clm-37e3b0ffdfe06707
quantitative
captured over 160,000 high-resolution aerial images in 41 days for a nationwide 0.1-metre reality mesh
Singapore spent about S$73 million on a national 3D city model that made headlines worldwide. A decade on, no source documents a single decision an agency has taken with it — and the twins Singapore actually operates are different platforms.
This overview was generated with AI from the public sources listed below and checked against them before publication. It has not had a complete human fact-check — treat it as a starting point and follow the sources.
How this atlas is made · What “last verified” means · What “source-checked” means · The full record behind this page
✎ Updated 20 Aug 2026 Publication date recorded for 3 sources. · 3 revisions in all What changed →
The most striking number in this story is not the price. It is 41 — the number of days it took to photograph the whole country from the air.
Virtual Singapore launched on 3 December 2014 as part of the country's Smart Nation drive — a research programme led by the National Research Foundation in the Prime Minister's Office, with the Singapore Land Authority supplying the survey data and GovTech the engineering. In 2015, Dassault Systèmes signed on to build the platform on its 3DEXPERIENCity software. The scale of the underlying mapping remains remarkable:
The promise matched the scale. Government material from 2017 described what agencies would be able to do: find every roof surface above a certain height and compute the sunlight falling on it, simulate crowd evacuations, test how a new building changes wind and shade at street level, plan barrier-free routes for the elderly. The platform was "targeted to be ready by 2018 for use by government agencies. Thereafter, it will be deployed to the public progressively."
| When | What happened | Who says so |
|---|---|---|
| Dec 2014 | Launched as part of the Smart Nation drive, led by the National Research Foundation | Government |
| Jun 2015 | Dassault Systèmes signs on to build the platform on 3DEXPERIENCity — its press release names no budget | Vendor |
| Mar 2017 | "Ready by 2018 for use by government agencies. Thereafter, it will be deployed to the public progressively." | GovTech |
| Sep 2018 | The one global headline: S$73 million, security fears, rollout on computers cut off from the internet | Reuters |
| Sep 2021 | Public 3D access finally arrives — as OneMap3D, a separate SLA platform built on open-source technology, without Dassault | Government |
| Feb 2022 | The national 3D mapping is completed; this is the concrete thing "finished in 2022" provably refers to | Trade press |
| Mar 2024 | Singapore's ten-year Geospatial Master Plan is announced — it does not mention Virtual Singapore at all | Government, by omission |
| Mar 2025 | The port gets an operational twin of its own, the Maritime Digital Twin — a new platform, not Virtual Singapore | Government |
Internationally, Virtual Singapore had exactly one big media moment. In September 2018, Reuters reported that the model was so detailed that officials worried about it falling into the wrong hands — "planners and plotters", as the syndicated headline put it. The programme's own director said the quiet part aloud:
This information will help our daily lives, but it could also fall in the wrong hands and create problems for Singapore.
— George Loh, director of programmes, National Research Foundation (Reuters, September 2018)
The government's answer was internet separation: Virtual Singapore would run only on computers physically cut off from the web. Searches for standalone coverage by the Guardian, the BBC or Wired come back empty; the world's most famous digital twin is famous, in large part, on the strength of one wire story and a vendor's marketing.
The most-cited real-world use is more modest than the legend. In the Yuhua housing estate, SLA staff modelled solar panels and connected them to energy-bill data — panels that were already installed under an existing HDB programme. In the SLA's own words: "We went to the site and looked at the solar panels and modeled them in Virtual Singapore. Then we worked with local authorities to get data from energy bills." That is a visualisation of a decision someone else had already made — genuinely useful, and categorically different from the twin choosing where the panels should go. The other documented users are research institutions: A*STAR ran large-scale wind, noise and traffic simulations; NUS received access for research.
Put the standard claims about Virtual Singapore next to what a source check finds, and the gap is consistent in one direction:
| The claim you will read | What the record shows |
|---|---|
| "Agencies simulate flood risk with it" | The water and aviation authorities part-funded the underlying data collection; no flood decision taken in the twin is documented anywhere |
| "It plans rooftop solar" | The Yuhua showcase modelled panels that were already installed — the installation decision was made elsewhere |
| "The public will get access" | Promised in the future tense since 2017; what shipped in 2021 was OneMap3D, a separate platform |
| "It transforms planning decisions" | The OECD's 2024 case study — the best independent account — lists capabilities only: no adoption figure, no documented decision, no outcome |
A 2024 peer-reviewed survey of urban digital twins names the field's central gap in one line: "a significant gap identified is the limited number of real-world implementations." Then the name itself went quiet. The NRF's Virtual Singapore programme page is gone; so is the Smart Nation initiative page. There was no cancellation notice and no closing report — the project simply stopped being something the government talks about.
