When the Auckland City Rail Link finally opens to passengers in the second half of 2026, the country will gain three and a half kilometres of twin-bore tunnel, two new underground stations and the single most expensive piece of transport infrastructure ever built in New Zealand. The final bill, last revised in 2023, is NZ$5.493 billion. Or, put another way, NZ$1.57 billion for every kilometre of new line.
That figure has been the subject of polite agreement for years. It became a national argument earlier this month when the project’s own former chief executive, Dr Sean Sweeney, told NZ Herald that the same line could have been delivered for “half the cost”, described the finished product as a “Rolls-Royce” where a Corolla would have done, and walked through, point by point, where he believed the money had gone wrong. Infrastructure Minister Chris Bishop responded within 48 hours by ordering a formal review.
It is the most candid public autopsy of a New Zealand mega-project ever delivered by the person who ran it. It is also, deliberately or otherwise, a diagnosis of a much bigger problem than one rail tunnel.
A century in the making
An underground rail link through central Auckland is not a new idea. The first formal proposal was made in 1918 by the New Lynn Town Board, which floated the idea of a tunnel from Beach Road through the central city to Morningside. By 1923 the Minister of Railways, Gordon Coates, was publicly backing the scheme. By 1936 his successor Dan Sullivan had abandoned it, citing an estimated cost of £1 million, “engineering difficulties” and doubts about whether suburban rail had a future as Auckland sprawled outwards.
The Halcrow report of 1950 revived the concept. So did studies in 1965, 1972, 1986, 2003 and 2008. Each time the answer came back the same. The line would be useful but it would be expensive, and Auckland was not yet quite ready to commit. Serious modern planning began in 2008. A funding deal was not signed until 2016, by the Key government and Auckland Council on a 50/50 split.
Construction began the following year. Tunnelling finished in 2022. Fit-out, signalling and integration testing have absorbed the four years since.
What the project actually delivers
The CRL is more than a new line. By looping the existing rail network through the central city it converts Britomart, now renamed Waitematā, from a dead-end terminus into a through-station. That single structural change roughly doubles the practical capacity of Auckland’s entire rail system. The two new underground stations, Te Waihorotiu in the central business district and Karanga-a-Hape near K Road, are designed for nine-car trains and reach 42 metres below street level at the deepest point.
Auckland Council’s modelling suggests the central business district will become reachable by rapid transit for roughly twice as many people as today. Around 21,000 residents, 17,200 students and 37,000 jobs will fall within a twelve-minute walk of the two new stations. The journey from Henderson to midtown is forecast to shorten by 24 minutes. Patronage on the rail network is forecast to climb to 42 million trips a year by 2030.
None of that is in dispute. The project’s defenders, and they are numerous, are right that it is a serious piece of infrastructure that will quietly reshape how Auckland moves for decades. The argument is about what it ought to have cost.
The bill, told in three instalments
The price of the CRL has not jumped once. It has roughly doubled, in three distinct steps.
The first formal estimate, in 2014, was around NZ$3.4 billion. The 2016 funding deal was signed on the basis of a project costing somewhere between NZ$2.5 billion and NZ$3.4 billion.
In 2019, after the Link Alliance contract was awarded for main works, the price was revised to NZ$4.419 billion. That was attributed at the time to scope refinement, tender pricing and the late decision to lengthen platforms from six-car to nine-car capacity to future-proof the line.
In 2023 the price rose again, by a further NZ$1.074 billion, to the current NZ$5.493 billion. CRL Ltd and its sponsors attributed that increase to Covid-related construction delays, supply-chain inflation, labour shortages and the lockdown disruption that hit the project in 2020 and 2021. Most external commentators accept that some portion of the third blowout was genuinely unavoidable.
The argument now is about the first two billion, and the structural conditions that produced them.
How that compares to the rest of the world
Before he resigned as chief executive in 2024, Sweeney used a slide pack on the international cost of urban rail in several public talks. The data, drawn from the Transit Costs Project at New York University’s Marron Institute, painted a picture that became more uncomfortable each time he showed it.
The New Zealand cost per kilometre, on the basis used by the Transit Costs Project, sat at roughly US$922 million. The equivalent United States figure, calculated across the country’s recent urban rail projects, was around US$602 million. Australia came in at US$321 million. France at US$256 million. Portugal, South Korea, Spain and Finland built urban rail at around US$100 million per kilometre, less than a ninth of the New Zealand figure.
