Opening a restaurant in India takes roughly 60–90 days and, depending on format, ₹5 lakh (a cloud kitchen) to ₹1.5 crore+ (fine dining). The critical path is licences, FSSAI, the municipal trade licence, Shops & Establishment, fire NOC, and in some cities a police eating-house licence, started in parallel with the fit-out, never after it.
What does it cost to open a restaurant in India?
The honest answer is a range, because format is destiny:
| Format | Typical initial investment | The big line items |
|---|---|---|
| Cloud kitchen | ₹5–15 lakh | Kitchen equipment, deposits, packaging, licences |
| Takeaway / QSR counter | ₹8–25 lakh | Frontage deposit, kitchen line, branding |
| Café (25–40 covers) | ₹20–60 lakh | Interiors, espresso equipment, furniture |
| Casual dining (40–80 covers) | ₹40 lakh–1 crore | Fit-out, HVAC, kitchen, deposits, opening team |
| Bar & dining | ₹80 lakh–2 crore+ | Liquor licence, interiors, compliance |
Three rules survive every format. First, the deposit is not the budget: a ₹1 lakh/month space in a metro typically demands 6–10 months as deposit plus advance rent, before a single chair arrives. Second, keep 20% of the total as working capital for the first 90 days, most new restaurants die of cash flow, not concept. Third, price the licences and professional fees honestly: ₹50,000–2 lakh across the stack depending on city and format.
Fixed costs ÷ (average ticket × gross margin) = covers per day to break even. If the answer exceeds 60% of realistic capacity, the lease is the problem, walk away before the deposit does it for you.
Which licences does a restaurant need?
The core stack, in the order you should start them:
- FSSAI licence, the food-business licence everything else assumes. Registration for tiny turnover, state licence for most restaurants, central for large/multi-state operations. Apply on FoSCoS; see the full FSSAI guide for types, fees and documents.
- Municipal trade / health licence, from your city corporation (BMC in Mumbai, MCD in Delhi, BBMP in Bengaluru, GHMC in Hyderabad, PMC in Pune, GCC in Chennai). Usually the slowest step after fire NOC, start it the week you sign the lease.
- Shops & Establishment registration, the state labour registration; needed for hiring and often for opening a current account.
- Fire NOC, required above seating/area thresholds and enforced for basements and rooftops; your architect should design to it from day one, not retrofit for it.
- GST registration, mandatory past ₹20 lakh turnover (₹10 lakh in special-category states) or from day one if you're on aggregators. Restaurants bill 5% without input credit; the GST guide covers the trade-offs.
- City-specific extras, Delhi's police eating-house licence, excise for liquor anywhere, music licences (PPL/IPRS) if you play recorded music, signage permission, lift and boiler certificates where relevant.
Treating licences as sequential. FSSAI, trade licence and fire NOC have independent timelines, file all three the week the lease is signed, and chase them in parallel while the fit-out runs.
How long does it take to open?
A realistic 90-day map for a café or casual-dining room, assuming the location is finalised:
Days 1–15, paper and demolition. Lease registered; FSSAI, trade licence, fire NOC and Shops & Establishment all filed; kitchen layout frozen (changing it later is the most expensive decision you'll ever revisit); contractor mobilised.
Days 16–60, the build. Civil and electrical work; kitchen equipment ordered against the frozen layout (imported equipment adds 3–6 weeks); POS, billing and printer stack chosen and configured; menu engineered and costed per dish, not "roughly", per dish.
Days 61–80, the assembly. Staff hired and trained on the actual systems; supplier accounts opened with two vendors per critical ingredient; recipes standardised with yields; menu printed only after cost checks; soft-launch inventory purchased.
Days 81–90, the rehearsal. Friends-and-family service treated as a full dress rehearsal, billing, kitchen tickets, day-close, all of it. Fix the process, not just the food. Open when a full Friday-night simulation ties out at the cash drawer.
How many covers do you need to break even?
The math is simple and unforgiving. Run yours now:
Break-even covers = fixed costs ÷ (average ticket × gross-margin%). Gross margin = 100% − food-cost% (65–72% is typical for Indian casual dining). Every cover beyond break-even contributes that margin straight to profit. CountStand's day-close tells you nightly where you stand against this line.
Two patterns from real operations are worth internalising. Restaurants that reach operational break-even do it in 3–6 months; the initial investment typically returns over 18–36 months. And the single best predictor of survival isn't cuisine or décor, it's whether the owner knows their daily break-even number cold and watches it nightly. That habit is precisely what a day-close that reconciles automatically buys you.
What should the software stack look like from day one?
Opening with spreadsheets and a bill book means rebuilding your operations twice. The day-one stack for 2026: GST-compliant billing that works offline (networks fail on opening weekend too), kitchen tickets on a screen rather than paper, recipe-linked inventory so food cost is measured rather than felt, and reporting an owner actually reads. One system doing all four, from ₹999/month, costs less than the bill books, and the data compounds from the first bill: three months in, you'll know your stars and your leaks by name.
How much money do I need to open a small restaurant in India?
A small café or QSR realistically needs ₹8–25 lakh including deposits, fit-out, equipment, licences and 90 days of working capital. A cloud kitchen compresses that to ₹5–15 lakh by deleting the dining room. Budget 20% extra as working capital, undercapitalisation kills more restaurants than bad food.
Can I open a restaurant without an FSSAI licence?
No, operating a food business without FSSAI registration/licence is an offence with penalties up to ₹5 lakh, and every other registration downstream assumes it. It is also the licence customers and aggregators can check. Start it first; it is the cheapest item on the list.
Is a restaurant profitable in India?
Well-run rooms net 8–15% after everything; poorly-run ones lose money at full tables because margin leaks through food cost, theft and mispricing. The difference is measurement discipline, knowing the break-even, the per-dish margin and the variance weekly.
Should I start with a cloud kitchen instead?
If capital is under ₹15 lakh or the concept is unproven, yes, a cloud kitchen tests demand at one-fifth the capex. The trade-off: you inherit delivery economics (18–30% commissions), which makes a direct-order channel and per-dish costing essential from day one.