations, where the Rankine cycle survives intact and only the fire changes: the same steam, the same turbines, heat from splitting uranium instead of burning carbon. The decade's assignment is therefore a race of books: grow the clean regulated book faster than the coal book ages, keep availability high so the annuity never stutters, and prove to the regulator at each five-year reset that the allowed return is being earned honestly. If Munger were grading management, the single metric would be: what fraction of new regulated equity, each year, needs no coal?

The owner's meter reading

Now buy the whole company, Buffett-fashion: the market asks ₹3,45,104 crore — 12.7 times last year's profit of ₹27,546 crore, with a 2.35 per cent dividend yield on a payout of about 32 per cent.

What you own is the closest thing the Indian stock market offers to an infrastructure bond with a growth clause. Profit has climbed from ₹9,992 crore in Mar 2015 to ₹27,546 crore in Mar 2026 — about 10 per cent a year over the decade, accelerating to 17 per cent over three years — and earnings per share from ₹10.09 to ₹27.90 with only modest changes in share count. Operating margins sit calmly in the high twenties (28 per cent last year, 29 the year before), because the formula makes them sit calmly. Return on equity is 14 per cent — just under the regulated allowance, the gap being the friction of projects under construction and the odd under-recovery. Predictability, the Buffett virtue that usually costs a fortune, is here almost free: you can sketch this company's earnings five years out with more confidence than any other business in this book, because a regulator has pre-written them.

What you give up is everything above the line. There will be no surprise year when margins double, no franchise pricing, no operating leverage miracle. Owner earnings are further squeezed by the model itself: the company must perpetually reinvest — borrowings have grown from ₹1,02,252 crore to ₹2,71,005 crore across the decade funding the build-out — so the dividend, not retained compounding, is a large share of the realised return. The 33 lakh shareholders on the register — up from under 10 lakh three years ago — have evidently decided that a pre-written 14 per cent, in a country still short of electricity, is a fair trade for the absent dream. It is a defensible trade. Just be sure you are buying it for what it is: a formula with a construction company attached.

Where the formula cracks

The structural risks all attack the paragraph, not the machinery:

Regulatory reset. The allowed return is re-decided every five years, under permanent political pressure to cheapen electricity. A durable cut of even a couple of percentage points re-prices every plant at once. The cap that protects also binds.
Stranded fires. The coal fleet's later decades may collide with carbon policy, cheap storage, and dispatch rules that idle coal first. The formula pays for availability today; a future regulator can redefine what deserves to be paid for.
Counterparty rot. The state distribution companies are the system's chronic invalid. NTPC has always been paid eventually; "eventually" is the risk.
Directed capital. The majority owner can assign projects a private board would refuse. Fifty years of history say this recurs.
Execution at frontier scale. Nuclear and giant storage are new engineering ground; the regulated return assumes projects finish near budget, and frontier projects are famous for not doing so.

New fires, same formula

Stand in that turbine hall once more, but set the year to 2050. India by then will need perhaps three times today's electricity — the surest demand-growth story on earth, driven by air conditioners, factories, electric transport, and the data centres of the machine-intelligence age. The evening peak will be met by sunlight banked in batteries and pumped hydro; the deep steady base, in any credible Indian scenario, by nuclear steam. Which is the quiet punchline of this chapter: the Rankine cycle — pump, boil, expand, condense — will outlive coal comfortably. Uranium makes steam. Some of the halls will fall silent; the kind of room will not.

Whether NTPC captures that future is not a physics question. The physics transfers. It is a question of whether an organisation built to pour concrete under a guaranteed return can learn new fires faster than its old ones are retired — and whether its owner, who writes the formula, keeps writing it generously enough to make the construction worthwhile. The student's takeaway travels well beyond power: wherever you find a business whose profits are set by formula, study the formula-writer's incentives before the business's — because the pen that caps the dream also holds up the floor, and both can move.

An Omaha Investments chapter. Educational material, not investment advice. Figures from Screener.in and NSE data via Angel One as of the date above.