Seven hundred and fifty dark stores is a coverage figure. A billion dollars is a momentum figure. Neither tells a brand whether the item it paid to list is actually pickable when a customer orders it, which is why the Amazon Now disclosure reads as a coverage story with an availability question tucked inside. The programme crossed USD 1 billion in annualised gross sales over the trailing three months, with around 700,000 daily orders across more than 60 cities, through November 15, 2026.
What the disclosure counts, and what it rounds off
Three cautions belong next to the headline. The run rate is annualised from a three-month window, so it describes momentum rather than a finished year. The market-share figures, roughly 10% by orders and 14% by value, are the company own estimates in a category where rivals Blinkit, Instamart and Zepto built scale earlier. And centre counts are a coverage measure: they say where the network exists, not how deep the inventory is inside it. A node handling a few hundred orders a day in a smaller town cannot economically hold the same assortment as a metro node, so its range gets trimmed, and the trimmed items are the first to break availability.
700,000 orders a day across 750 nodes
Divide the daily order figure by the node count and the operating problem becomes visible. Roughly 900 orders per node per day is enough to justify a compact grocery range and not enough to justify a deep one. That arithmetic explains why the numbers a brand should watch are not city counts but order frequency and average order value, both of which determine whether a node can afford to stock deeper. Amazon has an average order value of around INR 550 on the service, with perishables close to half of orders, which is a basket built on frequency rather than ticket size.
| Metric | Disclosed value | Why it matters |
|---|---|---|
| Annualised gross sales | Above USD 1 billion | Describes momentum, not a completed year |
| Daily orders | Around 700,000 | Frequency per node decides stocking depth |
| Fulfilment centres | More than 750 | Coverage, not assortment depth |
| Cities covered | More than 60 | Reach; item availability still varies by node |
| Average order value | Around INR 550 | A frequency basket, not a stock-up basket |
| Year-end target | 1,000 dark stores | Adds nodes faster than it adds density |
Why value share runs ahead of order share
A 10% share of orders against a 14% share of value means the average Amazon Now basket is worth more than the category average, or that the mix skews to higher-priced items. Either reading points the same way: this channel is not purely an emergency top-up for small tickets. For a brand, that argues against treating quick commerce as a clearance shelf and for putting the same pack architecture in it as in the main marketplace, with a separate note on which sizes actually fit a 30-minute basket.
A marketplace listing is not a pickable unit
The availability gap, measured at item level
The commercial difference between a listing and a sale is whether the item is pickable when the order arrives. A brand can appear in 60 cities on paper and still be unbuyable in half of them, because the node that serves a given postcode stocks only two of five SKUs. The measurement that catches this is item-level fill rate by node, refreshed as the footprint moves; the measurement that hides it is a regional average, which stays healthy while individual nodes run dry. Category growth in this channel will be decided by that gap, not by the next city announcement.
Amazon is pushing toward 100 cities and 1,000 dark stores by year-end as part of a previously announced plan to cover more than 300 cities. Each new node in a low-density market dilutes the frequency that makes the earlier nodes work, which is why the market size story and the assortment story have to be read together. Growth at this speed is real. Depth at this speed is the part nobody has published a number for.
A note on comparability, because this is where most quick-commerce decks go wrong. Amazon order volume and the rival networks are reported on different bases, and the order and value shares are company estimates rather than audited figures. The one number that can be compared across platforms is node-level availability for the same SKU, tested with the same sample order method on the same day. That is also the only availability figure a brand can verify without depending on anyone disclosure, which makes it the right basis for a category review that has to hold up in front of a retail buyer.
FAQ
How big is Amazon Now in India?
It crossed USD 1 billion in annualised gross sales over the trailing three months, with daily orders around 700,000 across more than 60 cities and over 750 fulfilment centres.
Is the USD 1 billion a full-year figure?
No. It is an annualised run rate extrapolated from a three-month window, so it measures momentum rather than a completed fiscal year.
Why does order share differ from value share?
Roughly 10% of orders against about 14% of value suggests a higher average basket than the category, or a mix weighted toward higher-priced items; both figures are company estimates.
What should a brand measure in quick commerce?
Item-level fill rate by node, the gap between listed SKUs and pickable SKUs, and the substitution and cancellation rate from sample orders.
