clm-c146b87f64a6c1ba
descriptive
fully automated, with no driver
A driverless railway opened between Busan and Gimhae in 2011 against a forecast of 176,358 riders a day. About 30,000 came. Sixteen years later the two cities have paid 906.8 billion won and owe at least a trillion more — for a line that works exactly as built.
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.
16 sourced claims on this page, none individually verified yet — turn on “Sources in the text” to see each one.
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 8 sources. · 2 revisions in all What changed →
Part of 1001 Smart Cities — 1,167 documented projects · 🏙️ 9 more in Busan · 🚌 Public transport guide · 🚲 Mobility projects
There is nothing to see in the cab, because there is no cab. The two-car trains that shuttle between Sasang in western Busan and Gimhae run fully automated, with no driver, on an elevated guideway threaded above a river plain that was rice fields within living memory. The line is 23.9 km long and has 21 stations. It opened in September 2011 and has been running, punctually and unremarkably, ever since.
It is also one of the most expensive transport mistakes in South Korean local government — and the mistake is not in the railway.
Korean light rail of this generation was built through a concession model: a private consortium designs, finances and builds the line, transfers ownership to the state, and operates it for a fixed term. The private side will only accept that deal if the revenue is predictable — so the public side guarantees it. The instrument is called a minimum revenue guarantee, and its entire logic rests on one input: how many people the forecast says will ride.
For this line the forecast came from the Korea Transport Institute, the state's own transport research body. It projected 176,358 riders a day in the opening year, rising to 322,545 by 2031. Those numbers went into the contract, and the contract turned them into an obligation: whatever the fare box failed to deliver against them, the two cities would make up.
On the day it opened, the line carried 30,084 riders — about a sixth of the promise. Nothing about the railway had gone wrong. The catchment simply did not contain the people the forecast had placed in it.
It is worth pausing on how that figure travels. The English-language Wikipedia article on this line describes 176,000 a day as the system's design capacity, which reads as an engineering specification — a ceiling, a thing the railway can do. It is not. It is the demand forecast, the number somebody expected to actually show up, and the near-identity of the two figures is the tell. A capacity ceiling and a demand projection landing within four hundred riders of each other is not a coincidence; it is the same number wearing a different label. Read as capacity, the line looks merely underused. Read as forecast, it is the whole story.
| When | What happened | Who says so |
|---|---|---|
| Sep 2011 | The line opens; 30,084 riders a day against a forecast of 176,358 | Regional press (Gyeongnam Domin Ilbo) |
| 2011 | First year of support payments — of the order of one billion won | Regional press |
| Mar 2017 | The minimum revenue guarantee is replaced by a cost-compensation arrangement running to 2041 | Regional press / operator |
| 2020 | Gimhae alone pays 48.6bn won (about $34m); ridership 34,752 against a forecast of 272,220 | Press (Busan Ilbo) |
| Sep 2021 | Ten-year mark: the line has never covered its costs in a single year | Press (Busan Ilbo) |
| 2024 | Ridership 45,821 a day — 14.9% of the 306,000 forecast for the year | Broadcast (SBS) |
| 2024 | Annual support payment of 84.1bn won (about $60m) | Broadcast (SBS) |
| Dec 2025 | Gimhae's accumulated light rail liability is national news | Press (Kyunghyang Shinmun) |
The guarantee did not promise the operator a profit. It promised that a gap between forecast and fare box would be filled from the public purse — which means the two cities did not take on the risk that the railway would break down or be badly run. They took on the risk that a consultant's spreadsheet was wrong. That risk has no engineering remedy. Running more trains cannot fix it, and neither can running fewer.
The bill grew in the shape you would expect from an obligation indexed to a gap that never closes. Gimhae's support payment for 2020 alone was 48.6 billion won (about $34 million), against ridership of 34,752 a day where the forecast had promised 272,220. By 2024 the annual payment across both cities had reached 84.1 billion won (about $60 million). The two are not equal partners in this: Gimhae carries 63.19 percent of the cost and Busan 36.81 percent — the smaller city, with the smaller budget, holds the larger share.
Across the whole period the totals are these. Busan and Gimhae paid 906.8 billion won in support payments over 2011-2025 (about $640 million), split 572.4 billion from Gimhae and 334.4 billion from Busan, and they are contractually liable for at least 1.0898 trillion won more through 2041 (about $770 million). The larger number is still in the future. It belongs to a contract signed before the line carried a single passenger.
