Quick commerce is usually framed as a speed race, but the disclosed numbers are about radius. Instamart added 69 net dark stores in the September quarter, taking its network to 1,240 outlets across 133 cities, while Blinkit added 308 stores to reach 2,751 and has said it wants more than 3,500 by March 2027 (Urbana Acres). Instamart's quick-commerce arm still posted an adjusted operating loss of Rs 895 crore for the quarter even as net order value rose 34.8 per cent. As of November 12, 2026, the race is measured in doors, and the business case is decided store by store.
What the network numbers actually show
A larger footprint is not the same as a better network. Instamart's contribution margin moved to negative 1.0 per cent of gross order value from negative 0.2 per cent in the prior quarter, and its adjusted operating loss widened to Rs 895 crore from Rs 849 crore a year earlier (Urbana Acres). Order value grew, losses grew faster, and the network grew fastest of all. For anyone building a store-selection model, that combination is the interesting part: the constraint is no longer demand, it is the cost of being close to demand.
Store expansion follows capacity, not ambition
On the October 22 earnings call, management was explicit about the trigger for a new store: sites are opened when an existing store crosses roughly 2,000 orders per day (ScanX). That is a capacity rule, not a market-share rule, and it changes the way a trade area should be scored. A site is not judged by how many households live nearby, but by whether the nearest existing store is already saturated and whether the new one can take volume without starving the old one.
Cannibalisation is the metric nobody publishes
Operator economics make the gap visible. One analysis of Zepto's filing notes that store openings fell from 699 in FY25 to 169 in FY26 while closures rose from seven to 59, and that 36.61 per cent of stores were run by growth partners as of March 2026 (Startup Chai). The same arithmetic explains why stockouts matter more than store counts: a store handling 2,200 orders a day can lose a third of its volume when a new site opens 700 metres away, while its rent, staff and equipment costs stay where they were. Platforms gain coverage; operators absorb the drop.
| Model | Who carries the asset risk | Key metric to watch |
|---|---|---|
| Company-operated store | Platform | Contribution margin per store |
| Growth-partner store | Operator | Orders per day versus breakeven |
| Adjacent new site | Whichever store loses volume | Volume shift within 1 km |
Where a store-level model has to start
Three inputs are enough to sanity-check a new dark store before signing a lease. First, the saturation level of the nearest existing store, expressed as orders per day against its practical ceiling. Second, the overlap of delivery polygons, since two stores covering the same buildings compete for the same orders. Third, the replenishment distance to the nearest mother hub, because a store that runs out of top sellers in the evening loses the orders it was built for.
None of these inputs appear in a quarterly results release. Instamart's disclosure that it reached 5.2 million square feet across 1,240 stores gives an average footprint but says nothing about how individual sites perform, what their leases cost, or how quickly they reach viability. The absence of that data is precisely why store-level diligence has to be done locally, with footfall counts, polygon maps and replenishment schedules rather than with headline store counts.
Speed is a location decision, not a promise
Ten-minute delivery is a radius problem dressed as a technology problem. Every additional minute of promised delivery time buys a wider polygon and therefore fewer stores per city; every minute removed pushes inventory closer to households and raises rent per square foot. When a platform says it will accelerate expansion without giving up margin, the claim rests on the assumption that denser coverage lowers delivery cost per order faster than it raises occupancy cost. That assumption is testable store by store, and it is the one number a brand or an investor should ask for.
What this means for brands selling through these networks
For a consumer brand, the practical consequence is that on-shelf availability now varies by store cluster rather than by city. Two stores three kilometres apart can carry different assortments, because the platform allocates SKUs by local sales velocity rather than by a national planogram. A brand that monitors availability at city level will see averages that hide both the best and the worst stores, and will misprice its trade spend as a result.
The fix is unglamorous: sample a fixed panel of stores, check availability and price weekly, and record which cluster each store belongs to. Over a quarter, the pattern shows where distribution is genuinely working and where the listing exists only on paper. Given that the largest networks are adding hundreds of stores a quarter, the panel has to be refreshed as the footprint moves, or the sample quietly becomes a history of last year's network.
FAQ
Is a bigger dark-store network better for a brand?
Not automatically. Coverage helps only where the store is stocked and the cluster has demand; empty shelves in a new store add cost without adding sales.
What is the single most useful store-level number?
Orders per day against the store's practical ceiling. It shows both whether the site is working and whether the next site nearby is needed.
How often should availability be checked?
Weekly, on a fixed panel. Quarterly reviews hide the stockouts that happen between replenishment cycles.
Does this apply outside India?
Yes. Any market where platforms add small-format fulfilment sites faces the same cannibalisation and replenishment questions.
