Cloud kitchen meaning — the real definition
The cloud kitchen meaning is deceptively simple: a restaurant that only exists to fulfil delivery orders. There is no dine-in. There is no takeaway counter customers walk up to. There is no signboard on the main road because a passerby is not the target customer — the target customer is a person scrolling Swiggy at 8:47 PM with a hungry family.
The kitchen itself is often in a basement, a first-floor unit above a garage, an industrial estate, or a shared commissary building rented from an operator like Rebel Foods, EatClub or Kouzina. The address on your FSSAI licence is real, but no guest will ever type it into Google Maps.
What matters operationally: every square foot inside is dedicated to cooking, packing and rider handover. No dining area, no washrooms for guests, no host stand, no bar. Rent per plate of output is a fraction of what a dine-in restaurant pays, and that ratio is the whole reason the cloud kitchen business exists.
Cloud kitchen vs ghost kitchen vs dark kitchen
New operators lose hours to this question. The short answer: they are the same thing. Different regions and different tech-industry eras coined different words for the identical operational model.
| Term | Region where it dominates | What it means |
|---|---|---|
| Cloud kitchen | India, UAE, Saudi Arabia, wider GCC | Delivery-only restaurant. Popularised by Rebel Foods and Swiggy Access from 2016 onwards. |
| Ghost kitchen | United States | Delivery-only restaurant. Popularised by CloudKitchens (Travis Kalanick), Kitchen United and Reef from 2017 onwards. |
| Dark kitchen | United Kingdom, Ireland, wider Europe | Delivery-only restaurant. Popularised by Deliveroo Editions and Karma Kitchen from 2017 onwards. |
| Virtual restaurant | Global — used by aggregators | Overlaps but not identical. A virtual restaurant is a brand that only exists on delivery apps; it can run out of a cloud kitchen OR out of the spare capacity of a dine-in restaurant. |
| Commissary kitchen | Global — legacy foodservice term | A shared production kitchen that supplies multiple downstream outlets. Not delivery-first by default, but many cloud kitchens are commissary-hosted. |
So the dark kitchen vs cloud kitchen and ghost kitchen vs cloud kitchen debate is really a vocabulary debate, not an operational one. When a Bangalore founder says "I run a cloud kitchen" and a London founder says "I run a dark kitchen," the day-to-day looks identical: aggregator tablets beeping, KOTs printing, packers sealing, riders walking in and out.
How does a cloud kitchen work — the 4 steps
The mechanics of how a cloud kitchen works are the same whether you are running one brand or eight.
Step 1 — Order capture
A customer opens Swiggy, Zomato, WhatsApp, or the brand's own ordering page and places an order. In India roughly 75-85% of a typical cloud kitchen's volume comes from Swiggy + Zomato combined; the balance comes from WhatsApp, direct-ordering, and a small trickle from Google Maps and Instagram DM enquiries.
Step 2 — Order lands in POS and prints a KOT
The order flows into the kitchen's POS or ordering suite. A Kitchen Order Ticket prints in the pass. The order is tagged by channel so the closing report can attribute revenue correctly. This step is where most first-time operators cut corners — they run the aggregator tablets separately with no consolidation — and it is exactly where the daily-close pain begins two months in.
Step 3 — Cook and pack to spec
The kitchen cooks and packages to a delivery-safe spec. Sealed containers, tamper-evident labels, insulated pouches for hot items, ice packs for cold. Packaging for delivery is materially different from plating for a table — biryani has to survive 25 minutes in a bike-mounted bag without going mushy at the bottom.
Step 4 — Rider pickup and delivery
An aggregator rider arrives, scans the order, and rides out. Some brands run in-house delivery for their WhatsApp and direct channel to avoid commission, but that only makes sense at 40+ direct orders per day. Below that, aggregator rider fleets win on cost every time.
The 4 cloud kitchen models
Every cloud kitchen business fits into one of four cloud kitchen model shapes. Choosing the right one on day one is the single most consequential decision the founder makes.
