How to Start a Taxi Service: The Complete 2026 Blueprint.

The U.S. taxi and limousine industry is roughly an $11 billion annual revenue market with an industry-average profit margin near 2.7%, according to IBISWorld data summarized in 2023. That headline number hides the real structure: Uber and Lyft took the discretionary, app-native, credit-card-paying urban rider a decade ago and are not giving that segment back. What remains for a traditional taxi company is the demand that rideshare serves badly or refuses outright. That is the market you are buying into, and you should underwrite it as a contract and dispatch business, not a hail business.

The paying segments in 2026 are specific. First, non-emergency medical transportation (NEMT) brokered through state Medicaid contractors such as Modivcare, MTM or Access2Care, which pays per-trip rates that are unglamorous but predictable and pays on net 30 to net 60. Second, municipal and county contracts: paratransit overflow, senior transport vouchers, and city taxi-voucher programs for late-night workers. Third, airport queue work at regulated taxi stands where the airport authority still issues a limited number of taxi permits, which is a real barrier to entry that protects you once you are inside. Fourth, hotel front-desk and hospital discharge accounts, where a concierge or case manager wants one phone number and one invoice. Fifth, cash and unbanked riders, elderly riders who will not use an app, and riders in neighborhoods where rideshare surge or driver refusal is chronic.

Size the local opportunity honestly before you spend. A well-utilized city taxi in a mid-size market grosses roughly $80,000 to $150,000 per year in fares, though that spread is wide and utilization-driven, and the low end is common in a market where you have no contracts. Three cars at the midpoint is roughly $300,000 to $400,000 in gross revenue. At a 2.7% industry net margin that is $8,000 to $11,000 of net profit, which is why almost every viable small taxi operator either owns the vehicles and drives one, or holds contracts that pay above street rates. The claim circulating online that a typical taxi company owner earns about $297,000 in annual profit is an industry-wide average distorted by large fleet operators and is not achievable at 3 to 5 vehicles.

Regulatory posture matters more than in most trades. Cities like New York, Chicago and Los Angeles have medallion or permit caps and detailed vehicle standards; mid-size cities often have an ordinary operator's permit process; some states preempted local taxi regulation when they passed TNC statutes, which can mean less paperwork but also no protection from unlimited competition. Verify which of those three worlds your city is in before you write a check, because the answer changes both your startup cost and whether the demand you are chasing exists at all.

Your siting decision is not about retail visibility, because nobody walks into a taxi office. It is about deadhead miles. Every mile driven without a paying passenger is pure loss at roughly $0.55 to $0.75 per mile in fuel, tires and depreciation. The yard should sit within a 10 to 15 minute drive of your densest trip origin cluster, which for most viable small operators means a hospital and dialysis corridor, an airport, or a hotel district. A cheap yard 25 minutes out of the demand zone costs you two dead miles on every trip and will quietly consume more than the rent you saved.

The physical requirement is modest: secure off-street parking for the fleet, a 200 to 600 square foot dispatch room with reliable broadband and a landline number, and ideally a bay or a nearby contracted mechanic. Deposit plus first month runs $2,000 to $5,000 in low-cost markets and $10,000 to $25,000 in expensive ones, with ongoing rent of $1,500 to $5,000 per month. Many one to two car operators skip the office entirely, dispatch from home and park at their residence, which is legal in most cities but check your zoning code for commercial vehicle parking restrictions, because a single code complaint can force a move mid-quarter.

Channel is the more consequential choice. There are four: street hail and stand, phone dispatch, contract and broker work, and your own app. Street hail is dead outside a handful of dense districts and airport queues. Your own app will not out-compete Uber and building one at $50,000 to $300,000 is capital destruction. The two channels that actually pay a small operator are phone dispatch to a repeat customer base of elderly, unbanked and mobility-limited riders, and contract work through NEMT brokers, county paratransit overflow, hotel accounts and school or agency transport.

Airport access deserves its own decision. Airport authorities typically issue a limited number of taxi permits with a separate application, a per-trip or annual fee, and a dedicated queue. That barrier cuts both ways: it costs you time and fees to get in, and it protects your trip volume once you are in. If an airport permit is currently open in your market, it is usually worth pursuing before you expand the fleet, because a stand position produces predictable trip flow that a cold phone line does not.

Work the numbers per trip, not per month, or you will not see where the money leaks. Take a typical mid-size-market metered trip of about 6 miles with a $16 to $22 fare, or an NEMT ambulatory trip that a broker pays at a base rate plus a per-loaded-mile rate. Against that fare you pay: driver compensation, which is the dominant cost at either 40 to 50% of the meter under a commission split or roughly $16 to $22 per hour loaded under a W-2 model in most markets; fuel at $0.12 to $0.20 per mile for a hybrid and $0.18 to $0.30 for a conventional sedan; maintenance, tires and depreciation at roughly $0.15 to $0.25 per mile on a used vehicle; and card processing at 2.6 to 3.5% on card fares.

