Everyone quoted Zepto's ₹5,905 Cr loss. The line that carries the filing is on page 162.

Zepto's IPO papers went public on a Monday. 690 pages. Every headline since has quoted the same number: ₹5,905 Cr of losses in FY26. It's a real number, and it's the least interesting one in the document.

The line that stopped me is on page 162. Advertising revenue: ₹1,636 Cr. Two years ago, this line was ₹49 Cr. That's 33x in 24 months, and it's still growing 151% a year. Read the filing closely and that single line quietly carries the entire business. Everything else in the P&L explains why it has to.

First, what the loss actually is

Before the ₹5,905 Cr scares anyone off, it's worth knowing what's inside it. A chunk of that headline is not cash leaving the building — it includes non-cash charges like share-based payments to employees (roughly ₹557 Cr of ESOP cost sitting inside the statutory loss). Strip the accounting noise and the company reports its own cleaner measure of operating performance: an adjusted EBITDA loss that works out to about ₹79 per order. That's the number that tells you how the engine runs. So the real question isn't "why is the loss so big" — it's "what is each order actually worth, and who pays to close the gap."

The goods business loses money before anyone is paid

Start with the thing Zepto looks like it does: sell groceries. The goods it moved last year cost ₹18,198 Cr to buy. They were sold on for ₹17,588 Cr. That's a ₹610 Cr hole before a single salary, lease, or delivery rider is paid.

One clarification worth making, because it trips people up: that ₹17,588 Cr isn't what shoppers paid at the app. It's largely business-to-business — Zepto supplying inventory to the merchant partners who actually sell to you. The customer pays the seller; Zepto makes its money around that transaction, not on it. Which is exactly the point. This is a marketplace structure doing what marketplaces do: supply the sellers at or below cost, keep the basket cheap, and earn everywhere except on the groceries themselves.

The delivery side is underwater too

If the goods don't make money, maybe the logistics do. They don't. Zepto collected ₹2,780 Cr in warehousing, packaging, and last-mile fees, and paid out ₹3,046 Cr to delivery partners alone. The part of the business everyone associates with quick commerce — the ten-minute promise — runs at a loss on its own line.

So two of the three things you'd assume are the business are structurally negative. Stack the revenue lines and the asymmetry is stark: platform commissions ₹564 Cr, subscriptions ₹28 Cr, and advertising ₹1,636 Cr — the last one growing at 151% a year, at margins the grocery side can only dream of.

The only fat-margin line is attention

2,468 brands paid that ₹1,636 Cr ad bill — an average of ₹66 lakh each. The most telling detail in the filing is who else is buying. The document names advertisers that don't sell a single product on the app: AU Small Finance Bank, CRED. They're buying slots on the post-order screen — the few seconds of attention after you've checked out. When a bank and a credit-card app are bidding for space inside a grocery app, the product being sold isn't groceries. It's attention.

That's why the platform's take-rate — the share of each order it keeps — has climbed so hard. The grocery business can't move it. The ad business can, because ad pricing is set by how many brands want the slot, not by the thin economics of a banana. Retail media is the highest-margin thing in commerce, and it's the one line in this filing compounding the way an investor needs it to.

I run finance at a D2C brand that sells on quick commerce. I sit on the other side of this line — the quick-commerce marketing invoices I approve every month land somewhere inside that ₹1,636 Cr. So this isn't an analyst's read for me. It's a supplier reading its platform's accounts.

The per-order math that made me sit back

Here's the whole document in four numbers. The average Zepto order carries roughly ₹390 of value. About ₹26 of that is advertising money from brands. The adjusted EBITDA loss — Zepto's own measure — is about ₹79 per order. Take the brand money out, and that hole widens to roughly ₹104.

Someone absorbs that ₹79–₹104 on every single order. For years it was investor money keeping your delivery cheap — the venture subsidy. What the filing shows is the subsidy quietly changing hands: brand marketing budgets have taken over a growing share of the job. It is the only subsidy in all 690 pages that is compounding rather than running down. Investor patience depletes; the ad line grows 151% a year.

What this IPO is actually asking

Put it together and the pitch comes into focus. This IPO isn't asking investors to believe in grocery margins. It's asking them to believe that attention inside a ten-minute app is worth more than the groceries moving through it. It's the model Amazon proved at scale — a retailer that makes its real money on advertising and third-party services, not on the products it ships. Zepto's filing is quietly chasing the same template, compressed into ten minutes.

As someone who pays for that attention every month, I think they might be right. That's exactly what bothers me. A subsidy funded by my marketing budget is a subsidy with a price that rises whenever the platform decides its ad inventory is underpriced — and that decision isn't mine to make.

If you sell on quick commerce, read it as a budget, not a milestone

For brands, the operator takeaway is concrete. The share of your platform sales that flows straight back as advertising is not a marketing line item you trim in a tight quarter — it's closer to rent. Branding doesn't get you out of paying it; at best, strong branding negotiates the rent down, because a brand people ask for by name needs less paid push to move off the shelf. But the rent itself isn't going away, and every filing like this one tells you it's going up. The brands that budget for retail media as a structural, rising cost of selling on these platforms will plan around it. The ones still calling it a discretionary spend will keep getting surprised by the invoice — and I know our number well enough to not enjoy looking at it.

Every figure here is drawn from Zepto's updated draft red herring prospectus — a public SEBI filing, freely citable. The FY26 summary P&L sits on page 71 and the KPI table on page 162, the latter certified by the company's auditors and approved by the audit committee. Derived figures (per-order economics) were recomputed from the filing and cross-checked against press coverage; nothing here is estimated beyond what the document states.