Model 1 — Single-brand delivery-only
One kitchen, one brand, one menu. Think a biryani specialist, a poke bowl brand, a Neapolitan pizza cloud kitchen. Focus is the advantage: procurement is simple, staff training is tight, brand recall is clean. The disadvantage is capacity. If your kitchen can cook 200 orders/day and biryani only sells at lunch and dinner, the kitchen sits idle for 6-8 hours daily. That idle capacity is the pain the next model solves.
Model 2 — Multi-brand (virtual brands)
One kitchen running 3-8 branded storefronts on Swiggy and Zomato. Each brand targets a different cuisine, price band or daypart. A single kitchen might run a rolls brand at lunch, a biryani brand at dinner, a wraps brand for evening snacks, a healthy-bowls brand for the fitness crowd, and a dessert brand as an add-on. Rebel Foods pioneered this at scale in India — Faasos, Behrouz, Oven Story, Sweet Truth all out of the same kitchen. The multi-brand cloud kitchen model squeezes more revenue out of the same rent and staff. The trade-off: menu complexity, procurement complexity, and staff being pulled across 5 brand SOPs on the same shift.
Model 3 — Aggregator-hosted (Rebel Foods, EatClub, Kouzina, Ghost Kitchens India)
The brand does not own the kitchen. It rents production capacity and operations from a host. The host runs the physical kitchen, hires the staff, holds the FSSAI, and takes a cut. The brand supplies the recipe, the menu, the branding and the marketing spend. Capex drops close to zero. Speed to launch drops to 2-3 weeks. Margin is thinner because the host takes 15-25% on top of aggregator commission, but for a first brand it is an honest way to test demand before signing a 3-year lease.
Model 4 — In-restaurant delivery-only expansion
An existing dine-in restaurant launches a virtual brand out of its own kitchen for a cuisine it does not serve in-store. A Punjabi dhaba adds a "healthy bowls" virtual brand for the office lunch crowd. A cafe adds a "midnight sandwich" brand that only takes orders 10 PM to 2 AM. Zero incremental rent, zero incremental staff on core hours, incremental revenue from a channel and time-slot the main brand was not touching. This is the fastest-growing shape of cloud kitchen in India in 2026 because the barrier to entry is one Swiggy Partner listing.
Why cloud kitchens exploded in India and the GCC
Between 2018 and 2026 every input to the launch equation moved in the operator's favour.
| Metric | Dine-in restaurant | Cloud kitchen |
|---|---|---|
| Rent as % of revenue | 10-12% | 2-4% |
| Fitout capex (India tier-1 city) | ₹20-80 lakh | ₹5-25 lakh |
| Time from lease to first order | 6-9 months | 30-45 days |
| Staff on day 1 | 12-25 | 3-6 |
| Break-even revenue/month | ₹10-25 lakh | ₹3-4 lakh |
| Rider fleet you need to build | Own delivery or aggregator | Aggregators already have it |
| FSSAI + trade licence effort | 3-6 months | 30-60 days |
The three tailwinds that mattered most: (1) Swiggy and Zomato built out a national rider fleet, so a new cloud kitchen inherits distribution on day one instead of building it. (2) Commercial rentals in tier-1 cities have kitchen-suitable units in industrial estates and basement floors at ₹40-70/sq ft — one-third of high-street dine-in rent. (3) Aggregator UI conditioning trained a whole generation of urban Indians to eat delivery-first, so a customer sees "no dine-in" as neutral, not a red flag.
Dubai and the wider GCC followed the same curve two years later on the back of Talabat, Noon Food, Deliveroo and Careem. The Dubai cloud kitchen founder guide covers the licensing and cost specifics for that market.
