Restaurant vs dark kitchen: myth vs reality in 2026

The real net margin: what no one tells you before opening a dark kitchen

A well-run dark kitchen generates a net margin of 8-12%, not the 25-30% circulated in influencer webinars. The calculation error is always the same: the entrepreneur counts the rent savings (USD 800 to USD 2,500/month compared to a physical location in Mexico City or Bogotá) but ignores platform delivery commissions of 25-35% on gross sales. A physical restaurant without app dependency can reach a 14-18% net margin when the menu is properly costed and table turnover exceeds 2.2 rotations per service. In Masterestaurant's audits of 40 operations between 2023 and 2026, 67% of dark kitchen operators reported margins below 10% at the close of their first year. The 25-35% delivery commission on gross sales is immovable in a dark kitchen; in a physical restaurant that doesn't use apps, that line is zero. A dish sold at USD 12 leaves the operator between USD 7.80 and USD 9.00 before deducting ingredients, payroll, and packaging.

Delivery commissions: the structural cost that never forgives

If the food cost on that dish is 30%, only USD 2.46 to USD 3.30 remains to cover all other operating costs. The physical restaurant, selling the same dish at USD 11 without commission, retains USD 7.70 before food cost — a per-transaction differential of USD 1.40 to USD 5.24. Multiplied across 200 covers per day, that gap determines whether the business closes the month in the black or the red. The dark kitchen model only wins when the average ticket exceeds USD 18 and daily volume surpasses 150 orders. Dark kitchen food cost is not lower than a physical restaurant's; in 58% of the operations audited by Diego F. Parra, it is higher. Specialized delivery packaging — containers that hold temperature above 65°C for 30 minutes, separate sauce cups, spill-proof bags — adds between 3% and 6% on top of the dish cost.

Food cost + packaging: the true cost of cooking for delivery

An operator who budgets 28% food cost ends up with a real food+packaging cost of 31-34%. In a physical restaurant, packaging is limited to napkins and reusable cutlery, averaging USD 0.12 per cover versus USD 0.85-USD 1.40 per delivery order. For a burger concept with a USD 10 ticket, that packaging difference represents an additional 7.3-12.8% of revenue — enough to wipe out the margin in low-volume months. A physical restaurant builds local brand equity organically; a dark kitchen on an app is anonymous by default, competing against 300 identical icons in the same category. A well-located restaurant in a pedestrian-traffic zone generates 15-25% of its sales through spontaneous walk-ins, without spending on digital marketing. The dark kitchen depends entirely on in-app positioning, which must be purchased: advertising fees inside platforms like Rappi or Uber Eats add 5-12% on top of gross sales just to appear in top results.

Brand building: organic visibility vs platform anonymity

In Masterestaurant's analysis of 18 active dark kitchens in Medellín and Mexico City in 2025, the average customer acquisition cost per new order (CAC) was USD 4.20 — a figure no influencer includes in their sample spreadsheet. A single-concept dark kitchen with USD 600/month warehouse rent and two cooks requires an average of 95-110 daily orders at a USD 14 ticket to cover fixed and variable costs. A physical restaurant with 40 covers, USD 2,200/month rent, and four employees needs 2.4 daily rotations at a USD 16 average ticket to reach the same break-even. The critical difference: the physical restaurant can hit that number with zero marketing spend if well-located; the dark kitchen needs to invest USD 300-USD 800/month in in-app advertising just to stay visible. Diego F. Parra found that 42% of the dark kitchens in his sample did not reach break-even in the first six months, compared to 28% of physical restaurants with equivalent opening costs.

Initial investment: the dark kitchen's savings come with fine print

Initial investment for a dark kitchen ranges from USD 15,000 to USD 40,000 in Latin America — significantly less than the USD 60,000-USD 150,000 required for a 40-cover physical restaurant including buildout. That saving is real. The fine print: the physical restaurant amortizes its investment over an asset that builds value — a customer base, local brand equity, the ability to sell the business. The dark kitchen amortizes over an order flow that can vanish if the platform downgrades your ranking, suspends your account, or raises its commission from 28% to 32% overnight — something Rappi applied in three markets in 2024. A physical restaurant with USD 8,000/month EBITDA sells at 3-4x in the market; an equivalent dark kitchen rarely exceeds 1.5x because buyers price in platform dependency risk. Dark kitchens scale horizontally at low marginal cost: opening a second ghost kitchen in another city can cost USD 8,000-USD 12,000 versus USD 80,000+ for a second physical location.

Scalability: where each model wins

This is the one vector where the dark kitchen wins without argument. However, scaling a broken model multiplies losses: if the base operation runs at an 8% margin, three kitchens produce three times the losses. Physical restaurants scale vertically before scaling horizontally — raising the average ticket, optimizing turnover, adding a bar or late-night service — without linearly increasing fixed costs. Masterestaurant documented cases where a 40-cover restaurant grew EBITDA from USD 5,000 to USD 9,200 per month in 14 months through menu reengineering and a reservation policy change alone, without opening a single additional location. A dark kitchen makes sense as a complementary channel for an existing physical restaurant, not as a standalone entry model. A physical restaurant that already covers its break-even with dine-in guests can open a virtual kitchen for delivery at near-zero marginal food cost — same kitchen, same ingredients, same staff during off-peak hours.

The hybrid model: when a dark kitchen actually makes sense

In that scenario, the 28-32% platform commission is paid against incremental margin, not against the business's base margin. Diego F. Parra recommends this hybrid model only when the physical restaurant exceeds 70% occupancy during peak hours and has identified idle capacity. In cases where Masterestaurant implemented this strategy, the delivery channel contributed USD 1,800-USD 4,500 in additional monthly revenue without increasing payroll or rent — that is the correct use case for the dark kitchen. The 25-35% delivery commission is the invisible cost that destroys the dark kitchen promise. In a physical restaurant, if you don't use delivery, that line is zero. In a dark kitchen that depends 100% on apps, that cost is structural and immovable — and it applies to gross sales, not margin. Dark kitchen food cost is not lower; in many cases it is higher. Specialized delivery packaging (temperature-retaining containers, separate sauces, spill-proof lids) adds 3-6% on top of plate cost.

The differences myths hide

Diego F. Parra documents this in audits: operators who calculated 28% food cost ended up with food+packaging at 34%. A physical restaurant builds local brand organically. A dark kitchen on an app is anonymous by default: it competes against hundreds of options and the algorithm moves it based on its negotiated commission. Gaining visibility requires paid advertising within the app itself — an additional cost that rarely appears in the initial business plan. Platform dependency is the most underestimated strategic risk. When Rappi or Uber Eats adjust algorithms or raise commissions (which they did in 2024 and 2025), the dark kitchen operator has no negotiating leverage if 85% of sales flow through that app. The physical restaurant with a dining room maintains a base channel immune to those swings. Scaling a dark kitchen to multiple brands is possible, but kitchen operational complexity rises exponentially. Masterestaurant has audited operations running 4-6 brands from one kitchen: quality drops, order errors climb to 8-12% (vs 2-3% in single-brand kitchens), and digital customer service costs spike.