The most uncomfortable point in the slide pack was this. The lead tunnelling contractor on the CRL is Soletanche Bachy, a French firm that is one of the leading metro tunnelling specialists in Europe. The same company is currently working on cheaper projects in Paris, Stockholm and Madrid. The contractor is not the variable.
Where Sweeney says the bloat is
Sweeney’s specific critique, given to NZ Herald and 1News in May, has four strands.
The first is station scale. The CRL’s two new underground stations are around 200 metres long and have between six and eight escalators each, with deep boxes excavated to accommodate nine-car trains. Copenhagen’s driverless metro, which Sweeney has used as a comparator for years, has stations around 60 metres long with two escalators each, one up, one down. The Copenhagen stations were built, on his numbers, for around a quarter of what Auckland has spent.
The second is the choice to future-proof for nine-car trains rather than build a higher-frequency line with shorter trains. Sweeney told 1News this decision alone “probably doubles the cost” of the underground civil works. A high-frequency, short-train metro of the Copenhagen model would move similar peak passenger volumes with a fraction of the structural overhead. He recounts asking a Copenhagen designer where the station toilets were. “He said, ‘There are none.’ I said, ‘Why not?’ He said, ‘Sean, we run trains every minute and a half. Why should we provide a toilet when you’re going to be here less than 80 seconds?’”
The third is what he describes as a New Zealand habit of over-specifying every job. “I do believe we’re not very good at specifying the scope,” he told NZ Herald, “and I do believe in New Zealand we’re over-specifying jobs, and it’s making them too expensive, which means we do less of them.” His comparison: “In China, every railway station is exactly the same. You’re catching a train, you’re not getting married.”
The fourth is procurement. In a 2023 interview, before he left, Sweeney flagged that the prevailing New Zealand model loads risk onto the contractor even when the contractor cannot reasonably price it. Below-ground geology risk, change-of-legislation risk, programme risk on the wider network. The contractor responds by pricing defensively, or by walking away. Fletcher Building, then the country’s largest construction company, had by that point stopped tendering for transport projects. They had told him, he said, that the projects were “too risky”.
The Fletcher footnote
That last detail carries a particular sting in 2026. Earlier this year, Fletcher Building announced the sale of Fletcher Construction, the country’s flagship civil-engineering firm, to VINCI Construction for NZ$315.6 million. The completion is expected before the end of 2026. The same VINCI is the French parent of VINCI Grands Projets, the lead partner in the Link Alliance that built the CRL.
The country’s biggest construction company, in other words, has effectively been absorbed by the French contractor on the project that was supposed to demonstrate why New Zealand could no longer build at New Zealand prices. Whatever else the CRL ends up signifying, that transaction is part of its legacy.
The counter-arguments
Sweeney’s critique has not gone unchallenged. CRL Ltd’s current chief executive, Patrick Brockie, told NZ Herald that the cost reflected “a wide range of factors beyond architecture alone”. Auckland councillor Richard Hills, who has sat on the CRL governance committee for years, said he was “totally furious that someone can be part of this project for six or seven years and then jump out” before publicly criticising the decisions made on his watch. Sweeney’s response was that, at the time many of the choices were made, he “probably didn’t have the experience in building metro stations” to push back.
The Greater Auckland transport blog has made the most substantive technical objection. It points out that running the central tunnel with three-car trains rather than nine-car trains would require frequencies of around 32 trains per hour through the central section, against the six to eight per hour the network currently runs through Britomart. Reaching that frequency would require hundreds of millions of dollars in signalling upgrades, hundreds of millions more in grade-separated junctions, billions in level-crossing removals across south and west Auckland, and a significant programme of freight-line extensions to get KiwiRail’s freight trains out of the way. A cheaper tunnel, in this account, would have shoved much of the saving back onto the surface network.
Both critiques can be true at once. The CRL was probably over-specified for its narrowest task. It would also have been very hard to specify it more lightly without rebuilding much of the rest of the Auckland network around it.
The Link Alliance question
The CRL’s main works were delivered through what the industry calls an alliance contract. The Link Alliance brings together VINCI Grands Projets, Downer, Soletanche Bachy, WSP, AECOM and Tonkin and Taylor as a joint venture sharing risk and reward with CRL Ltd as the client.
The model has clear advantages on a project this complex. It also has well-known weaknesses. Alliances are not strongly incentivised to deliver under budget, because the client carries part of the cost risk. They tend to operate as a black box from the outside, with limited transparency into how forecasts are built. And when things start to go wrong, the conversation between client and contractor can become long and unproductive.
