How to Analyse a Power Utility Stock (PLF, RoE, Merchant)

In FY25, NTPC's coal stations ran at a plant load factor of 77.44% while the rest of India's coal fleet managed 67.23% — a ten-percentage-point lead, and the company's highest PLF in seven years (Inve data, Q4 FY25; NTPC FY24 investor presentation confirms the same "consistent lead over All-India in PLF"). A casual reader sees a number near 77 and thinks the plant is running flat out. It is not. A new supercritical coal unit can physically run above 90%. The gap between what NTPC's plants can do and what they did do is not inefficiency — it is demand and despatch, decided by load-despatch centres and the discoms that buy the power, not by the plant manager. (Illustration of how to read the numbers, not a view on the stock.)

That one fact reorganises the whole sector. A power utility does not really sell electricity the way a steel mill sells steel. The regulated ones — NTPC, NHPC, most state generators — earn a return on the capital they sunk into the plant, set by a regulator, and they collect it whether the plant runs hard or sits idle, so long as the plant is available. The merchant ones — and the merchant slice of the hybrids like JSW Energy and Torrent Power — sell uncontracted power at a spot price that swings with the season. These are two completely different businesses wearing the same "power" label, and the single most common error in the sector is valuing one as if it were the other.

This is how to read a power utility the way an infrastructure analyst does: the handful of operating numbers that decide the outcome, where most of them are buried (not on the income statement — in the concall and the investor deck), what "good" looks like, and the multiple that fits each kind of utility. A boundary first: from the outside you will not rebuild a tariff order. What you can do is read whether the capital is regulated or merchant, whether the plant is available, and whether the cash is actually coming back from the people who bought the power.

Why is a regulated utility a "return on assets" business, not a "sell more units" business?

Picture a toll road built under a contract that pays you a fixed return on what the road cost to build, provided the road stays open — whether ten thousand cars cross or two thousand. That is a CERC-regulated power plant. The regulator lets the generator earn a normative return on equity on its approved capital base, plus pass-through of fuel and a few other costs, as long as the plant clears an availability bar. Run the plant harder and you earn a small incentive; run it less and, crucially, you keep the return — until availability slips below the norm, at which point the fixed-cost recovery starts to leak.

That framing fixes the beginner error of cheering revenue or generation growth. For a regulated utility, the engine is regulated equity — the capital base the return is earned on — and it grows by commissioning new plants, not by despatching more units from old ones. NTPC's consolidated regulated equity grew from roughly ₹90,902 crore (standalone, Q4 FY25) to ₹1,18,970 crore (consolidated, Q3 FY26) as new capacity came online (Inve data). That is the number that compounds the earnings, and it sits in the investor presentation, not the P&L. The income statement just shows you a thin, stable slice off the top — NTPC's EBITDA margin runs around 87–90% (Inve data, Q3 FY26 / Q4 FY25), which looks absurd for an industrial company until you remember fuel is a pass-through, so the "margin" is really regulated return plus depreciation on a huge asset base.

The merchant business is the opposite animal: no assured (regulated) return, price set by the market, and the value lives in how much of the fleet is uncontracted when prices move. Across the listed space, SJVN and NLC India sit toward the regulated end, while renewable pure-plays like Adani Green and NTPC Green live mostly on long-term PPAs — each at a different point on the regulated-to-merchant spectrum. Hold both pictures at once and the sector stops being confusing.

The metrics that matter — and where they hide

Most of what decides a power utility never appears on the income statement. Here is the spine, metric by metric.

PLF and plant availability factor — and why they are not the same number

Plant load factor (PLF) is how much a plant actually generated against its theoretical maximum over a period. Plant availability factor (PAF) is whether the plant was available to generate if called. For a regulated utility this distinction is the whole ballgame: the regulated return is tied to availability, not to load. A plant can have a low PLF (nobody despatched it) and still earn its full fixed cost if availability stayed above the norm.