Here is the detail that turns an absence of evidence into a finding. When Singapore needed a digital twin that runs something, it did not reach for Virtual Singapore — it built new platforms. Punggol Digital District's Open Digital Platform, built by JTC and GovTech, is described by its operator as "completely operationalized … fed live data mined from the sensors of the physical facility management systems". The island's heat-planning questions went to a separate Digital Urban Climate Twin developed with the Singapore-ETH Centre. In March 2025, the port got its own Maritime Digital Twin. And the mapping foundation — the genuinely finished piece — lives on as maintained public infrastructure beneath all of them.
Fairness requires the counterweight: a classified system that runs on air-gapped computers produces no public usage trail by design. Virtual Singapore may well be consulted inside agencies in ways no press release records. But that is precisely the point for anyone citing it as proof that digital twins transform city government: the world's flagship case offers no public evidence of a single transformed decision, and the twins Singapore demonstrably operates were built elsewhere.
Virtual Singapore succeeded completely at the thing its surveyors could control — the model — and left no public record of the thing its famous name implies: operation. For cities shopping for a twin, three questions follow directly from this story:
Our comparison of Singapore, Tokyo, Seoul and Rotterdam takes the access question across four operating models; the table just below sets this entry beside its three closest peers.
The Singapore Land Authority captured over 160,000 high-resolution aerial images in 41 days for a nationwide 0.1-metre reality mesh, plus roughly 25 TB of street data from 600 billion point-cloud points and 3 million panoramic images — refreshed every five years by air and every two years at street level. What no source documents is a single dated decision a planning agency has taken with the platform; the OECD's 2024 case study describes capabilities only.
Fame is not operation: when Singapore needed twins that actually run something, it built new platforms — Punggol's sensor-fed Open Digital Platform and the port's Maritime Digital Twin — while the world's most cited city model stayed a sealed simulation environment. Its security classification cuts both ways: the absence of a public usage trail is not proof of non-use.
Two clocks, kept apart: when this entry’s sources were published, and what has happened to its record in the atlas. Documented span: 2014–2022. 6 of 10 sources carry a publication date, from 28 March 2017 to 26 March 2025.
Ten sources carry this entry. The load-bearing ones are the OECD’s 2024 case study, the best independent account available; GovTech’s own 2017 description of what the platform was to do; the single Reuters wire story from 2018 that produced almost all of the international coverage; and the operator pages for the three separate twins Singapore actually runs. Dassault’s press release is used only for the fact that it signed, never for a capability claim. A word about method: much of this page is an argument from absence, and absence is the hardest thing to source. What was checked is that the OECD study, the government material and the trade coverage each describe capabilities rather than decisions, and that searches for standalone coverage by the Guardian, the BBC or Wired come back empty. Four things the record cannot answer: whether agencies use the platform internally, which an air-gapped system would not reveal; whether it was ever formally wound down, since no notice exists; what the S$73 million figure covers, as it appears in one wire story and no budget document; and whether the removed government pages were retired or merely moved. Anyone who can supply a primary document for these improves this page — the “Report an error” link at the foot of this page is how.
Evidence Passport · public beta 14 addressable claims, 10 source dossiers, one permanent baseline. The passport is unreviewed — an AI-assisted structural migration, not a human or expert endorsement. Every revision keeps a permanent copy under revision history.
The text below is copied verbatim from the published record. Each claim names the source dossiers that carry it; caveats stay attached to the claim they qualify.
