A smart district is a single neighbourhood run as one technical system — shared energy, sensors, a common data platform and mobility wired at the scale of a few blocks instead of a whole city. It is the smallest unit a city can rebuild from scratch, and the scale at which you can still see who moved in and who paid.
This guide covers 35 documented projects in 27 cities and 19 countries. The energy machinery beneath them is in the smart grids guide, ownership in our ownership essay, the wider argument in why smart cities fail.
| City | Instrument | What the record shows |
|---|---|---|
| Barcelona | Rezoning plus underground rebuild, 2000 | ~37 km of streets relaid on €180m+ of public money — the only peer-reviewed grade here, and it states no impact figure |
| Zurich | Utility-run district energy, 30-year contract | ~1,800 t CO₂/year, a city projection; the label it was built to earn was scrapped in 2024 |
| Helsinki | Agile-piloting programme, 2013–2021 | 25 pilots in the district, 60+ replicated elsewhere; then it closed on schedule |
| Helmond | 1,500 homes sharing energy, water and data | About 52 built; a 2023 reset traded much of the technology for affordable housing |
| Manchester | £10M national IoT demonstrator, 2016–2018 | Ended on schedule; its own domain has lapsed and now redirects to a gambling site |
Less than the brochures suggest, and the gap is countable: only 12 of the 35 entries state an impact figure at all, and 23 of the graded entries rest on official or operator sources. Exactly one is peer-reviewed — 22@ Barcelona, the oldest here at 2000 — and that grade covers the sourcing of what was built, not a result, because the entry states no impact figure either.
Where numbers exist, the developer usually produced them. Fujisawa SST reports a 70% cut in CO₂ against a 1990 baseline — and the company that built the town is the only party measuring it. Ülemiste City's savings and Virtual Shibuya's 3 million visits are operator counts; Punggol's 3,000 MWh is a projection. The firmest figure comes from a retrofit with a baseline — GrowSmarter cut 3,333 MWh a year across three 1960s buildings — a comparison a district on empty land cannot make.
Read the outcome labels the same way. Six entries are time-boxed programmes past their end year, and what separates the four marked succeeded from the two marked scaled back is whether a standing organisation picked up a piece: Helsinki's piloting model went to other Finnish cities, and what outlasted Greenwich's €25m lighthouse project was EV charging built into existing lampposts. CityVerve's £10M award named nobody to own the platform after month 24. "Succeeded" here usually means the programme reached its contracted end, not that the neighbourhood still runs on what it built.
Six entries are marked scaled back, and not one shrank because the technology stopped working. The atlas's failure taxonomy files three under announced far larger than built: The Line went from a 170 km blueprint for 1.5 million residents to a 2.4 km first phase, then to suspended construction; Brainport has about 52 of 1,500 planned homes; Senegal's digital city restarted only in late 2025 after a multi-year stall. Porto Maravilha is filed under the contract: all 6,436,722 development-rights certificates went in one 2011 auction to a single public fund, and fewer than 10% were ever resold. Buiksloterham was stopped by physics — the grid operator could supply about a quarter of the connection capacity requested, so peer-to-peer trading became flexibility management. None of the six is filed under people said no.
The rest of the pattern is occupancy: of the ten from-scratch districts here, six are still under construction and one suspended. Woven City's first residents are Toyota's own employees, toward a phase-one target of about 360; Tengah has more than 12,000 of roughly 30,000 launched flats occupied.
The practical takeaway: a smart district is a real-estate project with a data layer, and it fails on real-estate terms — occupancy, financing, a grid connection, an owner for the platform once the grant ends. The ones still working are mostly those added to an existing city rather than raised on empty or reclaimed land, where a standing utility or agency had reason to keep them running.
A single neighbourhood operated as one technical system: shared energy, sensors, a common data platform and mobility wired at the scale of a few blocks rather than a whole city. In this atlas, 35 documented projects across 27 cities and 19 countries carry the tag. Energy is the most common category, accounting for 14 of them — typically a district heating, cooling or micro-grid system serving every building on the site.
The record favours the retrofit. Of the ten districts here built from scratch, six are still under construction and one is suspended, and three of the six scaled-back districts come from that group. Districts added to an existing city — 22@ Barcelona's rezoned Poblenou, Stockholm's refurbished 1960s blocks in Årsta — are far more often operating today, and they also have an existing building stock to measure results against, which a district on empty land does not.
Six entries are time-boxed programmes that have reached their end year. None was cancelled; all ran to their agreed finish. What separates them is whether a standing organisation picked up a piece: Helsinki's piloting model was carried to other Finnish cities, Copenhagen's utility extended flexible district heating to another neighbourhood, and Greenwich's lamppost EV charging was widely copied. Manchester's CityVerve named nobody to own and pay for its platform after month 24, and its project domain has since lapsed.
Thinly. Only 12 of the 35 entries state an impact figure at all, and 23 of the graded entries rest on official or operator sources. Exactly one carries a peer-reviewed grade — 22@ Barcelona — and that grade covers the sourcing of what was built rather than a measured outcome, since the entry states no impact figure. Headline numbers from Fujisawa SST, Ülemiste City and Virtual Shibuya are all self-reported by the operator, with no independent audit located.
Four European districts were wired up as time-limited demonstration projects with an agreed finishing date. That is the honest way to run an experiment — and it puts the interesting question at the end rather than the start.
| Project | Since | Status | Outcome | Evidence | Run by | Cost | What survived the funding |
|---|---|---|---|---|---|---|---|
| Smart KalasatamaHelsinki, 🇫🇮 | 2013 | Completed | Succeeded | Independent | City government | not public | The method outlived the site: the agile-piloting model was carried to other Finnish cities after the district programme closed. |
| EnergyLab Nordhavn Smart Energy DistrictCopenhagen, 🇩🇰 | 2015 | Completed | Succeeded | Official | Research institution, City government, Utility or public operator | not public | The practice moved rather than stopped — the utility later extended flexible district heating to another neighbourhood. |
| GrowSmarterStockholm, 🇸🇪 | 2015 | Completed | Succeeded | Official | City government | not public | The retrofits stayed and the replication did not: deep renovation pays off, but ordinary landlords had no financing model to copy it. |
| CityVerve IoT Smart City DemonstratorManchester, 🇬🇧 | 2016 | Completed | Scaled back | Official | City government, Private company | £10M (Innovate UK/DCMS award) | Nothing said who would own or pay for the platform after month twenty-four, so the most durable artefact was a set of case studies — and the project's domain now redirects to a gambling site. |