Where to find them: PLF and PAF are almost never on the income statement. They live in the investor presentation and the concall — and frequently you have to read the transcript to get the PAF and the reason behind a move. NHPC's plant availability factor for FY25 came in at 73.94%, down about four points from 77.60%, and the transcript spells out exactly why: "mainly due to lower water availability, complete shutdown of Teesta-V Power Station and outages of units at TLDP-III, Uri…" (NHPC Q4 FY25 concall) — the Teesta-V station had been knocked out by a flash flood in the Teesta basin in October 2023. That is a hydro-specific lesson: a single asset going dark can drag the whole fleet's availability, and the income statement won't tell you why — the call will.

What "good" looks like: for a regulated thermal generator, availability comfortably above the CERC norm (historically ~85% for full incentive) and a PLF that beats the all-India average is the sign of a well-run fleet. NTPC's 77.44% coal PLF against 67.23% for the rest of India (Inve data, Q4 FY25) is exactly that signature. For a merchant thermal player, watch PLF directly — it is the revenue. (Illustration, not a view on any stock.)

Installed and under-construction capacity — the growth pipeline

Capacity is the one number that is easy to find, and the one investors over-rate in isolation. Installed capacity tells you the size today; capacity under construction tells you the growth. NTPC reported total group capacity of 86,565 MW (Q3 FY26) with roughly 16.5–17 GW under construction (Q1 FY26), and a long-dated target of 149 GW by FY32 — revised up from an earlier 130 GW plan — and 244 GW by FY37 (Inve data, Q1 FY26). The caveat: under-construction MW only matters once you know whether it is regulated (return locked in on commissioning) or merchant (return still hostage to the spot price). Capacity without that label is a vanity number.

Regulated RoE vs merchant exposure — the most important single distinction

This is the metric that decides which valuation lens to use. A pure regulated utility earns the CERC normative return on equity (the base norm has long been set at 15.5%, with a small extra for timely commissioning) on its regulated-equity base. A merchant player earns whatever the market pays. Most large utilities are now hybrids, and the number to hunt for is the share of capacity that is merchant or "open" — uncontracted, exposed to spot prices.

That number is buried in the concall. JSW Energy put it plainly: "Currently, our open capacity stands at approximately 8%, which will further reduce to around 5% from April 1, 2026, following the securing of a 400 MW 25-year PPA for our Utkal plant…" (JSW Energy Q2 FY26 concall). Read that as a company deliberately dialling down its merchant exposure by signing long-term PPAs — trading upside in a high-price year for earnings you can underwrite. The lower the open capacity, the more "regulated-like" the earnings, and the higher the multiple the market will pay for them.

Fuel cost and availability — the pass-through that isn't always clean

For a regulated thermal plant fuel is a pass-through, so headline margin is insulated from coal prices. But there is a trap hiding in the despatch: when a plant runs below the level the tariff assumed, its fixed costs are under-recovered — the "disincentive." This is the number management would rather you skim past, and it is only in the call. NTPC guided in Q2 FY25 that it hoped to bring FY25 coal under-recovery into "the range of Rs.250 crore to Rs.300 crore" (NTPC Q2 FY25 concall). The actual, disclosed at year-end: "the disincentive would be INR464 crore for the year as a whole" (NTPC Q4 FY25 concall). The guidance was missed by half again — not a scandal for a company this size, but a clean example of why you read the under-recovery line and not just the pass-through assumption. (Inve's Promise Tracker marks that FY25 under-recovery guidance missed at ₹464 crore, and the separate FY26 ₹250 crore target as at risk, with ₹454 crore already accumulated on the FY26 line by December 2025 — Inve data, Q3 FY26. Note the two figures belong to different fiscal years: ₹464 crore is the closed FY25 number, ₹454 crore is the running FY26 total.)

Discom receivables — where a profitable utility quietly bleeds

Here is the number that has done more damage to power-sector returns than any operating miss: the utility books the sale, recognises the profit, and then waits — sometimes for years — to be paid by financially stretched state distribution companies (discoms). Profit on paper, cash stuck in receivables. So read receivable days against the regulatory norm. NTPC's receivable days improved to 26 by Q3 FY26, "vs regulatory norm of 45 days" (Inve data; the company explicitly benchmarks itself to the 45-day norm, Q1 FY26 concall). A utility collecting inside the norm is being paid; one whose receivable days are climbing is financing its customers with shareholder money, and no profit ratio shows it. For the merchant-leaning names the equivalent is debtor days — JSW Energy reported 73 (Inve data, Q2 FY26). The instinct here is the same one that separates reported profit from real cash earnings in any business.

