Quick Commerce Unit Economics, and Why They Could Destroy Zepto

Key takeaways

• Zepto doubled FY26 revenue to Rs 22,624 crore and widened its loss 26 per cent to Rs 5,905 crore, on 64.02 crore orders across 1,139 dark stores.
• Work the DRHP back to one order: Rs 353.4 of revenue, Rs 69.1 of gross profit, Rs 81.2 of delivery plus dark store cost. The store is under water before a rupee of marketing.
• Zepto lost Rs 78.75 an order on an adjusted EBITDA basis in FY26, down from Rs 136.15. Blinkit lost Rs 3.02.
• The mark fell from 7 billion dollars in October 2025 to a reported 2.3 billion on the IPO anchor book, and the listing was deferred by two to three quarters.
• The one real fix is own brand. In-house margins are reported near three times branded resale, and BigBasket already runs private label at 36 to 37 per cent of turnover.
• But own brand cannibalises advertising, the highest margin line these platforms have, and India bars an inventory-led model under majority foreign ownership. Swiggy board has approved capping foreign holding at 49.5 per cent for exactly that reason.

Frequency is the religion of quick commerce. Get a customer ordering four times a week instead of once and revenue compounds. That is true, and it is half a sentence. Frequency compounds whatever the unit economics already are. If an order makes money, frequency builds a business. If an order loses money, frequency builds a valuation and burns the cash faster.

Zepto is the cleanest live test of the second case, because it had to publish a draft red herring prospectus and show its working. What follows is that prospectus divided by one order, then the only strategy that can actually fix it, and the three reasons that strategy is harder than it looks.

The dark store profit and loss, one order at a time

Cost of goods is Rs 284.3 against Rs 353.4 of revenue, so gross margin is 19.6 per cent, or Rs 69.1 an order. Notice where that margin comes from. Rs 1,636 crore of FY26 revenue was advertising income, about Rs 26 an order and 37 per cent of all gross profit. Strip the ads out and the product margin is roughly 13 per cent. Zepto is already part media business, and that matters a great deal later.

Now the fulfilment side. Delivery and handling is Rs 47.6 an order and the dark store network another Rs 33.6, so Rs 81.2 of fulfilment against Rs 69.1 of gross profit. The store is short by about Rs 12 before marketing, salaries or software. At a 19.6 per cent margin, covering fulfilment alone needs revenue per order near Rs 415, about 17 per cent above today. Covering the whole cost base at current spend needs about Rs 865, two and a half times. The first is reachable. The second is not.

Why frequency makes it worse before it makes it better

This is the trap. Order value up 40 per cent reads as progress in a deck and as accelerated burn in the cash flow statement. It is the same error as celebrating gross order value in food delivery while the take rate is under attack from Flipkart and Rapido. Growth is not evidence of a business. Contribution per order is.

Zepto is efficient, and still a long way behind

Be fair to the company. Zepto ran about 1,540 orders per dark store per day in March 2026 against roughly 1,470 for Blinkit in June 2026, on less than half the stores, and cut average delivery distance from 2.05 km in FY24 to 1.78 km through what it calls densification. Loss per order improved 42 per cent year on year and was Rs 59.4 in the March quarter. The operating team is good. The problem is the starting point. Blinkit lost Rs 3.02 an order in the same year. Zepto runs the same race with roughly 26 times the per-order handicap and far less capital behind it.

The capital squeeze

A business with negative unit economics needs continuous capital by definition. Zepto held Rs 5,680 crore of cash at 31 March 2026 and burned Rs 4,330 crore of free cash flow through FY26. At that rate the balance sheet covers roughly sixteen months. Burn did improve, from Rs 5,332 crore in FY25, and a pre-IPO round of about Rs 1,000 crore extends it. I am not predicting insolvency. I am saying the company no longer controls its own timeline, and it is asking for money in a year when India recorded zero rounds above 100 million dollars in the first quarter and the median late stage cheque fell 68 per cent.

The route out is own brand, not a bigger basket

Raising the basket needs the customer to change behaviour. Raising the margin on the basket you already sell does not. When you stock a branded biscuit you are a distributor with a warehouse earning a thin retail spread. When you sell your own biscuit you collect the manufacturer margin and the retail margin on the same unit. Inc42 puts in-house brand margins at roughly three times branded resale. That is the entire reason every grocery operator eventually does this.

India already has the templates. BigBasket has spent a decade on bb Royal, Fresho and bb Popular and is now at 36 to 37 per cent of turnover, targeting near 40, with Rs 500 crore committed to expansion. DMart is the offline proof: store brands around 20 to 25 per cent of sales on a gross margin near 16 per cent, used to turn inventory faster rather than to replace price discipline. Quick commerce is well behind. Zepto has Relish in meat and seafood, reported heading to Rs 500 crore of revenue in 2026, and Daily Good in staples. Blinkit has Whole Farm. Instamart is building exclusive SKUs and own labels. None of them is close to BigBasket.

