Depreciation: The Cost Hiding in Reported Profit

A cement plant near where I grew up ran the same kiln for as long as anyone could remember. To the people who worked there it was just the kiln — a fixed thing, like the hill behind it. But on the company's books, that kiln was quietly getting cheaper every single year. Not because anyone sold a piece of it. Because of an accounting rule called depreciation. And the speed at which the accountants decided to make it "cheaper" changed how much profit the company got to report.

That is the whole idea you need to hold onto: depreciation is a real cost that nobody writes a cheque for this year — and the company gets some say in how big it looks. Get that, and a lot of "earnings quality" stops being mysterious.

A delivery van, spread over its life

Forget kilns for a second. Picture a small bakery that buys one delivery van for ₹12 lakh. The van will last roughly ten years.

Here's the question the bakery's accountant has to answer: in which year is that ₹12 lakh an expense?

It would be silly to dump the whole ₹12 lakh into year one. The van isn't used up in year one — it'll be hauling bread for a decade. So accounting does the sensible thing: it spreads the cost across the years the van actually works. ₹12 lakh over 10 years is ₹1.2 lakh a year. That ₹1.2 lakh is depreciation — the slow, annual recognition that a long-lived asset is being used up.

Two things follow, and both matter to you as an owner.

First: the cash already left in year one, when the van was bought. The depreciation in years two through ten is a bookkeeping expense — it lowers reported profit, but no rupee leaves the bank that year. (That's why cash flow and profit drift apart for asset-heavy firms — a separate idea we cover in cash is truth.)

Second — and this is the one almost nobody watches — the bakery chose "ten years." What if it had said the van lasts five years? Then depreciation jumps to ₹2.4 lakh a year, and reported profit falls. Say twenty years, and depreciation drops to ₹60,000, and reported profit rises. Same van. Same cash. Different profit, just by changing one assumption.

Now imagine a business whose "van" isn't ₹12 lakh but ₹99,000 crore of plant and machinery. That's where this stops being a bakery story.

A real Indian van: UltraTech's ₹99,000 crore of machinery

UltraTech Cement is India's largest cement maker — kilns, grinding units, conveyors, trucks, the lot. On its books sit about ₹99,260 crore of fixed assets, with another ₹8,276 crore of capital work-in-progress (plants being built), per its FY26 (Mar 2026) balance sheet (Inve data, 2026; figures per Screener's UltraTech consolidated balance sheet). This is a capex-heavy business in the purest sense: to earn a rupee, it must first sink enormous sums into steel and concrete that wear out slowly.

So how big is its "van cost"? In FY26 UltraTech's depreciation was about ₹4,644 crore (Screener, UltraTech consolidated P&L), up from ₹4,015 crore the year before. Hold that figure against its profit:

• Operating profit (EBITDA — earnings before depreciation) for FY26: ₹17,005 crore
• Depreciation: ₹4,644 crore
• Net profit (after depreciation, interest and tax): ₹8,188 crore (Inve data, 2026)

Read those three lines slowly. Depreciation alone is 27% of EBITDA — more than a quarter of operating earnings vanishes before you reach the bottom line. Put differently, depreciation equals 57% of the entire net profit the company reported. (EBITDA, by the way, is just "earnings before interest, tax, depreciation and amortisation" — the profit a business makes before the van-cost and the other below-the-line items.)

So picture the lever. If UltraTech's accountants stretched the assumed life of all that machinery and shaved depreciation by even a fifth — roughly ₹930 crore — pre-tax profit would jump by the same amount, and the headline would read "profit up sharply," with not one extra bag of cement sold. That is how a capex-heavy company can flatter earnings without doing anything in the real world.

Aggressive vs lenient: the same machine, two profits

"Aggressive" and "lenient" here don't mean fraud. They mean which direction the assumption leans.

• Lenient (slow) depreciation — assume assets last a long time. Smaller annual charge, higher reported profit now. The risk: the company is under-counting how fast its plants actually wear out, so today's profit is borrowing from tomorrow's repair-and-replace bill.
• Aggressive (fast) depreciation — assume assets wear out quickly. Bigger annual charge, lower reported profit now. This depresses earnings, but it's conservative: the company is honestly admitting its machines won't last forever.

Crucially, this lever isn't a hidden dial. The assumed useful lives — and any change to them — must be disclosed in the notes to the accounts, where management estimates are reviewed at least annually under Ind AS 16 and signed off by the auditor and the board. So management can't silently shorten or stretch an asset's life undetected; a careful reader can see the assumption, and any revision to it, in the filing.

A beginner sees a company with low depreciation and high profit and thinks "great margins." An owner asks the harder question: are these earnings real, or are they propped up by an optimistic guess about how long the kilns last? The fix isn't to memorise a formula. It's to compare like with like — does this cement maker depreciate its plants faster or slower than peers like Ambuja or Shree Cement? A company that's the outlier on the lenient side, year after year, is one whose profit deserves a second look.

Buffett's jab is exactly this point. Some managers love to wave you toward EBITDA — profit before the van-cost — as if the machinery renewed itself for free. It does not. The kilns wear out, the trucks die, the grinding units need replacing. Depreciation is the accountant's honest estimate of that. Ignore it and you're letting the tooth fairy run your model.

Today's capex is tomorrow's depreciation

Here's the part that turns depreciation from a backward-looking footnote into a forward-looking signal.

Every rupee of capex a company spends today becomes depreciation in future years. So when management tells you on a concall how much it plans to invest, it's also telling you — if you listen for it — how much its future depreciation, and therefore its future reported profit, will be weighed down. (The harder question is whether that spending will actually earn its keep — see the capex that never earns.)

UltraTech is doing exactly this. On its Q4 FY26 call, management guided to annual capex of ₹8,000–10,000 crore per year from FY27 to FY31 (Inve data, 2026; management's own stated figure). That's a deliberate, multi-year building spree to push capacity past 200 million tonnes. Wonderful for long-term volumes — and also a near-certainty that the depreciation line, already ₹4,644 crore, climbs steadily for years. An owner reading that guidance knows reported profit will carry a heavier van-cost tomorrow than it does today, and won't be spooked when it does.

This is the kind of forward commitment worth tracking rather than forgetting by next quarter. (Take UltraTech's earlier FY26 capex guidance, first set at ₹9,000 crore: it was revised up to ₹10,000 crore a couple of quarters later, then reaffirmed as on-track mid-year — ₹7,100 crore spent in nine months against a ₹9,500–10,000 crore full-year figure — before the final actual FY26 capex number was simply left unstated on the closing call. That last dropped thread, after quarters of diligent updates, is exactly what Inve's Promise Tracker is built to catch across a whole portfolio, quarter after quarter, so you don't have to.) None of this is a view on UltraTech the stock — only an illustration of how to read the cost.

So what does an owner actually do?

You don't need to recompute depreciation schedules. You need three habits.

Look at depreciation relative to profit and to peers — if it's unusually low for an asset-heavy business, ask why. Never accept EBITDA as "profit"; the gap between EBITDA and net profit is the van-cost, and for a cement or auto firm that gap is the whole game. And listen to capex guidance as a preview of tomorrow's depreciation, not just today's growth story.

Depreciation isn't a line to skip on the way to the bottom number. For a business made of steel and concrete, it's the closest thing the accounts have to an honest answer to one question: how fast is this company using itself up?

Frequently asked questions