Why the 2 a.m. train matters more than the 8 a.m. one.

At eight in the morning, a transit system is easy to notice. Platforms fill, buses arrive in lines, and the city’s schedules seem to agree that this is when the day begins. At two in the morning, the same system reveals who it believes the city is for. The cleaner finishing a shift, the nurse heading home, the hotel worker arriving before dawn, the warehouse employee whose clock does not follow the office calendar—all still need a way to get there.

I am Ines Halvorsen, a former transit planner. I learned to look at the first and last trips as more than timetable details. A network that stops at midnight is not a network for someone whose work ends at 12:30. “There is a bus tomorrow” is not much comfort if tomorrow’s bus leaves before tonight’s shift is over.

The riders outside the peak.

Night and early-morning riders are often working hours that many other jobs do not require. Hospitals run around the clock. Cleaning crews work after offices close. Hospitality, food service, logistics, and freight depend on staff outside the conventional commute. Their trips can be difficult to shift by an hour, and many do not have another person or vehicle waiting if transit stops running.

The familiar rush-hour picture can hide this. If agencies evaluate service only by the busiest period, they risk making the network excellent for a narrow slice of the day and unavailable to people whose work keeps the rest of the city operating. A route can look quiet at 2 a.m. and still be essential. The question is not simply how many people are on one vehicle at one moment; it is whether people can reliably reach jobs and return home.

A city is not finished when the office lights go out. Neither should its transit network be.

The overnight network.

Cities solve late-night coverage in different ways. London introduced the Night Tube in 2016 on selected Underground lines, alongside its established night-bus network. New York’s subway, with 24-hour operation, remains a classic example of a system that does not close for the night. Neither model is a universal template: the shape of demand, infrastructure, staffing, and maintenance needs differ.

Some cities use “owl” buses: routes designed for overnight travel, often with a simpler network and timed connections at a hub. The goal is not necessarily to duplicate every daytime route. It is to connect major job centers, hospitals, residential areas, and transfer points in a pattern a tired person can understand at the end of a shift.

The real cost, including maintenance.

Running through the night is not free. Vehicles, operators, supervisors, cleaning crews, and control staff all have to be scheduled. Rail infrastructure also needs maintenance, and overnight hours are often the most practical window to inspect tracks, repair equipment, and complete work that cannot safely happen with trains moving. A 24-hour service can mean shorter or more complicated maintenance windows, or a need to plan work around service.

That is a real operational trade-off, not a reason to dismiss night riders. Agencies can use planned closures with clear replacement service, maintain selected corridors continuously, or offer overnight buses where rail work requires a shutdown. The answer should be based on the network’s needs and the trips people make, not an assumption that everyone can wait until morning.

Trust is part of the service.

People decide whether they can live without a car based on the whole day, including the trip home. If the last bus leaves before a shift ends, a rider has to arrange a backup: a ride, a taxi, a long walk, or a car they may not otherwise need. That makes transit less dependable even during the hours it runs. A published last trip is not enough if it cannot be trusted to arrive or connect.

Agencies can start by publishing the actual service span and headways in plain language, then coordinating night routes around the trips that workers need. They should ask employers and riders about shift times, include early-morning connections, and explain maintenance-related changes well in advance. Where service is sparse, a dependable hourly route can be more useful than a faster pattern that disappears without warning.

The 2 a.m. train may carry fewer people than the 8 a.m. one. It may still matter more to the passenger standing on the platform at 2 a.m., because there is no next departure in ten minutes and no easy alternative at home. A transit network proves itself not only at the busiest hour, but at the hour when the people who need it have the fewest choices.

— Ines