How do you value a power utility — and why the multiple depends on the mix?

There is no single right multiple for "power," and using one is the costliest mistake in the sector.

For the regulated business, use price-to-book against return on equity. The earnings are a regulated return on a known equity base, so the book value is the engine and the question is simply: what return is being earned on it, and how fast is the base growing? A regulated generator earning its ~15.5% normative RoE and steadily expanding regulated equity is a book-value compounder; you pay a multiple of book scaled to how far the RoE sits above the cost of equity and how reliably the regulated-equity base grows. This is the same P/B-vs-RoE logic you'd bring to a bank or NBFC, for the same reason — the balance sheet, not the sales line, is the business.

For the merchant business, use EV/EBITDA — and normalise it. Merchant earnings swing with the spot price, so a single year's EBITDA can flatter or scare you. The trap: capitalising a peak-price year's EBITDA at a fat multiple, as if it were a regulated annuity. Normalise across the cycle, then discount for the receivables and merchant-price risk you are taking on. The more of a hybrid's capacity that is open/merchant, the more its valuation should lean on this lens — and the wider the band of outcomes.

The practical move for a hybrid: split it. Value the regulated/PPA-tied slice on P/B-RoE, the merchant slice on a normalised EV/EBITDA, and add them — rather than slapping one blended multiple on a business that is really two. JSW Energy's deliberate march from 8% to 5% open capacity (JSW Energy Q2 FY26 concall) is, in valuation terms, the company shifting weight from the second lens to the first.

A worked case: when the PLF told the story before the profit did

Torrent Power is a hybrid — regulated distribution, plus thermal and renewable generation with a merchant slice — which makes it a clean teaching case for how a merchant-exposed plant's economics can lurch in a single quarter (illustration, not a view on the stock; figures from Inve data and the company's concall).

In Q1 FY26, Torrent's thermal plant load factor "significantly deteriorated to 39% in Q1 FY26 from 60% in Q1 FY25" (Torrent Power Q1 FY26 concall). That is a 21-point collapse in one year. For a regulated plant that would barely dent earnings — the return is on availability, not load. For the merchant portion, PLF is the revenue: fewer units despatched at a thinner spread. The same quarter, merchant sales were 714 million units contributing ₹327 crore at an average realisation of about ₹4.50 a unit (Torrent Power Q1 FY26 concall) — and management noted it had locked a 300 MW merchant block under SECI-XVIII at ₹3.97 a unit, below that ₹4.50 spot realisation. Read those two numbers together and you can feel the merchant trade-off: ₹4.50 today with all the volatility, or ₹3.97 contracted and predictable. The plant that ran at 39% is the one whose economics you cannot annualise from a good quarter.

Now the said-vs-did, the part the spreadsheet hides. Torrent guided in Q4 FY25 to transmission capex of "Rs. 900 to Rs. 1000 crores" — Inve's Promise Tracker marks that one missed — while its distribution-capex and total-capex guidance held on track (Inve data). On the regulated side it delivered; on a discretionary line it didn't. That is the texture you only get by checking guidance against what followed, across quarters, rather than trusting any single confident call. The honest read is not "management over-promised" — it is that a hybrid utility has regulated commitments it almost always keeps and merchant/discretionary ones that move with conditions, and you should weight them differently. (A read on how management communicated through recent quarters, not a lifetime verdict.)

This is precisely the work Inve's Promise Tracker is built to do — pin every forward commitment to the quarter and quote it was made in, then mark it as later calls confirm or contradict it, so you see the sequence rather than re-reading a year of transcripts by hand.