By the middle of the decade the annual payments had grown past what either council could absorb quietly, and the cities went back to the table. The original minimum revenue guarantee ran from September 2011 to March 2017, when it was replaced by a cost-compensation arrangement running through 2041. The distinction matters and is routinely lost in summary: a revenue guarantee tops up income against a forecast, while cost compensation covers the operator's costs against what it actually earns. The second is cheaper for the public side because it is no longer chained to a number that was wrong in 2011 and has stayed wrong.
It was cheaper, and it was not cheap. The renegotiation cut the projected remaining burden by roughly 304 billion won (about $215 million) — in the framing used in the regional coverage, from something on the order of 1.8 trillion won down to about 1.49 trillion, or roughly 1.3 billion dollars down to 1.05 billion. A third of a trillion won is a genuine saving, and it is also an admission: after six years the cities were negotiating not over whether to pay, but over how much of a thirty-year mistake they could buy back.
It would be easy, and wrong, to file this under failed technology. The automation works. The trains are two-car sets built by Hyundai Rotem and they run the timetable. Ridership has not collapsed; it has grown, slowly and steadily, in the way an ordinary suburban railway grows — from 30,084 a day at opening to 45,821 in 2024. On any measure that does not involve the contract, this is a modest, functioning piece of public transport serving a corridor that genuinely needed one.
That is precisely what makes it useful. A smart-city failure is usually legible: the sensors stop reporting, the app is switched off, the pilot never becomes a service. Here nothing switched off. The instrument that failed was a projection, and projections do not visibly break. They simply sit inside a contract, quietly converting an optimistic afternoon of modelling into a thirty-year transfer from two municipal budgets to a private consortium.
The pattern repeats across the same generation of Korean light rail. Uijeongbu's light rail, which opened in 2012, carried well under half its projected ridership and costs its city on the order of 20 billion won a year (about $14 million). And in one case the question of responsibility was actually put to a court: in Yongin, a former mayor's liability in damages over the city's light rail was confirmed by the Supreme Court. That is the rarest thing in this entire genre — a forecast with a named person attached to its consequences.
For Busan-Gimhae, the record contains no equivalent. The forecast was produced by a state institute, adopted into a contract by two elected councils, and is now paid for annually by both. Everyone in that chain acted within their remit. The bill arrives regardless, and will keep arriving until 2041.
Busan and Gimhae paid 906.8 billion won in support payments over 2011-2025 (Gimhae 572.4bn, Busan 334.4bn) and are contractually liable for at least 1.0898 trillion won more through 2041. Daily ridership in 2024 was 45,821, or 14.9% of the 306,000 forecast for that year.
A revenue guarantee written against a consultant's demand forecast hands thirty years of forecasting risk to the taxpayer. Renegotiating it in 2017 cut the remaining bill by about 304 billion won, but could not undo the fourteen years already committed.
Two clocks, kept apart: when this entry’s sources were published, and what has happened to its record in the atlas. Documented span: since 2011. 6 of 8 sources carry a publication date, from 5 April 2017 to 6 April 2026.
Eight sources carry this entry: the operator’s own site, a Busan city announcement, four Korean press accounts (Gyeongnam Domin Ilbo, Busan Ilbo, Kyunghyang Shinmun) and an SBS broadcast report, plus the English Wikipedia article, which carries three uncontested structural facts here (length, station count, rolling stock) and is also the exhibit for how the forecast figure gets mislabelled as capacity. No financial figure on this page rests on it. Dollar equivalents on this page are approximate and converted at a single reference rate — 1,409.94 won to the US dollar, the Federal Reserve H.10 rate for 7 August 2026, close to this entry’s last verification. They are a reader’s aid, not a source figure: no source quoted here states a dollar amount, and a sum accumulated between 2011 and 2025 spans years in which the won traded very differently. A caution about the numbers: Korean reporting counts in 억 (hundred million) and 조 (trillion), and machine translation of these articles reliably inflates 억 figures by a factor of a hundred. Every won figure on this page has been read back against its unit marker. Four things the record honestly cannot answer: who at the Korea Transport Institute produced the 176,358 forecast and on what assumptions; whether the on-board attendants reported at opening were later withdrawn; who holds the concession equity today; and what the Busan city announcement of the guarantee’s abolition says in full, because that page returned a server error on every attempt. 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 16 addressable claims, 8 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-c146b87f64a6c1ba
