How to start a coffee shop in 2026

Not a café, no — a sit-down café with seating starts around $80,000 in the US and runs to $350,000, and no lender funds that on an idea. What you can start close to nothing is the coffee business underneath it. Coffee Shop Startups prices event and farmers-market espresso catering, and a “piggyback” coffee bar operating inside somebody else's shop, at $5,000 to $15,000 — one grinder, one machine, a table and a permit. A used cart or trailer sits above that, roughly $25,000 to $95,000 depending on how much of the fit-out is already done. Keep the job while you run it, learn what your town actually buys and at what price, and let the cart pay for the lease later. Your website costs nothing either: you can build one free with Zarla and list your Google Business Profile for nothing. What you cannot do on no money is sign a lease and start a build-out, because the rent runs for months before the first coffee is sold.

It is the Pareto principle — named after the Italian economist Vilfredo Pareto, who noticed in 1906 that a small share of causes produced most of the effect — applied to a coffee shop. In practice it shows up three ways. Roughly a fifth of your customers, the regulars who come most mornings, produce most of your revenue, which is why remembering someone's order is a financial act and not just a nice one. A handful of drinks account for most of your tickets, which tells you what to keep on the menu, what to prep ahead and what to stop stocking. And a short window of the day produces most of the sales, which is where your staff should be rostered. Treat it as a prompt to go and look, not a law: pull your own sales mix out of your point-of-sale system after a month and you will know your real numbers rather than a rule of thumb.

It can be, and it is thin. Coffee has a high gross margin per cup and a small net margin per year, because rent, labor and waste eat the difference. Crimson Cup's published planning model shows the shape: a $4 average purchase and 175 purchases a day gives $21,000 a month, from which drink ingredients take up to 40% of sales, payroll 30% or less and rent 15% — leaving 15% to cover utilities, insurance, supplies, professional fees, borrowing costs, taxes and profit. Across US restaurants generally, the National Restaurant Association put a typical pre-tax margin at roughly 5% of sales before the pandemic, and reported in July 2026 that 42% of operators said their restaurant was not profitable in 2025. Four levers decide which side of that you land on: your average ticket (food attach rate is the biggest one), transactions per day, labor as a share of sales, and rent as a share of sales. A shop that gets all four right is a good business; a shop that gets rent wrong cannot fix it with better coffee.

Yes — but not as a café with tables. $50,000 buys you a cart, a trailer or a small prefabricated drive-thru stand, all of which start around $25,000 in the US, and it leaves you something for a deposit, an espresso machine and the first months of operating cash. It does not buy a sit-down café: those run $80,000 to $350,000 because the build-out alone is roughly $150 to $300 a square foot once plumbing, ventilation and an accessible restroom are involved. The trap with $50,000 is spending all of it on the opening. Crimson Cup's advice is to hold six months of operating expenses in cash, so a realistic $50,000 plan looks more like $25,000 to $30,000 of cart and equipment and the rest untouched in the bank. Taking over a second-generation space — somewhere that was already a food business, with the hood, traps and sinks in place — is the other way $50,000 stretches further than it should.

Lower than the number you have heard, and the honest answer comes with a caveat: nobody publishes a survival rate for coffee shops specifically. The closest official US measure is the Bureau of Labor Statistics' Business Employment Dynamics data for accommodation and food services, where recent cohorts of new establishments show roughly 85% surviving their first year and a little under 60% still open after five. The famous claim that 90% of food businesses fail in year one is a myth: H.G. Parsa's study of independent restaurants in Columbus, Ohio, published through the Cornell Hotel and Restaurant Administration Quarterly, found about 26% closed in the first year, 19% in the second and 14% in the third, and Parsa said he could find no evidence for a 90% rate anywhere in the literature. What closes shops is not exotic — rent set above what the site's traffic can carry, a build-out that ate the working capital, and an owner who never got a day off. Both data sets are revised, so check the current BLS release rather than a figure printed in any guide.

In the US the list is set by your city, county and state rather than by Washington, but it is almost always the same eight items. A business license or registration (Seattle's 2026 tiers, for instance, are $73 under $20,000 of taxable revenue and $147 up to $500,000). A food service establishment permit with a health inspection — the core one, at $258 to $773 a year in Texas depending on sales volume where the state issues it, and $280 in New York City. Health department plan review of your floor plan before you build. Food handler cards for staff: California requires one within 30 days of hire from an ANSI-accredited provider, valid three years, and ServSafe's online course and assessment is $15. A certified food protection manager, which many states require on site, following the FDA Food Code's person-in-charge rule — the Food Code is a model that states adopt rather than federal law, so the version in force is your state's. A sales tax permit. A certificate of occupancy and a sign permit from your city. And a liquor license only if you intend to pour alcohol. Which of these apply, and what they cost, genuinely varies by county — confirm your own list with your local health department and your accountant before you sign a lease.

Three things, in this order. A professional website carrying your hours, your address and a map, your food and drink menu, what you are brewing this week and photographs of the actual room — fast on a phone, because that is what somebody is holding on their commute. A Google Business Profile connected to that website, which is what puts you on Google Maps and in the local pack for “coffee shop near me”, with hours you keep accurate, including holidays. And a steady flow of reviews, much of the “prominence” half of Google's local ranking, asked for while the customer is still holding the cup. Zarla builds that site free in about a minute — menu, hours, photos — and wires it to the Google Business Profile people actually find you through.

Between about $25,000 and $350,000 in the US, and the format decides where you land. Reconciling the published cost guides that show their numbers — Crimson Cup's 2025 breakdown, KORONA POS's 2026 one and Coffee Shop Startups — a mobile cart or trailer runs $25,000 to $95,000, a kiosk or walk-up stand $50,000 to $150,000, a purpose-built drive-thru $100,000 to $300,000 (small prefabricated stands start nearer $25,000), a full sit-down café $80,000 to $350,000, and seating plus a drive-thru $120,000 to $400,000 and up. Most of the spread is build-out at roughly $150 to $300 a square foot, cheapest in a space that was already a food business. The itemized pieces you can price today: a new two-group commercial espresso machine at $5,500 to $23,000 at US dealers, or a refurbished one at $4,000 to $14,500; grinders at $1,000 to $4,200 each; a food service establishment permit at $258 to $773 a year in Texas or $280 in New York City; a business license at $73 to $147 in Seattle; food handler cards at $15 each. And on top of the build: six months of operating expenses in cash, which is Crimson Cup's rule and the one most new owners skip.

You can, and the owners who answer this question on public forums all give the same advice: go and work in one first. Their reasoning is practical rather than gatekeeping — you cannot price a menu, roster a morning rush, judge a lease or read a roaster's quote from the outside, and the people who skip that step tend to discover it after the build-out money is spent. Two lower-risk ways in exist. Start with a cart, a trailer or event catering for a year, where the downside is tens of thousands rather than hundreds of thousands and every mistake is cheap. Or buy an existing shop with a trading history, then go and count its customers yourself before you believe the seller's numbers — one poster weighing two drive-thrus in 2023 did exactly that and watched three cars come through in fifteen minutes. Either way, plan your own pay down: experienced owners told one prospective buyer who needed $50,000 a year that it was the wrong expectation for a first-year shop, and that $30,000 would be doing well.