The cloud kitchen pitch was compelling: skip the real estate premium, skip the front-of-house staff, skip the ambience capex, and capture the delivery boom. EBITDA margins of 20–25% were the number floating around in 2021. In 2026, we have enough data to interrogate that number.
The original promise
A 400 sq ft dark kitchen in an outer-ring suburb of Bengaluru costs ₹40,000–70,000/month in rent. A 1,500 sq ft dine-in venue in the same suburb costs ₹1.8L–2.5L. The rent saving is real. The question is what fills the gap.
Real cost structure
Add these up: 72–88% of GMV is absorbed before EBITDA. The promised 20–25% margin assumes a food cost of 28%, commission of 22%, and labour of 12% — all at the optimistic end of each range, simultaneously. In practice, most cloud kitchens land at 8–13% EBITDA if they are well-run, and negative EBITDA if they are not.
Aggregator dependency
A dine-in restaurant can build its own customer base. The regulars walk in because they love the place. A cloud kitchen has no footfall. 95%+ of orders come from Swiggy or Zomato. If the aggregator delists you (for low ratings, non-compliance, or commercial reasons), the revenue goes to zero overnight. This concentration risk is the biggest operational vulnerability in the model.
Breakeven math
At a ₹400 average order value and 30% effective commission, each order nets the kitchen ₹280 before food cost. At a 28% food cost, contribution per order is ₹168. At ₹70,000/month rent + ₹1.8L in labour and utilities, the kitchen needs 1,488 orders per month (≈50/day) to cover fixed costs. Most cloud kitchens in Tier 1 cities do 35–55 orders/day — so breakeven is achievable, but not easy.
Multi-brand strategy
The practical fix is running multiple virtual brands from the same kitchen. A single kitchen running three brands — North Indian, biryani, and healthy wraps — can serve the menu needs of each segment with 70% overlapping ingredients. Each brand has its own Zomato listing. Volume per square foot goes up. Fixed costs are distributed across three revenue streams.
Before launching a second brand, map ingredient overlap. If two brands share fewer than 70% of their raw materials, the inventory complexity will eat your margin.
Who should do it
Cloud kitchens make sense for operators who already have a successful dine-in kitchen with spare prep capacity, or for culinary entrepreneurs testing a brand concept before a full dine-in investment. They do not make sense as a first restaurant — the lack of direct customer feedback, the aggregator dependency, and the thin margins make failure too quiet and too fast.
15 years in restaurant operations across 3 continents. Former GM of a 5-star hotel restaurant in Mumbai.