descriptive
fully automated, with no driver
clm-1641ddd8ec354d98
quantitative
The line is 23.9 km long and has 21 stations
clm-d51adc16fbc0fbf3
quantitative
It projected 176,358 riders a day in the opening year, rising to 322,545 by 2031
clm-5a486192c78904a8
quantitative
On the day it opened, the line carried 30,084 riders
clm-6971ce29b4ec1bf2
quantitative
Gimhae's support payment for 2020 alone was 48.6 billion won
clm-62c33287a7b57542
quantitative
against ridership of 34,752 a day where the forecast had promised 272,220
clm-2220b670135f5898
quantitative
By 2024 the annual payment across both cities had reached 84.1 billion won
clm-d391f2ab0dd7bc97
quantitative
Gimhae carries 63.19 percent of the cost and Busan 36.81 percent
clm-d0039e28a7bb5ba5
quantitative
Busan and Gimhae paid 906.8 billion won in support payments over 2011-2025
clm-7cb5af254cce4956
quantitative
they are contractually liable for at least 1.0898 trillion won more through 2041
clm-47888f53eb33ce56
quantitative
The original minimum revenue guarantee ran from September 2011 to March 2017, when it was replaced by a cost-compensation arrangement running through 2041
clm-bc0955a3ff20e6f2
quantitative
The renegotiation cut the projected remaining burden by roughly 304 billion won
clm-9babc150669e37d8
descriptive
The trains are two-car sets built by Hyundai Rotem
clm-15075be2ed3db22a
quantitative
from 30,084 a day at opening to 45,821 in 2024
clm-ff55c3ebc98c686c
quantitative
Uijeongbu's light rail, which opened in 2012, carried well under half its projected ridership and costs its city on the order of 20 billion won a year
clm-e65fa8b4f3f0e034
descriptive
in Yongin, a former mayor's liability in damages over the city's light rail was confirmed by the Supreme Court
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.
sha256:4cdf18e2ad9afa4b1c5f7f4de0a6debbc018f959d777e69a22cff280e6c21b11
src-c2af622a9a9f920cSentence-level source for the forecast and its author: the Korea Transport Institute projected 176,358 riders/day at 2011 opening, 211,147 by 2014 and 322,545 by 2031; actual opening-day ridership was 30,084 (20% of forecast); 2024 ridership 45,932/day; 2024 support payments 84.2bn won; April 2017 switch from MRG to 25-year cost compensation (MCC) running to 2041; public liability cut from 1.7963tn to 1.4919tn won.
src-ef495553f07849e9Authoritative city source for the contractual mechanism: press release of 5 April 2017 announcing abolition of the minimum revenue guarantee and its replacement by minimum cost compensation; revised agreement signed 24 March 2017, effective 1 April 2017; concessionaire, operating company and maintenance contractor merged into direct operation; private return cut from 14.56% to 3.34%; savings of about 304.4bn won to 2041.
src-aed164cc97a8ee6cMost recent source, April 2026: the two cities paid 906.8bn won in support payments over 2011-2025 (Gimhae 572.4bn, Busan 334.4bn) and must pay at least 1.0898tn won more from 2026 to 2041; confirms the subsidy is an active political issue in the 2026 Gimhae mayoral race, including a 17 March 2026 meeting with the transport minister.
src-1510a1616ac15422Independent confirmation of the payment split and the 2024 gap: cumulative support of 519.1bn won (Gimhae) and 302.8bn won (Busan) since 2011; 2024 payments of 52.9bn and 31.2bn won; forecast of 306,000 riders/day for 2024 against about 45,000 actual; Gimhae's request for national cost-sharing in the 2027 budget.
src-c84463cc3335f411Source for the 2024 and 2025 ridership figures used in the entry: 45,821 riders/day in 2024, equal to 14.9% of the 306,000 forecast for that year, and 45,689/day over January-May 2025.
src-4647d109f777759cSource for the opening date (revenue service from 17 September 2011) and for the COVID trough used as the recovery baseline: 34,752 riders/day in 2020, with 48.6bn won paid in support that year.
src-a0159a68aed0f6c5Official site of the BTO concessionaire that has run the line directly since the 2017 restructuring; confirms the line is in normal operation in 2026.
src-0a5ad2fb460befb7Used only for the technical specification: 23.9 km, 21 stations, standard gauge, 750 V DC third rail, Thales SelTrac CBTC moving-block signalling and fully automated driverless operation. Note that this page presents 176,000 riders/day as design capacity; the Korean sources above show it is in fact the opening-year demand forecast.
1001 Smart Cities (2026). “Busan-Gimhae Light Rail and Its Revenue Guarantee” — Busan, South Korea. The Smart City Atlas. https://1001smartcities.org/projects/busan-gimhae-light-rail/ (last verified 2026-07-26). Data: CC BY 4.0.
The underlying data is free to reuse with attribution — see the open data page.