By late 2024 those conversations had reached crisis point. CRL Ltd wrote formally to its sponsors, the Government and Auckland Council, expressing a “lack of confidence in the Link Alliance’s reporting and forecasting”. Senior VINCI executives flew from France to Auckland in December that year for direct meetings. The crisis did not produce another headline cost increase but it produced a steady backdrop of unease through the final year of the build.
Why it took as long as it did
The tunnels, it turns out, were the easy bit. They were finished in 2022. The four years since have been spent on the harder problem, which is making a brand-new metro work as a seamless extension of an ageing suburban network it was not originally designed to mesh with.
That has meant 16 kilometres of new track, 816 kilometres of signalling cables, more than 5,000 square metres of metal cladding in the stations, and an extended programme of pre-passenger testing during which trains run on real timetables, in real conditions, without carrying anyone. The same integration phase delayed London’s Crossrail by three years and was, in Sweeney’s own assessment, always the riskiest part of the CRL programme.
In parallel, KiwiRail has been running its Rail Network Rebuild, replacing drainage and track foundations across the wider Auckland network. The CRL cannot open in earnest until that rebuild is finished. The two programmes have been running on overlapping deadlines for years.
The bill that arrives next
The capital cost is only half of the financial picture. Once the line opens, Auckland Council will be on the hook for around NZ$220 million a year, net of fares and government contributions, to operate and maintain it. Auckland Transport’s own share of that is roughly NZ$26 million a year for additional services. The gross operating cost is closer to NZ$265 million, with about NZ$44 million coming back via Waka Kotahi NZ Transport Agency.
That bill is the principal reason the average Auckland household saw a 7.9 per cent rates increase this year, with annual rates rising from NZ$4,055 to NZ$4,375. Average business properties saw a 9.84 per cent increase. The CRL was promised to ratepayers, fairly, as transformative infrastructure. It was not always made clear that the transformation would also arrive as a line item on the rates bill.
What this means for the next project
Bishop’s review of the CRL is the second formal cost review of a major New Zealand transport project in eighteen months. The first, into the cancelled Auckland Light Rail, recorded that NZ$228 million had been spent over six years without producing a single metre of track. The CRL produced tunnels. It also produced, depending on which side of the argument is being made, between NZ$1 billion and NZ$2 billion of avoidable expense.
The Northwest Rapid Transit corridor, a revived Auckland to Hamilton commuter rail proposal, the long-promised Wellington mass transit project and an eventual second Waitematā harbour crossing are all queued behind the CRL. If they are scoped and contracted the same way, they will be priced the same way.
The Transit Costs Project’s standing recommendations to expensive countries are, in their own words, unfashionable. Keep the design work inside the public agency. Standardise station designs across multiple lines. Run a steady project pipeline so contractors can plan and price competitively. Resist the urge to make every station an architectural statement. Stop dumping unmanageable risk onto contractors. Build the next line shorter and faster, then the one after that, rather than agonising for a decade over a single grand project.
None of that is glamorous. It is, however, what the countries getting four times as much rail for the same money are doing.
An honest read
The “half the cost” line is partly true and partly retrospective. A genuinely Copenhagen-style line, with three-car trains, driverless operation and stripped-down stations, probably could have been delivered for NZ$2.5 to NZ$3 billion. That was never the line the country actually agreed to build in 2016, and trying to retro-fit Copenhagen onto Auckland’s existing suburban network would have created its own large second bill on the surface.
The more uncomfortable truth is the structural one. New Zealand pays roughly three times what Australia pays, and nine times what Spain pays, to build a kilometre of urban rail, regardless of scope, regardless of contractor, regardless of which government is in office. Sweeney’s most useful contribution is not the headline. It is the diagnosis underneath the headline. Until the pipeline, the scoping discipline and the risk-allocation problems are addressed, the next project will produce the next review.
New Zealanders will get to vote on a great deal of transport policy at the 2026 general election, from rapid transit revival to road tolling to KiwiRail’s future. Our voting tool compares the parties’ transport platforms side by side.
What do you think? Was the CRL worth the money, or was Sweeney right that a leaner project would have done? Tell us in the comments.
Featured image: City Rail Link construction on Albert Street, Auckland, 2018. Photo via Wikimedia Commons, CC BY-SA 4.0.
This article was written by AI, briefed to report the facts, hopefully without some of the bias people bring to the job 🙂
Leave a Reply