The app is the marketing. The lease is the company.
The costume
Quick commerce presents itself as a triumph of engineering. There are routing algorithms, demand forecasts, pickers trained to assemble a basket in under a minute, and a great deal of talk about the last mile as though it were a frontier requiring exploration.
It is a persuasive costume. It has convinced investors, journalists and, for a period, several billion dollars of Gulf sovereign wealth. But strip the app away and look at what these companies actually own, actually pay for, and actually fight over, and a duller business emerges. They are competing for small boxes of urban floorspace situated within a very short distance of people with disposable income. Everything else, the riders, the software, the ten-minute promise itself, is downstream of that.
Quick commerce is a land business. The technology is the leasing agent.
Ten minutes is not a time, it is a radius
Begin with physics, since marketing tends not to.
A rider on a scooter in an Indian city, allowing for order acceptance, picking, packing, traffic lights, a gate, a lift and a wrong flat number, can reliably cover about two kilometres in ten minutes. That number is not a target. It is a constraint imposed by the street.
Which means the ten-minute promise is not a claim about operational excellence. It is a claim about where the box is. A firm cannot algorithm its way to a shorter distance. It can only buy or rent its way closer. Once you accept that, the entire industry reorganises itself in your head: the core competency is not fulfilment, it is site selection, and the core asset is not the app, it is a portfolio of leases.
This also explains the otherwise strange fact that these companies do not compete on speed anymore. They compete on how many stores they have opened. Store count is the metric because store count is the business.
The tell is in the rent
If you suspect an industry of being something other than it claims, look at what it overpays for.
A conventional warehouse in India's top seven cities rents for roughly ₹18 to ₹31 per square foot per month, with Pune at the top of that range and the Mumbai region at the bottom. Dark stores, according to Savills India, command ₹40 to ₹200 per square foot per month, with Delhi at ₹150 to ₹200.
Consider what that premium buys. Not better shelving. Not a taller ceiling, superior loading bays or cheaper power. A dark store is, physically, a worse warehouse: smaller, more awkward, frequently in a basement, often in a building designed for something else entirely. Quick commerce firms are paying up to five times conventional warehouse rent for inferior warehousing.
They are not buying storage. They are buying coordinates.
The effect on the neighbourhoods supplying those coordinates has been predictable. Reporting from Bengaluru's Domlur has rents moving from around ₹50 to ₹75 per square foot, with some deals struck 40 percent above the asking rate, and commercial space in prime residential pockets of Bengaluru, Delhi and Mumbai rising by as much as 35 percent. Landlords who could not previously let a basement showroom now have a queue. A former tuition centre in Andheri East is now a fulfilment node. Ground-floor units that once went to a chemist or a bakery are being outbid by operations managers with a spreadsheet and a two-kilometre circle.
None of this is a logistics story. It is a bidding war for well-located land, conducted by firms whose investor decks contain the word "network" a great deal and the word "landlord" not at all.
Case study: Getir, or how to lose ten billion dollars to urban form
The clearest proof that geography, not technology, decides this business comes from the company that tested the same technology in several different geographies.
Getir was founded in Istanbul in 2015 and became the sector's global standard-bearer. It raised $768 million in 2022 at a valuation of $11.8 billion, acquired the German rival Gorillas for £1.2 billion, bought FreshDirect in the United States, and sponsored Tottenham Hotspur, which is the traditional final stage of a European technology bubble. By September 2023 it was reportedly raising at $2.5 billion, a discount of roughly 79 percent. In April 2024 it withdrew from the United Kingdom, the United States, Germany and the Netherlands entirely, cutting around 6,000 jobs, and retreated to Turkey, where it said the prospects for sustainable profitability were stronger.
Note carefully what did not change during this collapse. The app was the same app. The riders were riding the same scooters. The picking process, the inventory software, the ten-minute promise: identical across every market.
What changed was the city.