That leaves a contribution margin of roughly 20 to 35% of the fare before any fixed cost. Now load the fixed costs. Per vehicle per year: insurance $3,000 to $12,000, per-vehicle licensing $150 to $550, dispatch software allocation $600 to $4,000, and your share of rent, phone staffing and administration. On a car grossing $80,000 to $150,000 a year, insurance alone consumes 3 to 12% of gross revenue. That is why the industry-average net margin is about 2.7% and why small operators who guess at these numbers instead of tracking them per vehicle usually discover the loss at tax time.

The two variables that decide whether you clear anything are utilization and deadhead ratio. A car that produces 12 to 18 paid trips per shift is profitable; a car producing 5 to 7 is not, because the fixed cost per trip roughly doubles. Similarly, a deadhead ratio of 30 to 40% of total miles is normal and survivable; 55% or higher, which happens when you cover a sprawling territory with too few cars, converts a profitable rate card into a losing one. Track paid miles divided by total miles per vehicle every single week.

One practical consequence: contract and standing-order work is worth taking at a lower per-mile rate than street work, because it arrives with a scheduled pickup, a known destination and a return leg. A dialysis run three days a week for the same patient produces near-zero deadhead on the return and predictable receivables. A random hail at a higher rate that ends 9 miles outside your zone can be the worse trip once you count the empty drive back.

At one to two vehicles you are the owner, the primary driver, the dispatcher and the bookkeeper, and that is the correct structure. Do not hire before you have volume. At three to five vehicles the structure changes: you need enough credentialed drivers to seat every licensed vehicle across the shifts you intend to run, plus at least one spare. Running two shifts on four cars realistically means seven to nine drivers on the roster given absences and turnover. Whether they are W-2 employees or lease-drivers who pay a daily or weekly cab lease is a serious legal question in states with strict worker classification tests such as California's ABC test; get a labor attorney's opinion before you set the model, because a misclassification finding carries back wages, penalties and payroll taxes.

The role most owners skip and most regret is the dispatcher. A competent dispatcher answering the phone from 6 a.m. to 8 p.m., assigning trips by proximity rather than by seniority, and calling passengers back when a driver is running late is the difference between a hotel account that grows and one that quietly stops calling. At $17 to $24 per hour in most markets, one full-time dispatcher plus your own coverage on evenings and weekends is the minimum credible staffing. Software assigns trips; a person keeps accounts.

You also need a maintenance relationship, not necessarily a maintenance employee. A contracted independent shop that will take your cars on same-day priority, at a negotiated fleet labor rate, is usually cheaper than an in-house mechanic until you pass roughly 10 vehicles. Get preventive maintenance on a schedule (oil at 5,000 miles, brakes and tires inspected monthly) because a taxi accumulates 40,000 to 70,000 miles a year and reactive maintenance means unplanned idle days.

Finally, one part-time bookkeeper or an outsourced bookkeeping service at $300 to $900 per month. Taxi accounting has specific traps: per-vehicle cost centers, driver split reconciliation, cash fare handling, broker receivables aging at net 30 to net 60, and fuel card fraud. Owners who run this on memory and a shoebox do not find out they are losing money on two of four cars until the year is over.

Days 1 to 30 are institutional, not consumer. Walk in, in person, with a business card, a certificate of insurance and your permit numbers, to every one of these: the discharge planning office of each hospital within your zone, every dialysis and infusion center, every skilled nursing and assisted living facility, the two or three largest hotels, and the regional network manager for your state's Medicaid NEMT broker (Modivcare, MTM, Access2Care or the state equivalent). Ask one question: are you short of capacity on any run right now, and what would it take for us to be on your list. Case managers and front desk managers are the actual buyers here, and they are reachable without an appointment.

Days 30 to 60 are about proving reliability at small scale. Take the unglamorous runs the incumbents refuse: 5 a.m. dialysis pickups, Sunday discharges, wheelchair-assist trips if you have an accessible vehicle. Report your own on-time percentage back to each account weekly in a two-line email. Institutional buyers switch vendors on missed pickups, not on price, and a documented 95% on-time record after six weeks is a stronger sales asset than any advertising you can buy for the $1,000 to $5,000 most operators spend on branding.

Run the consumer channel in parallel but cheaply. Claim and fully complete the Google Business Profile with the phone number as the primary action, because a large share of local taxi demand still starts with a maps search for taxi near me. Get a memorable local number, put it on the doors and roof light in numerals large enough to read from a sidewalk, and leave printed cards at senior centers, VFW and American Legion halls, dialysis waiting rooms, bars that close at 2 a.m., and hotel front desks. The unbanked and elderly rider who will not use an app is your most loyal repeat customer and is reached by phone number and paper, not by digital advertising.

Days 60 to 90: convert the accounts that responded into written standing orders and negotiate a corporate account with monthly invoicing for the two or three highest-volume ones. A signed standing order that produces 8 to 25 trips per day is what makes the fourth vehicle safe to buy. If by day 90 you have no contract, no airport permit and no repeat phone base, your revenue is street-hail dependent in a market Uber already owns, and the honest move is to shrink to one owner-driven vehicle rather than finance more capacity into a channel that is not there.