Smart city project failures
Every entry in this atlas that failed or was scaled back — 58 of 651 published projects — grouped by the way it went wrong, with the sources that say so.
Most writing about smart cities is written by the people selling them, so the record is missing its most useful half. This atlas keeps failures permanently, checked against public sources like everything else, because the projects that went wrong are the ones with something to tell a city that has not spent the money yet.
Two labels, kept apart. Failed (21) means it stopped, was cancelled, or demonstrably did not do what it was built to do. Scaled back (37) means it runs, or ran to its end, but delivered materially less than promised — most of these are live services today. Together that is 8.9% of the atlas, which is a floor and not an estimate: an entry only earns a failing outcome when a named source says so, and quiet abandonment rarely has one.
✔ Every published entry is checked against the public sources it cites before publication.
📉 The measurement worked. Nothing followed it.
The test: Did the instrument do its job, and fail because nobody was obliged to act on what it said?
This is the most common way a smart-city project disappoints, and the least visible from the control room: the sensors report, the dashboard updates, and no lever is attached to the other end. Chicago heard the gunshots and its own Inspector General could not say what the police response produced. Los Angeles ran crime forecasting for nine years and its own watchdog still could not measure what it changed. Miami-Dade lowered the temperature at which a heat alert fires, and then the state removed the county's power to attach any obligation to that alert. In each case the technology is not the thing that failed.
9 entries in this pattern
🧪 The pilot never became a service
The test: Did it work at trial scale, and stop when the trial, the grant or the vendor did?
A pilot is designed to end. That is not a criticism of pilots — it is the reason so few of them turn into anything, and the reason a press release about a trial tells you almost nothing about a city. Stockholm's driverless bus is the instructive case in this group, because it did graduate: five years as a scheduled public route, and then cancelled for the oldest reason in transport, too few passengers. Wellington retired 3,800 in-ground parking sensors after eight years. Helsinki's Whim was widely called the world's first commercial mobility-as-a-service app, and its operator filed for bankruptcy in 2024.
13 entries in this pattern
🧩 Half a system
The test: Was the piece that would have made it work cancelled, deferred or never funded?
These projects were delivered. What was not delivered is the other half, and without it the delivered half does the wrong job. Accra bought the buses and the contactless fare cards, then cancelled the 9.1 km of dedicated lane — so a service designed to beat the shared minibuses now queues behind them. Baghdad installed 149 water meters that were never networked, so readings still needed a site visit twice a year. Dhaka wired 29 junctions with automated signals that went largely unused while police kept directing traffic by hand. Buenos Aires bought signalling that can run a train every 2 minutes 15 and runs one every 3, because the fleet is what sets the service.
9 entries in this pattern
✋ People said no
The test: Was it stopped by residents, a court or a regulator rather than by the technology?
Nothing in this group broke. Sidewalk Labs' Toronto waterfront was abandoned in May 2020 after years of privacy backlash, before construction began. Hong Kong had to disable its smart lampposts' number-plate recognition and Bluetooth speed detection after protesters sawed several down. A court suspended Buenos Aires's facial recognition system, and the reason it could is worth reading closely: the statutory limit existed, but the oversight commission was never seated. Stockholm legislated a zero-emission zone, installed the cameras, and an appeals board blocked it days before launch.
8 entries in this pattern
📜 The contract decided it, not the technology
The test: Did the money structure — concession, guarantee, permit, commission — set the outcome before the system was switched on?
The most consequential decision in these projects was signed years before anything was installed. Busan and Gimhae guaranteed a private operator's revenue against a consultant's forecast of 176,358 daily riders; 30,084 came, and the two cities have been paying the difference ever since. Manila paid its camera contractor out of the fines, in a programme whose stated purpose was removing the financial incentive from enforcement. Three South American bike shares — Bogotá, Santiago, São Paulo — cost their cities almost nothing and, as each entry works out in its own words, bought them almost nothing to govern.
11 entries in this pattern
🏗️ Announced far larger than built
The test: Is the gap between what was promised and what stands the main fact about it?
These are the projects the phrase "smart city" is usually arguing about, and they are a minority of the failures in this atlas — worth saying plainly, because they dominate the coverage. The Line went from a 170 km linear city to a 2.4 km first phase and then to suspended construction inside three years. Masdar was launched as the world's first zero-carbon city and had its target downgraded to low-carbon by 2050. Delhi built two 24-metre air-purifying towers; its pollution control committee later told the National Green Tribunal that cutting the city's pollution 17% this way would take about 40,000 of them. The gap does not always open at the announcement — Barcelona built four of a planned one-in-three green streets before a new administration ruled out the rest, and Madrid's participatory budget halved after a change of mayor.
8 entries in this pattern
Common questions
How many smart city projects fail?
Nobody can answer that honestly, because failures are the entries that never get written up. What this atlas can say is what its own record contains: of 643 published, source-checked projects, 21 are recorded as failed and 37 as scaled back — about 9% — and those figures describe our reading of the evidence we could verify, not a global failure rate. A project only earns a failing outcome here when a named source says so, so the true share is certainly higher than 9%: the quiet abandonments are exactly the ones with no source to cite.
What is the most common reason smart city projects fail?
In this record it is not cost, procurement or technical difficulty. It is that the measurement works and nothing is obliged to happen next. Chicago's gunshot sensors, Los Angeles's crime forecasting and Miami-Dade's heat alerts all functioned as instruments; what none of them had was a body required to act on the reading, or an evaluation that could say what the response achieved.
What is the difference between a failed and a scaled-back project?
Failed means the thing stopped, was cancelled, or demonstrably did not do what it was built to do. Scaled back means it is still running, or ran to its end, but delivered materially less than it promised — a smaller footprint, fewer functions, a narrower mandate. Most scaled-back entries are still live services. The distinction matters because the two are routinely blurred: a bike share that shrank from fourteen districts to five is not the same story as one whose operator went bankrupt.
Are famous smart city failures like NEOM and Masdar typical?
No, and that is the most useful thing on this page. Megaproject collapses get the coverage, but they are eight of the fifty-eight entries here. The failures that actually recur are quieter: a pilot that ends with its grant, a system delivered without the half that would have made it work, a contract signed against a demand forecast that never arrived.
Why does this atlas publish failures at all?
Because a catalogue that only contains successes is a marketing brochure, and because the entries that went wrong are the ones with something to teach a city that has not spent the money yet. Failed and scaled-back entries stay in the record permanently, with the same source-checking as everything else and the same links to the primary evidence.
Where to read next
Three routes out of this page: the essay on why smart cities fail argues the pattern; a reading is not a decision takes the first group above apart in detail; and Failed & Instructive is the short curated route for people who want six entries rather than 58.