Capacity Utilisation: Pre-Capex Signal for Stocks

In the Q1 FY26 concall, HEG's management said something that, in hindsight, was the whole story. The graphite-electrode plant was running at "over 90% capacity utilization on its 100,000-ton expanded capacity" (HEG Q1 FY26 concall) — and in the same breath, the company "recently announced another expansion plan to increase our existing capacity from 100,000 tons to 115,000 tons, which will require a capex of about INR650 crores" (HEG Q1 FY26 concall). The 90% came first. The ₹650-crore capex was just the paperwork catching up. (HEG is used here only as an illustration of how to read the metric — not as a view on the stock.)

Capacity utilisation is the most quietly predictive operating KPI in manufacturing, and it is hiding in plain sight on every concall. It is the pressure gauge of a factory-based business: how much of the plant's installed capacity is actually being used. Read it right, and you can often see a capex cycle — and the revenue it eventually brings — coming a couple of quarters before management formally commits to it. This is a field guide to reading the tell, built on companies that lived it.

What is capacity utilisation?

Capacity utilisation is the ratio of what a plant is actually producing to what it could produce if it ran flat-out.

A plant rated for 100,000 tonnes a year that produces 80,000 is running at 80% — exactly where HEG sat for FY25, on its own description: "Our capacity utilization for the year 2024-25 was close to 80% based on our expanded capacity of 100,000 tons" (HEG Q4 FY25 concall). The remaining 20% is headroom — capacity the company already paid for, sitting idle, waiting for demand.

That headroom is the key to why the metric leads. When utilisation is low, rising demand can be met for free — the plant simply runs harder, and because the fixed costs (the factory, the machinery, the core staff) are already paid for, every extra tonne is highly profitable. But headroom is finite. As utilisation climbs toward the practical ceiling — rarely 100%; maintenance, changeovers, and bottlenecks usually cap it, as a general manufacturing rule of thumb, around 85–95% depending on the industry — the company runs out of room. To grow further, it must add capacity. That means capex. And capex means a new line that, once commissioned, drives the next leg of revenue.

One thing to keep in the back of your mind before trusting any utilisation figure: it is almost always self-reported by management in the Q&A, not audited and not standardised. The person quoting "over 90%" is the same person who wants you to believe a capex is justified and growth is coming — so the gauge you're being handed is an incentivised number, often unaudited, and defined however that management chooses to define it. Useful, but supplied by an interested party; weigh it accordingly.

So the sequence is mechanical: rising utilisation → approaching the ceiling → capex announcement → new capacity → future revenue. The investor who watches utilisation sees step one. The investor who waits for the revenue sees step five — years late, and after the price has moved.

One caveat up front, because the sequence is seductive: high utilisation is not a buy trigger. Knowing a capex cycle is likely tells you where a company sits in its cycle — not whether the stock is cheap. The market often prices the coming expansion in long before the line is commissioned; the capex can be mistimed into oversupply so the new capacity arrives just as demand softens; and, as the HEG example below shows, a full plant can still lose money if prices fall. Reading the gauge earlier than the crowd is an information edge, not a valuation edge. Treat what follows as a way to understand a manufacturer's position, not a signal to front-run its capex.

Why does high utilisation precede capex guidance?

Because a rational management does not announce a major expansion until its existing assets are nearly full — building capacity ahead of demand is expensive and risky, so most managers wait for the evidence that demand is real and durable. That evidence is high utilisation. By the time a plant is running at 90%-plus across several quarters, the case for expansion has made itself, and the capex announcement becomes almost a formality.

HEG is the textbook walk. Its own utilisation, read across the calls Inve parsed, escalated quarter by quarter — and the expansion announcement landed exactly when the gauge crossed the line.

All figures and quotes: HEG concalls Q3 FY25–Q4 FY26, as parsed by Inve.

Read the gauge and the language together and the capex is visible a quarter early. Utilisation crossed from 80% to "90% plus" between Q4 FY25 and Q1 FY26 — and the 115,000-tonne, ₹650-crore expansion was announced in that same Q1 FY26 call, not a quarter later in a press release. The investor watching utilisation saw it form; the investor waiting for the new line's revenue will wait until early calendar 2028, when production is slated to begin (HEG Q2 FY26 concall) — nearly three years after the tell.