clm-37e3b0ffdfe06707
quantitative
captured over 160,000 high-resolution aerial images in 41 days for a nationwide 0.1-metre reality mesh
clm-36eeabc4fae78a81
quantitative
roughly 25 TB of street data from 600 billion point-cloud points and 3 million panoramic images
clm-037052c47cff7e99
descriptive
refreshed every five years by air and every two years at street level
clm-484eead4348af846
quantitative
the OECD's 2024 case study describes capabilities only
clm-6ffe39cf774c1670
descriptive
when Singapore needed twins that actually run something, it built new platforms — Punggol's sensor-fed Open Digital Platform and the port's Maritime Digital Twin
clm-7bba6c4228eeef03
quantitative
Virtual Singapore launched on 3 December 2014 as part of the country's Smart Nation drive
clm-f0cdfef268ffc7b9
quantitative
In 2015, Dassault Systèmes signed on to build the platform on its 3DEXPERIENCity software
clm-8ab98527dd132b4c
quantitative
In September 2018, Reuters reported that the model was so detailed that officials worried about it falling into the wrong hands
clm-97d702ba4e260f5d
descriptive
Virtual Singapore would run only on computers physically cut off from the web
clm-02c971dc7c23d9b8
descriptive
In the Yuhua housing estate, SLA staff modelled solar panels and connected them to energy-bill data
clm-dd772c139f9d52e6
descriptive
A*STAR ran large-scale wind, noise and traffic simulations; NUS received access for research
clm-f29fd43deba48da2
descriptive
The NRF's Virtual Singapore programme page is gone; so is the Smart Nation initiative page
clm-63f1d71e2f110b06
descriptive
Punggol Digital District's Open Digital Platform, built by JTC and GovTech
clm-bd1a4fe91df25a82
quantitative
In March 2025, the port got its own Maritime Digital Twin
This layer records what is known about each source, what is still missing, and the exact snapshot this page is rendering. A missing date, archive or hash stays visible as missing.
Awaiting operator and domain-expert review; publication of this revision implies neither.
sha256:051acdbab24b522d0e6a1cc78b50837f29a37692cd24303c33373f32d7bf32e6
src-c8cce305aea3273eThe government's own 2017 framing: ready for agencies by 2018, public access 'progressively' thereafter — every application described as a capability, not a completed use.
src-a5362d572693ca0fBest independent case study; names SLA as lead and a planned subsurface twin, but records no documented agency decision and no outcome figure.
src-1262fbdd835ade0fThe programme's one big international media moment: S$73 million figure, NRF director George Loh's misuse warning, and the 'internet separation' rollout.
src-3d63098b8831abfaDocuments the 2022 completion of SLA's national 3D mapping: 160,000+ aerial images in 41 days, 0.1 m reality mesh. Bentley (vendor) savings claims inside are not used here.
src-0e0dd5ffff3873fdSLA Land Survey Division authors: 600 billion point-cloud points, 3 million panoramic images (~25 TB), and the five-year aerial / two-year street refresh cycle.
src-5610cc8e574b6b63OneMap3D launched 15 September 2021 as SLA's public 3D platform — the publicly delivered piece of the 3D-Singapore vision, separate from Virtual Singapore.
src-d4f2bcea1081ead1Singapore's first Maritime Digital Twin, launched 24 March 2025 by MPA with GovTech — an operational twin built as a new platform.
src-6b936743b75acefdJTC/GovTech's district twin described by its operator as 'completely operationalized' and fed live sensor data — the operational contrast to Virtual Singapore.
src-c60afdac64b8e688The platform vendor's own 2015 announcement; notable for what it omits — no budget figure appears in it.
src-4e7ff17397602dcfFullest account of the Yuhua pilot with SLA's own words: existing solar panels were modelled and linked to energy bills — a visualisation of what was already built.
The same label covers a sealed national security asset, a browser tab anyone can open, a bookable simulation lab and a live operations dashboard. Access is the axis that separates them — and it decides what each twin can honestly claim to do.
| Project | Since | Status | Outcome | Evidence | Run by | Cost | Who gets in |
|---|---|---|---|---|---|---|---|
| Virtual Singapore: National 3D City Model and Digital TwinSingapore, 🇸🇬 | 2014 | Completed | Scaled back | Official | National government, Private company | About S$73 million (~US$53 million), as reported by Reuters; the former NRF programme page is offline. | Sealed: rolled out to agencies behind 'internet separation'; the promised public access never launched. |
| Tokyo Digital TwinTokyo, 🇯🇵 | 2021 | Live | Ongoing | Official | City government | not public | Open to anyone: a free browser viewer, with ward point clouds released as open data in December 2024. |
| S-Map — Seoul 3D Digital Twin PlatformSeoul, 🇰🇷 | 2021 | Live | Ongoing | Official | City government | not public | The lab in between: agencies and companies run simulations on roughly 600,000 buildings through its Open Lab. |
| One Network Unified Management (city operations digital twin)Shanghai, 🇨🇳 | 2019 | Live | Ongoing | Peer-reviewed | City government | not public | The operations pole: a live twin coordinating transport and emergency response rather than planning studies. |
1001 Smart Cities (2026). “Virtual Singapore: National 3D City Model and Digital Twin” — Singapore, Singapore. The Smart City Atlas. https://1001smartcities.org/projects/singapore-virtual-singapore/ (last verified 2026-08-07). Data: CC BY 4.0.
The underlying data is free to reuse with attribution — see the open data page.