The cloud kitchen pitch is seductive: no dining room, a third of the capex, launch in six weeks. The restaurant pitch is older: real margins, brand gravity, customers who are yours and not an app's. Both models mint winners and busts — the difference is fit between model, market and operator. Here is the honest financial comparison, line by line, and the decision logic.
Capex: the headline gap
| Item | Cloud kitchen (600 sqft) | Casual dine-in (1,200 sqft, 40 seats) |
|---|---|---|
| Deposit & fit-out | ₹6–12 lakh | ₹25–60 lakh |
| Kitchen equipment | ₹5–10 lakh | ₹8–15 lakh |
| Licences & tech | ₹1–2 lakh | ₹2–4 lakh |
| Launch marketing | ₹1–3 lakh | ₹2–5 lakh |
| Typical total | ₹13–27 lakh | ₹37–84 lakh |
The kitchen wins the entry ticket decisively — which is exactly why its competitive moat is shallow: what costs ₹15 lakh to enter, a hundred others can also enter. The restaurant's capex is partly a fortress.
The P&L truth: where the margins actually land
- Cloud kitchen revenue line: aggregator commissions of 18–30% + ads/visibility spends of 5–10% come off the top — an effective 25–35% channel cost before you cook anything; discounts often ride on top
- Restaurant channel cost: near-zero on walk-ins; its 'commission' is rent — but rent buys the acquisition machine (location footfall) and the margin on dine-in beverages/add-ons kitchens never see
- Typical steady-state EBITDA: well-run cloud kitchens land 8–18% (single brand) — multi-brand operations from one kitchen push this up; well-run dine-ins land 12–22% with alcohol pushing higher
- Ticket economics: kitchens live on ₹250–450 AOVs with delivery constraining category (biryani travels; soufflé doesn't); restaurants monetise occasions — the same customer pays 2–3× per visit dine-in
A cloud kitchen's demand faucet belongs to Swiggy/Zomato: commission hikes, ranking-algorithm shifts and ad-auction inflation arrive by notification, and your 'customers' are the platform's data, not yours. Restaurants own their footfall but marry a location — one bad street decision is unfixable. Choose which dependency you can manage: platform risk or property risk.
Break-even math, worked
Cloud kitchen: fixed costs ~₹1.8–2.5 lakh/month (rent, staff of 4–6, utilities); at ₹350 AOV and ~30% contribution after food cost + channel cost, break-even sits near 55–75 orders/day — a real number that takes most kitchens 4–9 months of ad-fuelled climbing. Restaurant: fixed costs ~₹4.5–7 lakh/month (rent, staff of 12–18); at ₹300/cover contribution (~55–60% gross margin, no commissions), break-even needs ~60–80 covers/day — achievable at one good location from month one, or never at a bad one. Payback: kitchens 18–30 months when they work; restaurants 24–48 — slower, but on a bigger, more defensible profit pool.
The decision logic
- Choose cloud kitchen when: capital is tight, you're testing a cuisine/brand, the category travels well, you'll operate multi-brand from one kitchen (the model's real margin unlock), or you're extending an existing restaurant's reach
- Choose dine-in when: you hold a strong location insight, the concept is experience-led, alcohol/beverage margins apply, or you're building a brand for franchising (dine-in brands franchise at multiples kitchens don't)
- The hybrid that increasingly wins: a compact dine-in (15–25 seats) doing 40–60% delivery — location acquisition + platform reach, one rent
- Whatever the model: own-channel orders (WhatsApp/website, even 15–20% of volume) are the profitability swing factor — build them from day one
How Aidwish helps
Aidwish models both formats for food founders — location and catchment analysis, format-wise P&L projections, licence stacks (identical FSSAI core, different municipal layers), and the aggregator/own-channel strategy — then executes the chosen build end to end.