But that clean walk — utilisation up, capex announced, revenue eventually lands — is the trophy version, and showing only it would be dishonest. The same tell has a graveyard. The far more common path is the one discipline #4 below concedes: utilisation climbs, the capex is duly announced, and then the project slips — the cost creeps, the commissioning date moves out, and at some point the line simply stops being mentioned, with the expected revenue never arriving. Even within HEG's own calls, the anode project (below) drifted ~10% on cost across three quarters; Dixon's refrigerator expansion later in this piece inflated three times and then went silent entirely. So read the table as one outcome out of several, not the guaranteed one. A 90% reading tells you a capex is likely — it tells you nothing about whether that capex will be built on time, on budget, or into demand that is actually there.

Where utilisation goes quiet: the capex you can't see coming

The inverse is just as useful, and easier to miss. Falling utilisation — or, more often, a brand-new plant that won't fill up — is an early demand warning that doesn't show in headline revenue for a couple of quarters.

Timken India is the cautionary version of HEG. It built a new bearings plant at Bharuch, then spent five quarters cutting its own ramp target as the demand it had built for failed to arrive on schedule. In Q2 FY25 management guided "to reach 80% capacity utilization within 18 months of commissioning" (Timken Q2 FY25 concall). By Q4 FY25 that had been pared to "around 45-odd percent… by the end of FY26" (Timken Q4 FY25 concall). And the actual number, a year on, was lower still: "Management expects utilization to climb from the current 30% to over 50% by Q1 FY27" (Timken Q3 FY26 concall). Eighty became forty-five became a real thirty. The capex was already sunk; the demand case it rested on kept getting smaller. (Illustration of the metric, not a view on Timken.)

A new plant whose utilisation guidance only ever moves down is telling you the same thing falling utilisation tells you at an old plant: the demand wasn't where the spending assumed it was. That is how manufacturers destroy capital — by building into a forecast and discovering the forecast was the most fragile number in the deck.

How do you read utilisation without getting it wrong?

Utilisation is a powerful tell, but it rewards skepticism. Five disciplines from the field.

Watch the trend across quarters, not a single number. One quarter at 90% can be seasonal. HEG's signal wasn't the first 90% — it was 80% sustained through FY25, then 90%-plus held across the whole of H1 FY26 ("achieved over 90% capacity utilization in the first half of FY26," HEG Q2 FY26 concall), with the expansion language hardening alongside it. Sustained is the word that matters.

Distinguish a capex of expansion from a capex of desperation. Capex announced into rising utilisation — HEG's, into 90%-plus — is a growth signal. A plant whose ramp targets keep getting cut — Timken's Bharuch, from 80% to a real 30% — is the opposite, and management will rarely frame it that way. Read what the spending is actually being justified by: a full plant, or a hopeful one.

Respect the realistic ceiling. "100% utilisation" rarely exists. Each industry has a practical maximum. HEG itself called 90%-plus a high number and guided "maybe between 85% and 90%" for the full year even while running hot (HEG Q2 FY26 concall) — because changeovers and the order book cap the real number below 100. A plant at 90% in this business is, for practical purposes, full; don't wait for a 100 that never comes.

Check whether the capex is funded and on time — because the number drifts. An announced capex is guidance, not a fact. Watch HEG's other project — the new graphite-anode plant for EV batteries. The total capex it guided crept across three calls: "INR 1,700-1,750 crores" (Q3 FY25) → "INR 1,850 crores" (Q4 FY25) → "INR 1,800-1,900 crores" (Q1 FY26), and Inve's record marks that commitment missed on cost. A 10% overrun on a flagship project, narrated so gently across calls that no single quarter looked like a miss, is exactly the kind of slippage the high utilisation can't warn you about. It's a pattern across the calls Inve parses: capex guidance frequently slips quarter by quarter and then goes quiet — the target drifts, and at some point it simply stops being mentioned. That behaviour, not the high utilisation gauge, is what the Promise Tracker is built to catch. A capex target that drifts, then goes silent, is telling you something the utilisation reading never could.

