Target Network 2032 · launch report
Your first 12 days of designing Europe's night trains
The tool opened to everyone on Tuesday 22 September 2026. Here is what happened until 3 October — what you built, what you told us, and what we are doing with it. The September product update has the second half of that story.
451
people joined
1 321
routes proposed
2.2 M km
of route drawn, one way
34
countries
01 · You showed up
451
people signed in to design a night train
179 opened an account, 272 tried it as a guest. On day one alone, 149 of you were building at the same time — and 85 came back on a second day or more.
02 · You built
1 321
night train routes, by 449 different people
55 % of you proposed one route and left it at that. The rest kept going — the most prolific of you published 108 routes.
And you iterate: 46 % of all proposals were re-published after a change, 3 122 times in total. One route alone was refined through 39 versions.
03 · Where
94 %
of your routes cross at least one border
Only 6 % of proposals stay inside one country; the typical route runs through 4. You did not design national night trains — you designed a European network. That is the whole point of Target Network 2032: the gaps are between countries, and no single ministry closes them on its own.
34 countries and 959 stations appear. Germany is in 54 % of all proposals, France in 46 %; the most-served country pair is Germany–France (145 routes), then Germany–Sweden (108) and France–Italy (81). Bruxelles-Midi is the station your trains pass through most (205 routes); where they start and end is a different list, led by Barcelona - Sants.
Countries shaded by how many proposals touch them; bubbles are stations, sized by use. The darker a country, the fewer routes it has so far — click a country or a station to open it in the gallery, and design a route from there.
Most used stations
Most popular termini
04 · How far
2.2 M km
of night train route, one way
That is 54 times around the equator. The typical route is 1 473 km long and takes 16 hours — a proper night train.
Not all of them are: Sevilla - Santa Justa – Söğütlüçeşme at 9 421 km and 5 nights, or Zagreb Glavni kolodvor – Zagreb Zapadni kolodvor at 2.1 km, were most likely people testing where the tool breaks (113 proposals have just two stops). Everything stays in the counts above — the next card is where the sieve starts.
05 · From a crowd of proposals to one network
924
of your proposals are typical night trains
Assuming a typical night train, one way, covers 500–2 000 km, takes 7–21 hours and averages at least 50 km/h — the yardstick of our general position paper as the gallery's typical night trains filter applies it — 70 % of your proposals qualify. Of the rest, 27 are shorter than 500 km, 295 longer than 2 000 km, 315 would take more than 21 hours one way, 38 fewer than 7, and 4 crawl below 50 km/h — overlaps included.
This is a first sieve, not a verdict on anyone's route. Between here and a study we can hand to the EU, every proposal goes through filtering — duplicates, tests, the envelope — and refinement with the people who know the line. The cards that follow use only the 924 typical night trains.
Many of them are the same train drawn twice: the 924 boil down to 531 distinct city pairs (termini within 50 km count as one city), and grouping routes that share most of their stops leaves about 319 corridors — each with one to 31 variants to refine.
Our position paper puts the number Europe needs at about 300 — so are we nearly done after two weeks? No. These 319 come out of a first sieve run on a first-generation model. The real demand model, the refined cost model and the routing fixes that are coming (see the September product update) will all tighten the mesh, and a fair share of today's corridors will not survive that reality check. So keep drawing — above all where the map above is still dark.
06 · The trains you agree on
12
of you drew Paris – Praha — independently
Duplicates are not noise, they are votes. 531 city pairs for 924 typical night trains means 43 % of them already existed when they were published — and almost always by someone else: only 32 are the same person re-proposing their own connection.
Paris – Praha was drawn 15 times by 12 different people, with anything from a non-stop run to 24 intermediate stops. That is exactly the material the refinement step needs: the connection is agreed, the stopping pattern is the discussion.
