Türkiye has begun buying construction capacity for one of the largest railway projects in its history. On 21 July 2026, the Ministry of Transport and Infrastructure, through the Directorate General of Infrastructure Investments (AYGM), invited international proposals for the design and construction of infrastructure and superstructure works — including electrification — for the İstanbul North Rail Crossing Project, known as INRAIL.
The tender is the point at which a long-discussed alignment stops being a map and starts becoming a procurement problem. And the procurement problem is unusually interesting, because INRAIL is not a single railway: it is a 127 km corridor that stitches together two airports, a motorway bridge, a Marmaray junction and a European main line — and the riskiest element in the whole chain was built for something else.
What is actually being built
INRAIL is a greenfield, double-track, electrified, fully signalled mixed-traffic railway running across the northern edge of the İstanbul metropolitan area, from Çayırova on the Asian side (Kocaeli province) to Çatalca on the European side. It is designed to let freight and intercity passenger traffic bypass the congested Bosphorus crossings and the urban core entirely.
The alignment, as set out in the advance procurement notice, is split into four geographic corridors:
| Lot | Corridor | Approx. length in the RFP |
|---|---|---|
| Lot 1 | Çayırova – Sabiha Gökçen – Kurtdoğmuş | 28.49 km |
| Lot 2 | Kurtdoğmuş – Alibahadır | 25.76 km |
| Lot 3 | Alibahadır – YSS Bridge – Mithatpaşa | 32.14 km |
| Lot 4 | Mithatpaşa – İstanbul Airport – Çatalca | 39.49 km |
The four lot lengths sum to 125.88 km. That is not the 127 km figure quoted in most official material, and the gap has caused confusion. The reconciliation sits in the tender documents: the scope is 122.3 km of double-track main line plus 4.7 km of single-track connection lines, totalling 127.0 km. In other words, the headline length includes spur and connection mileage that the lot breakdown does not fully capture. The numbers are consistent once you know which number means what — a small thing, but the kind of thing that matters to anyone pricing a bid.
What sits under those kilometres is substantial civil engineering. Across the four lots the RFP anticipates:
- 5 twin-tube TBM tunnels totalling 36.2 km
- 22 double-track NATM tunnels totalling 23.3 km
- 17 cut-and-cover tunnels totalling 3.7 km
- 2 viaducts and 40 bridges totalling 22.1 km
That is roughly 63 km of tunnel and 22 km of bridge on a 127 km railway — close to half the route is underground, and most of the rest is elevated. The Ministry's February 2026 statement described "44 tunnels of 59.1 km and 42 bridges of 22.4 km"; the difference from the RFP figures is small and most likely reflects different counting conventions at different design stages rather than a change of scope.
The tunnels are not just a cost line. The World Bank, in approving its financing, framed them as climate adaptation: putting half the route underground reduces exposure to extreme heat, flooding, high winds and wildfire — a real consideration for a corridor crossing the Thracian and Kocaeli hinterlands.
The bridge is ready. The railway is not.
The Bosphorus crossing is planned over the Yavuz Sultan Selim Bridge (YSS), the 2016 cable-stayed road bridge better known as the Third Bosphorus Bridge. The bridge was built by an İçtaş–Astaldi joint venture to a design competition won by Michel Virlogeux and Jean-François Klein, with final design by T-Ingénierie and Bureau Greisch. Crucially, it was built to carry four road lanes in each direction plus two railway tracks — deck space for rail has existed since opening, but no railway has ever been laid on it.
This is the distinction that Turkish procurement documents are careful to preserve: the bridge is railway-ready, not railway-equipped. The future contractor's scope includes installing track, connections, electrification, signalling and associated works on the deck. That is a very different construction problem from a greenfield alignment in open country — it involves working on a live motorway bridge, under cable-stayed load paths, over one of the busiest and most environmentally sensitive waterways in the region.
It also explains why INRAIL's financing has been structured as a sovereign operation. The bridge is operated under a concession; the railway on it is not the concessionaire's project. Layering a 127 km international railway across a motorway PPP is exactly the kind of interface that sends lenders to governments rather than corporates for repayment.
Who is paying
The money has been assembled with unusual coordination. Six multilateral institutions are providing roughly US$6.75 billion alongside Turkish government co-financing, under a Full Mutual Reliance Framework with the World Bank as lead lender — a structure designed to let the borrower run one procurement and one set of lender safeguards instead of six parallel processes.
| Institution | Indicative amount |
|---|---|
| World Bank (IBRD) | US$2.0 bn (approved 31 March 2026) |
| AIIB | Up to US$1.5 bn |
| ADB | US$750 m – US$1.5 bn (see note) |
| Islamic Development Bank | US$750 m |
| EBRD | Up to €500 m |
| OPEC Fund | Balance of package (~US$500 m) |
A note on the ADB figure: the AIIB financing table attributes US$1.5 bn to ADB, while ADB's own project page describes a US$750 million equivalent loan. Both are reported here rather than reconciled, because the sources genuinely differ.
EBRD's exposure is a sovereign loan of up to €500 million, classified Environmental Category A — the highest environmental and social risk category — with disclosure in March 2026 and the project described as having passed structure review and pending final review. Category A is appropriate: the alignment crosses forested and agricultural land on both sides of the Bosphorus, involves extensive tunnelling, and touches the protected status of the Bosphorus strait itself.
Total project cost is put at roughly US$8.3 billion by the World Bank and about €7.799 billion by EBRD — a difference explained by currency, timing and FX assumptions rather than by different scope.