Istanbul is dense, vertical and congested, with a habit of small and frequent shopping. A single dark store there sits within scooter range of an enormous number of households who buy little and often. Suburban Britain, Germany and the United States are horizontal. Households own cars, have garages, own large refrigerators, and do a weekly shop. The same two-kilometre circle drawn on a map of Chicago contains a fraction of the customers, each ordering less frequently, each order needing a delivery cost that American wages will not permit.
Getir did not fail because it executed badly. It failed because it exported a business model that was really a bet on urban form to cities with the wrong urban form, and then discovered that a scooter cannot make a suburb denser.
The Indian exception, and why it is geography rather than genius
India is where the model works, which is routinely presented as evidence of Indian operational brilliance. It is mostly evidence of Indian population density and Indian wage levels.
Blinkit, the market leader and now the majority of its parent Eternal's revenue, reported a genuinely impressive quarter in April to June 2026. Net order value rose 86 percent year on year to ₹17,132 crore. It added 200 net new dark stores to reach 2,443. Monthly transacting customers nearly doubled to 31.8 million, and it recorded 331 million orders, about 3.6 million a day. Adjusted EBITDA turned positive at ₹102 crore.
Now read the same quarter as a property investor would. Average order value fell slightly, to about ₹518. Growth came from more customers, more frequent orders and more stores, not from anyone spending more per visit. Net order value per store per day was ₹8.27 lakh, against a management assumption of around ₹11 lakh for the target economics. Adjusted EBITDA margin sits at 0.6 percent of order value against a modelled 6 percent, and contribution margin is about 4 to 5 percent, which slipped slightly as state minimum wage rises pushed direct costs per order up.
In other words: the business is rent, labour and a very thin spread, and the entire path to profitability runs through pushing more orders per day through the same box. That is not a technology flywheel. It is the operating logic of a shopping centre.
The clinching evidence arrived via Bernstein, which found that roughly 3,600 of the top 3,800 dark stores in India's eight largest cities are profitable, while tier-two stores continue to bleed cash. Same app. Same suppliers. Same ten-minute promise. Different map.
If this were a logistics innovation, it would travel. Innovations travel. Density does not.
The neighbourhood pays a fee it never agreed to
There is a second-order consequence that municipal authorities have been slow to notice, largely because the industry is filed under "startups" rather than "land use."
A dark store is a warehouse. It receives goods in bulk at odd hours, dispatches dozens of vehicles an hour, requires kerbside space it has not paid for, and generates no walk-in footfall in return. It is being sited, at scale, inside residential and neighbourhood retail zones that were planned on the assumption that ground-floor commercial units would be shops.
The costs of this are real and are borne by people who receive none of the revenue: parking congestion, pavement occupation, delivery traffic at night, and the steady eviction of low-margin neighbourhood retail by a tenant that can pay 35 percent more because it is monetising a radius rather than a storefront. India's quick commerce sector has been debated as a consumer protection question and a competition question. It is more accurately a zoning question, and no one has written the zoning.
How to value a company like this
If the thesis holds, the standard valuation framework is wrong. A technology business earns returns on code, which costs nothing to copy and can be deployed anywhere. This business earns returns on leases, which expire.
So the questions that matter are property questions. What is the weighted average remaining lease term across the estate? What happens at renewal, now that every landlord in Domlur and Andheri has read the same newspaper articles and knows exactly what the tenant's alternatives are, which is to say, the shop next door? Revenue-linked leases are already appearing in the mid-market, which is landlords politely announcing that they intend to be shareholders. Meanwhile, more than 6,000 dark stores now operate across India, several operators are chasing the same catchments, and the marginal store is being opened in precisely the places where Bernstein says stores lose money.
The bull case for quick commerce is that Indian cities are dense enough, and Indian labour cheap enough, to sustain a thin margin across an enormous volume of orders. That case is probably correct. It is also a statement about India, not about the companies.
The ten-minute delivery firm would like to be valued as a technology platform. It should be valued as a landlord's tenant with an unusually good customer acquisition strategy, in a business where the rent goes up every time it proves the location was a good one.