Traffic Lights Will Disappear From Smart Cities

The Question

The world's first electric traffic light was switched on at the corner of Euclid Avenue and East 105th Street in Cleveland, Ohio, in 1914. It was a brilliant solution to a specific problem: humans in fast machines cannot be trusted to negotiate an intersection by eye contact. So we built a system of coloured lamps that stops half the traffic so the other half can move — and then we accepted, for over a century, that sitting motionless at an empty crossroads is simply what driving is.

Add it up and the absurdity shows. The average urban driver loses days every year waiting at red lights, engines burning fuel to go nowhere, while the cross street sits vacant. The traffic light is not traffic management. It is a confession that vehicles cannot talk to each other. But now they can — and the question is when the first city admits that the coloured lamps, like the gas streetlight and the telephone box, have finished their work.

What the Evidence Shows

The theoretical case was made at MIT's Senseable City Lab, whose slot-based intersection research modelled crossroads where vehicles talk to each other — a technology called V2X, short for vehicle-to-everything — and book a "slot" to pass through, like aircraft landing at an airport. No lights, no stopping, just continuous interleaved flow. The models showed such intersections could handle roughly double the traffic of a signalised junction while nearly eliminating the stop-start waves that cause congestion and waste fuel.

The practical case is already on the street. Pittsburgh's Surtrac system — adaptive signals that watch traffic in real time and re-time themselves — cut travel times by about 25% and idling by over 40% in its deployment corridors. Google's Green Light project uses driving data to re-tune signals in cities from Seattle to Jakarta, reporting up to 30% fewer stops at optimised junctions. These systems still use lights, but they are the transitional species: signals that think are the step before signals that vanish.

There is even a low-tech preview. Dutch and Scandinavian towns pioneering "shared space" urbanism removed signals, signs, and kerbs from town centres — forcing everyone to negotiate by attention rather than obedience — and recorded fewer accidents, not more. Drachten in the Netherlands became the famous case: junctions that handled 20,000-plus vehicles a day flowed better and crashed less without lights. Human intersections can already work unsignalised at moderate speeds. Machine intersections can work unsignalised at any speed.

Why This Is Happening

The connected-car fleet is quietly reaching critical mass. V2X hardware is being built into new vehicles as a safety feature, letting cars broadcast position, speed, and intention to each other and to city infrastructure. Robotaxi fleets in San Francisco, Phoenix, and Wuhan already navigate complex junctions autonomously thousands of times a day. Once a district's traffic is mostly networked, the light becomes redundant hardware.

Cities are desperate for capacity they cannot build. No mayor can widen downtown streets, but MIT's numbers suggest signal-free coordination could double intersection throughput using the asphalt that already exists. That is the political jackpot: congestion relief with no construction, funded by removing infrastructure rather than adding it.

The climate math is embarrassing. Idling at signals burns billions of litres of fuel a year globally and every stop-start cycle spikes emissions. Adaptive-signal pilots have already documented double-digit emission cuts from simply reducing stops. For cities with legally binding climate targets, the intersection is low-hanging fruit that gets riper every year the vehicle fleet gets smarter.

What Could Happen

What Can We Do

The intersection of 2035 is being designed now, in city budgets and vehicle standards. There are practical ways to benefit early and shape what arrives.

Support adaptive signals in your own city today. Systems like Surtrac and Green Light are cheap, proven, and cut your commute now. When your council debates transport budgets, smart signal retrofits deliver more time back per dollar than almost any road project — ask for them by name.

When you next buy a car, check for V2X support. Connected-vehicle features will increasingly determine which lanes, districts, and priority schemes your car can use — and its resale value into the 2030s. It is the new equivalent of buying a car with GPS in 2005.

Defend the pedestrian in every consultation. The signal-free intersection is optimised for vehicles; walkers and cyclists must be designed in, not patched in. Residents who show up to planning meetings asking one question — "how does a child cross this street?" — will shape these districts more than any lobbyist.

Ask the equity question early. When intersections become algorithms, priority becomes a product someone can sell — express slots for those who pay. Cities that ban paid priority before deployment will avoid the ugliest version of this future. Raise it before the system ships, because rules are softest at the start.