The cheapest 14 kilometres of Alpine rail capacity in Europe is a hole in the ground that has carried nothing since the 2000s. It sits inside my mountain — the second bore of the Lötschberg Base Tunnel between Ferden and Mitholz, excavated, dry, and empty. No track, no traction power, no ventilation. A raw shell (Rohbau) that Switzerland paid to dig and then deliberately declined to fit out.

On 19 August 2026 BLS — the operator here, not SBB — put the amended plans for the full expansion on public display. The öffentliche Auflage under the Federal Office of Transport’s planning-approval procedure runs 17 August to 15 September, as bahnonline reported. When you read this, the objection window has about a week left. Say what you mean now, or hold your peace for a decade.

Here is the lineage the ribbon-cutting of 9 December 2007 skipped. The 34.6 km bore from Frutigen to Raron was designed twin-bore, double-track. Under NEAT cost pressure — and the political need to protect the Gotthard budget — only the southern 14 km got two running tracks. The other 21 km opened single-track: trains in opposite directions sharing one line under the Bernese Alps, mutually exclusive by geometry. Deferral was booked as a financing decision. It became a scheduling decision, and the schedule is the one you cannot renegotiate.

Within a few years the single-track section ran at roughly 80% of capacity, 100% on peak days, on one of two principal north–south corridors through the Alps. That is what a deferred fit-out actually buys: the concrete holds, the calendar does not. The 14 km shell waiting quietly for track is genuine option value on a balance sheet — the right bet. The 7 km between Mitholz and Frutigen is the option they did not buy. It was never bored at all, and buying it late now costs CHF 640 million extra, inside a total BLS forecast of CHF 1.7 billion.

The bill for late sequencing is not only money. That 7 km runs toward Frutigen through Mitholz — the mountainside that held a Swiss Army ammunition depot until 19 December 1947, when part of 7,000 tonnes detonated, killed nine people, and destroyed around 40 buildings. Clearing the debris cone — some 350,000 m³ of heavily polluted material — is costed at CHF 2.5 billion; SWI swissinfo reported on 29 May 2026 that contamination has pushed completion to 2047, with project manager Adrian Götschi warning of “enormous amounts of pollutants.” Roughly 150 residents evacuate from end-2032 for ten years. My construction decade (from 2028, ~10 years) lands squarely on top of theirs: two federal megaprojects, CHF 4.2 billion combined, one narrow valley, one decade.

←TODAY: A bored-but-unequipped tube is Europe’s cheapest Alpine capacity — and its most expensive one to finish late.
→3012: The network that survives is the one that bought its second path before the mountain got busy.
Fulcrum: Deferral prices the concrete correctly and the calendar not at all.

Every architect has run this experiment at a smaller scale. Core-and-shell delivery, oversized risers, a spare structural bay, a slab poured for a lift not yet ordered. PAZ has covered deferral-as-policy before — our piece on what CHF 1.7 million cuts from the Swiss e-ID made the same point in software: the line item you postpone quietly decides what the thing becomes. The Lötschberg says the shell bet is usually right about the material and usually wrong about the moment you get to install into it.

Atelier: For an office pricing a core-and-shell option this week, the empty bore’s real cost was never the concrete — it was booking the closure window ten years early, into a valley already full. Monday move: add a “when, and who blocks the site then” column to every deferred-scope line in your next tender, so a financing choice can never sneak through as a scheduling one.

Hack: Model the bottleneck before you assert it. On a single-track bore, throughput falls out of geometry, not demand — block occupancy is transit time plus clearing plus a headway margin, and opposing moves cannot overlap. Run the 21 km section and watch paths/hour roughly halve against a double track:

transit = 21 / 100 * 60           # 21 km bore at 100 km/h, minutes
block   = transit + 3 + 2         # + clearing + headway margin
single  = 60 / block              # opposing moves share ONE line
print(round(single, 1), "paths/h single vs", round(single*2, 1), "double")

Change the line speed or the margins and the 80%-of-capacity figure moves with the arithmetic, not with the timetable.

If you mean to lodge an objection, you have until 15 September. If you are building within any long-lived asset this year, ask for the second path on paper now — the diversionary route, the fitted bore, the spare bay — and ask who installs it in year sixty, when the valley is busy. Build the redundancy before the mountain is.