Mind what utilisation can't tell you — it measures volume, not money. This is the discipline HEG teaches most sharply, and it's where the gauge can fool you. HEG ran near-full all year — "over 90% capacity utilization throughout the year, reaching 95% production utilization in Q4" (HEG Q4 FY26 concall). And in that same Q4 FY26 it posted an operating loss of ₹148 crore on sales of ₹603 crore — a negative 25% margin (Inve data, Q4 FY26). A plant running at 95% lost money, because graphite-electrode prices, not its volumes, had collapsed. Utilisation is a volume gauge. It will sit at 95% while the economics fall out from under it — the same skepticism you'd apply when checking whether reported profit is backed by cash.

Where this signal fails

Capacity utilisation is a prior, not a conclusion, and it's worth being plain about its limits rather than selling it harder than it deserves:

• It's often not disclosed at all. Many manufacturers never quote a utilisation figure, and some quote it only when it flatters them. Absence of the number is itself a constraint — you can't read a gauge that isn't on the dashboard.
• Definitions differ — across companies, and even across quarters at one company. "Capacity" can mean rated, effective, or de-bottlenecked capacity; "utilisation" can be measured on volume or on value, on installed or on expanded base. HEG itself reports against an "expanded capacity" base. Two companies at "85%" may not be comparable, and one company's 85% this quarter may not mean what it meant last quarter.
• A single quarter is noisy. Seasonality, a maintenance shutdown, an inventory build, or a one-off order can move the figure several points with no change in the underlying trend. That's why discipline #1 is to read the trajectory, not the point.
• And, as the whole piece argues, it's a volume gauge that's blind to price and to execution. A full plant can lose money; an announced capex can slip and go quiet.

Used as one input among several — read across quarters, against the company's own history, alongside margin and the capex track record — it earns its place. Used as a standalone trigger, it will mislead you. If your read rests entirely on a single self-reported utilisation number, assume you could be wrong.

Where capacity utilisation fits in a research process

Utilisation belongs with the other operating KPIs that lead the reported result — read it on the shortlist, after the financial screen, to see where a manufacturer is in its capex cycle before the next result confirms it.

For each manufacturing holding: pull the utilisation figure from the latest call (it's usually in the Q&A, not the deck), read it across the prior few quarters for the trend, listen for the escalating expansion language, and — when a capex is announced — log it as a falsifiable commitment to verify next quarter. Then watch whether the number holds. Dixon Technologies, the EMS giant, guided a refrigerator-capacity expansion that only ever moved one way: "Targeted expansion to 2 million units" (Q4 FY25) → "2.5 million units" (Q1 FY26) → "3 million units" (Q3 FY26) → then, in Q4 FY26, "No update provided on refrigerator capacity in the Q4 call" — a commitment Inve marks ghosted (Dixon concalls Q4 FY25–Q4 FY26). The target inflated three times, then vanished. Guidance that only ratchets up and then goes silent is its own signal. (Illustration of the metric, not a view on Dixon.)

Inve's KPI Screener surfaces operating KPIs like capacity utilisation across companies, with their trend and a data-confidence flag, so you can read a sector's capex pressure at a glance rather than reading every transcript by hand. Whether the capex management commits to is actually built on time and on budget — HEG's anode plant drifting 10% over, Dixon's fridge line going quiet — is the long-memory check the Promise Tracker keeps. That cross-company, quarter-after-quarter assembly is the work no single transcript and no ratio screen does for you.

Frequently asked questions

Capacity utilisation is the gauge that tells you a manufacturer's story before the manufacturer does. HEG's plant grinding past 90% for several quarters was a company that had run out of road — and the ₹650-crore expansion, announced the moment the gauge crossed the line, was the slow paperwork confirming what the number already showed. But the same case carries the warning: that plant hit 95% utilisation and still lost ₹148 crore in a quarter, because the gauge measures how hard the factory runs, not what its output is worth. So here is the owner's question the metric forces you to ask: if a manufacturer you hold is running near full and about to spend its way to a new plant, what must you believe — about demand and price — for that new capacity to earn its return, rather than sit at 95% and bleed?

See how a manufacturer's capacity utilisation is trending — and whether its capex commitments have held up — in Inve's KPI Screener and Promise Tracker.

Related reading

• How to Analyse a Cement Company (India)
• How to Analyse a Pharma Stock in India
• How to Analyse a Real Estate Developer Stock </content>