You open an app, tap twice, and a packet of coriander, a phone charger, and a tub of ice cream are at your door before your tea has cooled. The delivery fee was ₹0, or close enough to feel like it. Nothing in business is free. Someone, somewhere, is covering the bill. The intriguing question is who.
This is a look at the money behind India’s fastest-growing way to shop: what one order really costs, why the buildings behind it matter more than the bikes in front of it, and why the winners are being decided by something much less glamorous than speed.
The ₹500 Problem
Start with the basket. The typical quick commerce order in India sits somewhere around ₹450 to ₹520 after discounts. Groceries are a famously thin-margin business, so the gross profit on that basket is small before a single cost is counted.
Then the costs arrive. Industry estimates commonly put the rider at roughly 8 to 12 percent of the order value, the dark store’s running costs at 5 to 8 percent, and packaging, technology, and customer support at another 2 to 4 percent. Add discounts, payment fees, marketing, and wastage, and a basket that looks healthy on the receipt can leave very little behind.
That is why the early years of the category looked the way they did. Platforms were not selling groceries at a profit. They were buying habits.
The Real Product is a Building
The most important asset in quick commerce is not the app or the bike. It is the dark store: a small, closed-to-the-public warehouse placed a few minutes from your home, stocked with only what your neighborhood is likely to want.
A dark store behaves like a restaurant that has to be full all day. Rent, electricity, staff, and cold storage cost roughly the same whether it handles 300 orders or 1,300. Analysts who track the sector put the point where a mature metro store starts to make a positive contribution at somewhere around 1,200 to 1,400 orders a day, with baskets of ₹500 or more.
Here is what that means in practice. Speed is not the thing being sold. Density is. The faster a platform can pack a neighborhood with orders, the cheaper each order gets. Ten-minute delivery is marketing. The number of orders per square foot is the key metric for the business.
The scale of the bet is large. The category leader was running more than 2,400 dark stores by mid-2026. A major challenger had around 1,170 across 131 cities, and it reported that a little over 45 percent of its stores were already profitable at the contribution level. The flip side of that number: more than half were not.
The Scoreboard, Read Carefully
For years the headline was losses. In 2026 the headline changed, but it deserves a careful read.
The category leader turned adjusted EBITDA positive in the October to December 2025 quarter and reached roughly ₹102 crore in April to June 2026. That sounds big until it is converted: about 0.6 percent of order value, or a few rupees per order. This is a significant milestone, but it also represents a very thin margin.
A major challenger reached contribution margin breakeven in May 2026. That means it stopped losing money on the marginal order. But contribution margin ignores rent, marketing, and head office costs, and once those were counted, estimates put the loss in that quarter at about ₹68 per order.
Two different words, “profitable” and “breakeven”, and a very long gap between them. It is worth remembering that these are company-defined metrics and not net profit.
So Where does The Money Actually Come From?
If groceries alone cannot pay for the speed, something else has to. One platform spelled out its own route to breakeven, and the numbers are revealing. It said it needs roughly ₹30 more per order. About ₹10 from a better product mix, ₹10 from advertising, ₹5 from density and automation, and ₹5 from higher store utilisation.
Read that again. A third of the fix is ads.
The shelf in your app is rented space. Brands pay to be the first biscuit you see, the default shampoo, the top result when you type “chips”. Commissions that brands pay to sell on these platforms typically run from 15 to 25 percent, and can climb towards 35 percent once visibility fees are added. The other levers are private-label products and higher-margin categories like beauty, electronics and pharmacy, which is why the app that once sold you milk now also sells earphones.
So a surprising amount of the answer to “who pays” is: brands do, for the right to be in front of you at the exact moment you are ready to buy.
The Shopper Pays Too, Just Not on The Invoice
You may not see a big delivery charge, but the model quietly recovers its cost in other ways. Small handling and surge fees. A minimum order value that nudges you from ₹200 to ₹350. And a design built around impulse: the cold drink at the checkout, the “people also bought” strip, the speed that removes the pause between wanting and having. A study last year found quick commerce had made many shoppers more impulsive.
The average basket is the lever the entire model leans on. You came for milk. The platform needs you to leave with milk and a chocolate bar.
The Part of the Bill that is Hardest to See
Speed has a cost that does not appear in a P&L line, and it falls on the people doing the riding.
In January 2026, after worker strikes and safety concerns, the government asked major platforms to stop advertising a fixed “10-minute” promise, and they removed the claim from their apps and branding. Reactions were mixed. Some analysts called it optics rather than a change to the model, and platforms said the business would not be materially affected. Some delivery partner unions said they had seen little change on the ground, and several riders said their main concern was falling payouts per order, more than the timer itself.
Whatever view you take, one economic point holds. When the whole system is tuned for minutes, the pressure to hit those minutes lands on whoever is closest to the doorstep. Dense stores, shorter distances and better batching can reduce that pressure. Pure pressure cannot be sustained. That is a business reason as much as a human one to design the speed out of the system rather than out of the rider.
Where Technology Quietly Decides Who Wins
This is where the story gets interesting for anyone who cares about how things are built.
One platform’s leadership has said inventory losses (expiry, damage, shrinkage and transit) run at about 1.8 percent of order value. Its adjusted EBITDA margin in the same period was about 0.6 percent. Wastage was three times the margin.
When the thing you lose is bigger than the thing you earn, a better forecast is not a nice-to-have. It is the business. Behind a ten-minute promise sit a lot of quiet decisions made by software:
-
Demand forecasting per store. How many litres of milk will this lane need by 8 a.m.? Too many and it expires. Too few and the order is lost.
-
Slotting. Which item sits on which shelf so a picker can assemble a 12-item order in under two minutes.
-
Batching and routing. Whether two nearby orders can share one rider trip without breaking the promise to either customer.
-
Store placement. Where the next dark store should open so it fills quickly, not slowly.
In each of these, the person on the floor still makes the order happen. The software simply makes sure they are working with the right stock, in the right place, at the right time. A percentage point of wastage saved or a minute shaved off picking, repeated across millions of orders, is worth more than any discount.