Three reasons that route is harder than it looks

One, you cannibalise your best customer. Advertising is the highest margin line these platforms have. Blinkit, Zepto and Instamart advertising revenue is put near Rs 4,900 crore for 2026, and Redseer has beauty brands spending up to 10 per cent of their quick commerce sales on ads. Every search slot you hand to your own label is a slot an FMCG company was paying you for. You are not simply adding margin, you are swapping a very high margin rupee for a slightly higher margin one and hoping the arithmetic nets out.

Two, the ownership rule. India separates two models. A marketplace connects a buyer to a third party seller and may take full foreign investment. An inventory-led model, where the platform owns what it sells, may not operate under majority foreign ownership. Private label is inventory by definition, so the strategy is unavailable to a company whose cap table is majority offshore. Swiggy is working this in public: its board has approved capping aggregate foreign ownership at 49.5 per cent to qualify as an Indian owned and controlled company, with a special resolution at the 18 August annual general meeting, after an earlier attempt failed in May. The stated purpose is to let Instamart go inventory-led.

Read that in reverse and it is uncomfortable. To run the only strategy that fixes your margin, you must cap the source of capital that built Indian e-commerce in the first place. Once you sit at 49.5 per cent, fresh foreign money can only arrive by one foreign holder buying out another. New capital has to be domestic, which means family offices, domestic funds and the public market, at exactly the moment those pools are also being asked to absorb an IPO pipeline. That is structural, not a sentiment cycle.

Three, inventory is a balance sheet. A marketplace holds no stock. An inventory-led grocer holds working capital, and in fresh it holds shrinkage. Zepto procurement was Rs 18,199 crore in FY26. Move even a fifth of that onto your own books and you have added a financing line to a company that already burned Rs 4,330 crore of free cash flow. Add expiry on fruit and dairy, quality control across dozens of contract manufacturers, and the reputational cost of one bad own-label batch landing on the platform that carries your name. BigBasket took ten years to reach 36 per cent for these reasons, not because nobody thought of it sooner.

What would have to be true

Here is the arithmetic, and it is my model, not a disclosure. Of Rs 353.4 revenue per order, about Rs 26 is advertising, leaving roughly Rs 328 of product revenue at about 13 per cent margin, or Rs 43. Add the ad rupee and gross profit is Rs 69.1 against Rs 81.2 of fulfilment. Closing that gap with margin rather than basket size means lifting product gross profit about 30 per cent. At a three times margin multiple on own brand, that needs private label to be roughly 14 per cent of product revenue. If a quarter of the ad income walks out with the shelf space, the requirement rises to about 22 per cent.

So my call is this. Fourteen per cent is reachable inside two years, because BigBasket already runs at 36. Twenty-two per cent while defending advertising revenue is much harder, and it is the number to watch. If Zepto or Instamart disclose own brand above 20 per cent of basket with ad revenue still growing, the economics are fixed. If own brand rises while ad revenue flattens, they have moved the loss rather than removed it.

Three other checkable things. Revenue per order above Rs 415 within four quarters turns store contribution positive on this arithmetic. Loss per order below Rs 30 by the March 2027 quarter extends the Rs 136 to Rs 79 to Rs 59 trend. A completed listing or a priced round at or above 4.5 billion dollars ends the capital question. Miss all of them and the pressure is structural.

What to do about it

If you are waiting on this IPO, price contribution margin, not growth. Ask what revenue per order is, what fulfilment costs, and what happens to both if discounting stops. Then ask for two numbers together and never one alone: own brand share of gross order value, and advertising revenue growth. A platform that shows you only the first is hiding the second.

If you are an FMCG brand, treat rising private label share as your notice period. Your distributor is becoming your competitor while still charging you for shelf space. If you are building here, scale does not fix a negative contribution margin, it multiplies it. Fix the order first, then buy the frequency. The alternative is the pattern in the audit of Indian unicorns that grew revenue and never found a floor under their losses.

Read next: how Flipkart and Rapido are attacking the Swiggy and Zomato take rate, and why a record global venture year never reached India.

Sources

• Entrackr and Inc42, Zepto FY26 revenue, loss and expense breakdown from the DRHP
• Forbes India, Zepto DRHP detail on cash, free cash flow, dark stores and market share
• INDmoney, loss per order for Zepto, Blinkit and Swiggy Instamart, FY26
• Bloomberg and The Week, the IPO anchor book near 2.3 billion dollars and the deferral
• Storyboard18 and Business Standard, Blinkit and Instamart store counts, and quick commerce advertising revenue near Rs 4,900 crore in 2026
• Inc42 and IBEF, private label margins at roughly three times branded resale, BigBasket own brand share and target, and the Zepto Relish and Daily Good brands
• Inc42 and Business Standard, the Swiggy board approving a 49.5 per cent foreign ownership cap and the route to an inventory-led Instamart
• India Dispatch, DMart store brand share and the role of private label in its model

Per-order figures are derived by dividing reported FY26 totals by reported FY26 orders. Figures described as modelled are my own calculation on those totals. All other figures are as reported by the sources named above at the time of writing.