The honest cost structure — a Mumbai cloud kitchen unit
Enough theory. Here is a real single-brand cloud kitchen P&L for a 400-600 sq ft unit in Andheri East, Kurla or a similar Mumbai industrial pocket in 2026.
| Line item | Monthly ₹ | % of ₹5 lakh revenue |
|---|---|---|
| Rent (400-600 sq ft, industrial pocket) | ₹40,000-60,000 | 8-12% |
| Staff — 1 head cook, 2 kitchen helpers, 1 packer | ₹80,000-1,20,000 | 16-24% |
| Food cost | ₹1,40,000-1,50,000 | 28-30% |
| Aggregator commission (all-in 25-32%) | ₹1,25,000-1,60,000 | 25-32% |
| Packaging | ₹20,000-30,000 | 4-6% |
| Ad spend inside aggregator apps | ₹25,000-40,000 | 5-8% |
| Utilities (LPG, electricity, water) | ₹15,000-25,000 | 3-5% |
| POS + Ordering Suite (Online eMenu) | ₹199 + ₹417/mo amortised | <0.2% |
| FSSAI + BMC compliance amortised | ₹2,000-3,000 | ~0.5% |
Break-even sits at roughly ₹3-4 lakh monthly revenue — that is where staff, rent and platform fees are covered but the owner takes home nothing. Genuine profitability starts around ₹6 lakh/month, and a well-run Mumbai single-brand cloud kitchen doing ₹8-10 lakh/month clears ₹1.2-1.8 lakh in owner earnings. Anything below ₹3 lakh/month is losing money; the founder is subsidising the aggregator.
The Mumbai cloud kitchen setup guide breaks the capex and location-selection down further. And because aggregator commission is the single largest post-food line item, most operators eventually work through the 7 ways to reduce Swiggy and Zomato commission playbook to drag it from 32% toward 25%.
When a cloud kitchen is the right choice
The cloud kitchen model earns its keep in four specific situations.
- First-time operator learning delivery. Lower capex means the tuition on menu, packaging, aggregator ops, unit economics and staff scheduling costs ₹5-25 lakh instead of ₹30-80 lakh. If the concept dies, the founder is bruised, not bankrupt.
- Existing chain expanding to a new city or new format. A Delhi brand testing Bengaluru does not need to sign a 5-year high-street lease. Cloud kitchen for 12 months, learn the demand, then decide whether to open a dine-in flagship or scale to 3 more cloud units.
- Virtual brand experiment on top of an existing kitchen. An existing dine-in has spare capacity between 3 PM and 7 PM; a virtual brand fills that white space with almost zero incremental cost. Fail-fast at aggregator-listing speed.
- Delivery-heavy cuisine. Biryani, momo, sandwich, wrap, poke bowl, salad bowl, dessert, ice cream — food that either travels well or is inherently packaged. This is 70% of Swiggy and Zomato order value already.
When a cloud kitchen is the wrong choice
Equally, cloud kitchen is a bad answer for three concept shapes.
- Chef-driven dine-in concept. If the value proposition is a tasting menu, live-fire cooking, plating theatre, or a chef's counter — delivery destroys everything the brand is selling. The concept needs a room.
- Brands that live on face-to-face guest relationships. Neighbourhood cafes, private-dining supper clubs, community-focused restaurants where the manager knowing every regular's order is the point. Delivery is transactional; those brands are relational.
- High-touch cuisine that fails in packaging. Crisp dosa, live-fire steak, tempura, soufflé, ice-sculpted plating, anything where the 20-minute rider journey turns a ₹600 dish into a soggy apology. Some food is not delivery food. Force-fitting it into a cloud kitchen model kills the brand's future dine-in prospects too.
The honest test: if a customer opening the packaging at their dining table gets 80% of what they would get at your dine-in table, cloud kitchen is viable. Below 60%, you are selling the customer a bad first impression and burning the brand slowly.
The Online eMenu 2-product stack for cloud kitchens
Every cloud kitchen needs two software layers: something that consolidates the digital order flow (aggregators + WhatsApp + direct), and something that handles the actual billing, GST-compliant receipts and end-of-shift reconciliation. Online eMenu ships both.