Red flags specific to power utilities

• Rising receivable days while profit grows. The classic discom trap: earnings up, cash stuck with state distribution companies. A utility drifting past the ~45-day regulatory norm is financing its customers — invert the income statement and look at whether the cash actually arrived.
• A high blended multiple on a merchant-heavy book. Capitalising a peak-price merchant year's EBITDA as if it were a regulated annuity is how investors overpay at the top of a price cycle. Demand the open-capacity number before you accept the multiple.
• Availability slipping below the CERC norm. For a regulated plant the danger isn't low PLF — it's low availability, because that is where the fixed-cost under-recovery (the disincentive) bites. NTPC's ₹464 crore FY25 under-recovery against a ₹250–300 crore guide is the visible edge of that mechanism.
• A single asset carrying the fleet's availability. NHPC's FY25 PAF dropped largely because one station, Teesta-V, was shut by a flood. Concentrated hydro or large single units mean a single outage can swing group numbers — read the asset list, not just the total.
• Under-construction MW with no PPA or regulated tariff named. Merchant capacity dressed up as growth. Capacity is only worth what its offtake contract is worth.

Where this lens can be wrong

The strongest case against everything above is that the regulated model is too comfortable a story. A reader who learns to trust regulated RoE and waves through availability and receivables can still be blindsided two ways. First, regulation is not permanent: tariff norms get reset, the normative RoE can be revised, and a pass-through can be challenged — the "annuity" is only as durable as the next CERC order, which no transcript can predict. Second, the merchant slice that looks like risk in a soft year is upside in a tight one; a reader who always discounts merchant exposure will systematically underpay for the hybrids in the years power is short, which is exactly when they earn the most. The honest claim is narrower than it looks: reading availability against under-recovery, and receivable days against the norm, tells you whether this utility is being run and paid well right now. It does not tell you where the spot price or the next regulatory order is going — and on those, the ceiling on anyone's confidence is low.

A hard limit worth stating plainly: we have not modelled any company's tariff order or its discom counterparty credit from the outside, and you can't fully do so either. A clean availability record and tidy receivables do not immunise a utility against a state-level payment crisis or a regulatory reset that takes good and weak operators down together.

A repeatable workflow

1. Split the book. What share is regulated/PPA-tied versus merchant/open? The open-capacity figure (in the concall) decides everything that follows.
2. Check availability, then load. PAF against the CERC norm for the regulated part; PLF directly for the merchant part. Both live in the deck and the call, not the P&L.
3. Read the under-recovery line. Is fixed-cost recovery clean, or leaking via the disincentive? Only the transcript has it.
4. Follow the cash. Receivable days against the ~45-day norm — is the profit being collected, or parked with the discoms?
5. Pick the lens. P/B-RoE for the regulated slice; normalised EV/EBITDA for the merchant slice; never one blended multiple.
6. Audit the commentary. Check capacity-commissioning, capex and under-recovery guidance against what actually happened, quarter by quarter.

Inve's KPI Screener lines up PLF, capacity, regulated equity and receivable days across utilities — with the value, trend, and a data-confidence flag per number — so step 2 takes minutes, not an afternoon. And the concall summaries pull every forward commitment into one guidance table per quarter, with speaker and quote, which is where the buried operating KPIs in this sector actually live.

Frequently asked questions

The discipline comes down to refusing to read a power utility like a manufacturer. The units it generates are not the business; the return on the capital, the availability that protects it, and the cash that comes back are. So invert the question you bring to a utility's results. Don't ask "did generation grow?" Ask: if this utility were quietly earning a paper profit it can't collect, or annualising a merchant price that won't last, what would the numbers look like — and do the receivable days and the open-capacity figure rule it out? A climbing receivable line under a growing profit, or a fat multiple on a merchant-heavy book, does not rule it out.

And the owner's question, the one to sit with before buying a single share of any power company: across the next regulatory cycle and the next soft year for spot prices — not this quarter's PLF — what must I believe about this utility's availability, its regulated base, and its ability to actually get paid, for it to still be compounding on the other side? If the answer leans on a good merchant year repeating, you've valued the cycle, not the business.

Related reading

• How to Analyse a Bank Stock (NIM, CASA, GNPA, ROA)
• How to Analyse a Building Products Company (Pipes, Tiles)
• How to Analyse a Cables and Wires Stock