What the tender mechanics tell you
Contract values by lot have not been disclosed. What has been disclosed is the proposal security required per lot, and that is a usable proxy for the employer's own view of relative risk and value:
| Lot | Proposal security |
|---|---|
| Lot 1 (Çayırova–Kurtdoğmuş) | US$5,000,000 |
| Lot 2 (Kurtdoğmuş–Alibahadır) | US$2,800,000 |
| Lot 3 (Alibahadır–YSS–Mithatpaşa) | US$2,800,000 |
| Lot 4 (Mithatpaşa–Airport–Çatalca) | US$5,000,000 |
Lots 1 and 4 carry the heaviest securities — the two lots that anchor the line's urban interfaces, the Sabiha Gökçen connection at one end and the İstanbul Airport–Çatalca approach at the other. Lot 3, which contains the bridge crossing itself, carries a comparatively modest security, which is worth noting: it is the shortest lot by civil scope in some senses, even if it carries the highest interface risk.
The bid deadline has already moved. It was reported as 14 October 2026 and then postponed to 11 November 2026 at 14:30, with technical proposals due to open at 14:45. Proposers may bid for one or more of the four lots. AYGM records an Addendum No. 1 dated 9 October 2026 — consistent with the deadline extension and the kind of clarification cycle a project of this complexity generates.
The works tender is not the whole procurement. Telecommunications and signalling are being tendered separately, which is a deliberate unbundling: it gives the employer a chance to buy a systems package on its own terms rather than let it be swallowed by civil contractors, and it keeps a single integration point for what will eventually be a very long, very fast railway. As of the evidence available, that procurement has not been launched and its timing is unclear.
The commercial picture: who is circling, and who is committed
No EPC or design-build contractor has been identified or awarded. That is the single most important fact about INRAIL's current status: it is a tender, not a project under construction.
Turkish trade press, citing ANKA, reports market interest from Chinese firms and from Turkish groups including Kalyon, Çelikler, Metgün, Özaltın, Türkerler, İçtaş, Cengiz and Makyol, with Turkish–foreign consortia considered likely. This is unconfirmed market reporting, not a bidder list, and should be treated accordingly. It is nonetheless informative: the names are the ones that consistently appear on Türkiye's largest transport infrastructure works, and several of them — İçtaş in particular — have prior history with the YSS Bridge itself.
AYGM's railways construction department, headed by Serkan Uçar, is the employer-side contact for the works tender.
Why this matters beyond Türkiye
Türkiye's rail freight problem is not a lack of track. It is a Bosphorus problem. The Marmaray tunnel carries only limited freight, and the existing crossing capacity puts a hard ceiling on how much rail cargo can move between Asia and Europe. The World Bank frames INRAIL's effect starkly: Bosphorus rail freight capacity rising from roughly 3 million tonnes a year to as much as 50 million tonnes — more than a fifteenfold increase.
The Ministry's own February 2026 forecast was more conservative, at around 33 million passengers and 30 million tonnes of freight per year. The gap between 50 Mt (capacity ceiling) and 30 Mt (traffic forecast) is normal — one is what the railway can carry, the other is what is expected to use it — but it matters for the business case, because a corridor justified on 50 Mt and delivered at 30 Mt has a different return profile.
The strategic logic is that INRAIL completes a continuous Middle Corridor chain: the Baku–Tbilisi–Kars line into eastern Türkiye, across to Kocaeli, over the YSS Bridge, west toward the Halkalı–Kapıkule line and on into the EU. Freight moving from Central Asia and China to European markets currently faces a bottleneck exactly where the geography is hardest. INRAIL is an attempt to widen it.
For the steel and industrial supply chain, the implications are concrete even before an award. A project of this shape — 63 km of tunnelling, 22 km of bridging, two airport connections and 127 km of electrified double track — represents a very large structural steel, reinforcement and rail procurement programme, plus catenary structures, signalling gantries and bridge steelwork. TBM tunnelling at 36 km in five twin-tube drives is also a segment-lining demand story. None of those specifications have been published. Anyone modelling the materials impact is currently modelling a range, not a bill of quantities.
What to watch
The award. With bids due 11 November 2026, awards across four lots would be the moment INRAIL becomes a construction programme rather than a procurement plan. Watch particularly how the lots are allocated — a single contractor taking multiple lots would change the supply chain geometry and the degree of coordination risk the employer is accepting.
The bridge interface. Lot 3 is where the project's technical uncertainty concentrates. How the track, electrification and signalling works are sequenced around a live motorway bridge, and who carries the interface risk with the bridge operator, is the detail that will determine whether the Bosphorus crossing is the project's showcase or its bottleneck.
The signalling tender. Bundling telecommunications and signalling into a separate procurement is defensible, but it creates a second award timeline that must land in the right order relative to civil works. If the systems tender slips, the civil contractors will finish alignment and find nothing to install on it.
Ministerial forecasts. Former statements suggested the line could be operational by 2028 if the tender proceeded promptly. That was a target, not a committed date, and with bids now landing in late 2026, it should be read as ambition rather than schedule. A 127 km greenfield railway with 63 km of tunnel and a bridge retrofit does not compress easily.
INRAIL's most striking feature is not its size — Türkiye builds large infrastructure constantly. It is the degree to which it depends on assets and institutions that already exist: a road bridge built with spare track beds a decade ago, a set of multilateral lenders willing to run a single joint procurement, and a national rail strategy that has been waiting for someone to solve the Bosphorus. The tender is where those threads either converge or come apart.