Ordering Suite — ₹199/month
Cloud-hosted, monthly billing, no lock-in. Connects Swiggy Partner, Zomato Partner, WhatsApp ordering, and a brand-owned direct-ordering page into a single kitchen queue. Multi-brand routing means one kitchen can run 3-8 virtual brands on the same POS with per-brand menus, per-brand pricing, per-brand reports. KDS (kitchen display) prints or displays KOTs by station. Channel-revenue report at end of shift attributes every rupee to Swiggy, Zomato, WhatsApp or direct — the number the owner needs to answer "should I invest more in direct-ordering next month?"
Desktop POS — ₹4,999/year
Windows on-premise billing. Works offline (relevant for basement kitchens with patchy internet). Prints GST-compliant receipts. End-of-shift Z-report groups cash + card + UPI + aggregator + KOT variance in one screen. Cash-drawer variance auto-flagged above ₹200. One-time yearly licence, no per-outlet fee.
Live in 48 hours
Setup for a single-brand cloud kitchen is typically 24-48 hours end to end — menu import, aggregator connections, POS install, one training session with the head cook and packer. See the pricing page for the full breakdown or the products page for the feature-by-feature spec.
Launch your cloud kitchen on both channels in 48 hours
Ordering Suite ₹199/month + Desktop POS ₹4,999/year. Swiggy, Zomato, WhatsApp, multi-brand routing, GST-compliant receipts. No lock-in.
Start 14-day free trial → WhatsAppFrequently asked questions
What is a cloud kitchen in simple terms?
A delivery-only restaurant. No dine-in, no walk-in, no signage the customer sees. It exists to receive orders on Swiggy, Zomato, WhatsApp or direct, cook, pack and hand to a rider. India had ~400 in 2018 and roughly 4,000 in 2026.
What is the difference between cloud kitchen, ghost kitchen and dark kitchen?
They mean the same thing. Cloud kitchen is India + GCC vocabulary, ghost kitchen is US, dark kitchen is UK + Europe. Same operational model, different regional labels.
How does a cloud kitchen actually work?
Four steps: order captured on Swiggy/Zomato/WhatsApp, order lands in POS and prints a KOT, kitchen cooks and packs to spec, aggregator or in-house rider delivers. Typical cycle 25-40 minutes.
What are the 4 cloud kitchen models?
Single-brand, multi-brand (virtual brands), aggregator-hosted (Rebel Foods, EatClub, Kouzina), and in-restaurant delivery-only expansion.
Why are cloud kitchens growing so fast in India?
Rent ratio 2-4% vs 10-12% for dine-in, fitout ₹5-25 lakh vs ₹20-80 lakh, launch in 30-45 days vs 6-9 months, and aggregators already have the rider fleet.
What does a cloud kitchen cost to run in India?
Mumbai unit: rent ₹40-60K, staff ₹80K-1.2L, aggregator commission 25-32%, food cost 28-30%, packaging 4-6%, ad spend 5-8%. Break-even around ₹3-4 lakh monthly revenue.
When is a cloud kitchen the wrong choice?
Chef-driven dine-in concepts, relationship-heavy neighbourhood brands, high-touch cuisines that suffer in packaging. Delivery destroys what those concepts are selling.
How much does aggregator commission actually cost?
Base 18-26% plus restaurant-funded discounts, ad spend, packaging fees and payment charges push the all-in rate to 25-32% of gross order value.
Do I need a POS or is the aggregator dashboard enough?
Above 2 orders/day per channel, yes. Aggregator dashboards do not consolidate, do not track KOT-vs-sales, do not manage multi-brand, and cannot power WhatsApp or direct ordering.
Does Online eMenu support multi-brand cloud kitchens?
Yes — Ordering Suite at ₹199/month routes 3-8 virtual brands from one kitchen with separate menus and reports. Desktop POS at ₹4,999/year handles billing and Z-reports. Live in 48 hours.