07 · What it would take to run them
427 M
train-kilometres a year
Added up, the 924 typical night trains come to 327 641 departures a year and 125.7 bn seat-kilometres offered. Read these as what the model says today, not as a forecast: every proposal is counted on its own, with a first-generation supply and demand model behind it, and the model will be reshaped considerably before any of this becomes a study. Supply side only, too — the demand assumption behind the revenue figures is still the simplest one that works, so we are not summing passengers yet.
Counted route by route they would need 1 047 trainsets — a theoretical figure, because it ignores every network effect. Most of your trains run three nights a week and sit idle the other four, when the same set could serve another corridor; if every set ran every night, 553 would do. The real fleet question can only be answered by looking at proposals together, as one network — the network analysis mode we are planning to build, precisely because this report cannot answer it yet.
What a train-kilometre costs is calibrated — as of today: the median proposal runs at 23.1 € per train-km, all costs included. We will keep pushing the cost algorithm towards real-world behaviour, component by component; the feedback button on every cost factor is how you help.
327 641
departures a year
1 047 → 553
trainsets, route by route → if every set ran nightly
125.7 bn
seat-km a year
Cost per train-km, €
08 · What you'd run
86 %
of you stayed with the default train
The next-generation Nightjet-style formation (NEW-BAL-7, 260 places in 7 coaches) is what almost everyone kept. It is a convention, not a recommendation: the catalogue holds 11 other concepts — pod trains, all-private-room formations, budget seat-and-couchette sets, double-length consists — and depending on the route, another concept may be the more efficient one, or the one whose supply profile fits the passengers it is for. Pick one on your route and compare; the tool shows the difference side by side.
8 % of you want your train to run every night; the default is three nights a week, and 89 % left it there.
09 · You talked
115
likes and 28 comments — a start
38 of you liked at least one other proposal and 17 wrote a comment. Honest reading: the building worked, the talking-to-each-other has not really begun — 1 321 routes and 28 comments means most proposals were never discussed. Making it easier to react to each other's routes is part of what comes next. The comments that were written are worth reading, though, without names:
“Would be a dream come true!”
on Kraków – Rotterdam
“500 thousand Czechs visit Austria every year, often for hiking or skiing in the Alps … This train would allow them to leave Prague in the evening and arrive comfortably in the mountains.”
on Praha – St. Anton am Arlberg
“Needed for members and staff of the European Parliament.”
on Strasbourg – Bruxelles-Midi
“I wouldn't have expected this route to not require subsidies!”
on Rotterdam – Roma
“The journey time is a crucial aspect to attract passengers — night trains should mean the whole day is spent at the destination.”
on Bruxelles – Milano
“Such an obvious connection. Berlin is a big vacation destination for many Scandinavians.”
on Berlin – Malmö
10 · You told us what's missing
37
pieces of feedback
By far the most frequent message: "I searched the stop list and found nothing." 32 stations and 3 lines you asked for, from Arlon to Vlissingen. 3 routing detours we need to fix, 1 documentation correction, and 1 feature request: editing the stop order of a trip.
What we did with it, and what comes next, has its own page: the September product update.
Stations you asked for
ArlonBerchtesgadenBratislava-Nové MestoBrixen/BressanoneBüchenČadcaCeljeDen Haag HSFishguard HarbourFolkestoneFredericiaHamburg-BergedorfHarwich InternationalIisalmiKajaaniKouvolaKristiansandLisboa Santa ApolóniaMantes-la-JolieMassy-PalaiseauMeran/MeranoMikkeliMonaco-Monte-CarloNeustadt (Weinstraße)Ostseebad BinzPasilaPieksämäkiPovažská BystricaSchleswigTinglevVersailles-ChantiersVlissingenNova Linha de ÉvoraLinha da Beira BaixaLinha da Beira Alta
Keep designing
The gallery is open, your proposals are there, and the next update will count what you do next. What we did with your feedback is in the September product update.
Figures from the production database, 22 September – 3 October 2026 (exported 4 October 2026). No personal data: counts, public proposal names and comment texts only. Under the hood: 7 844 route calculations, a median of 3.7 s each, 1 server error in 12 days; 30 % of all requests came from a phone.
