The Cook Strait Tunnel: Has Elon Musk made it feasible?
- Grant McLachlan
- 19 hours ago
- 9 min read
An analysis of technological evolution and strategic infrastructure planning
Every generation since the 1960s has studied a Cook Strait tunnel and concluded it was “too expensive.” Each generation was correct—based on the technology available at the time. But here's the critical question we're not asking: what if the technology that makes this project viable doesn't exist yet, but will exist by the time we need it?
In November 2025, New Zealand committed $1.86 billion to new Interislander ferries, scheduled for delivery in 2029. These vessels have a 30-year design lifespan, meaning they'll reach end-of-life around 2059. By then, we'll face a choice: spend another $3–5 billion on yet another ferry fleet, or commission a fixed link that eliminates weather risk permanently.
Major tunnel projects require 15–25 years from commitment to operation. If we want a tunnel operational by the late 2050s, serious investigation must commence by 2030 at the latest. This isn't a proposal to build a tunnel today. It's a proposal to plan intelligently for technological change.
The question shouldn't be “How much will it cost?” The question should be “At what cost does this become feasible?”
The technology revolution that changes everything
Two parallel technological revolutions are converging to make a Cook Strait tunnel increasingly viable: tunnelling technology and vehicle autonomy.
The Boring Company: Tunnelling cost disruption
Elon Musk's Boring Company has demonstrated construction costs of approximately NZ$47 million per mile ($29 million per km) with its Prufrock-4 machine, targeting further reductions to $17 million per mile ($11 million per km). Compare this to traditional tunnelling costs of $150–280 million per kilometre, and the implications are staggering.
The company has completed multiple projects demonstrating remarkable speed: the Encore-LVCC Connector in Las Vegas was completed in under 10 weeks from site access to tunnel completion. Clark County has approved 68 miles of Vegas Loop tunnels, and Nashville is proceeding with a 10-mile Music City Loop connecting the airport to downtown.
While current Boring Company tunnels are 12–14 feet diameter (designed for passenger vehicles), the company is actively exploring larger diameters up to 21 feet for freight applications. Elon Musk has publicly solicited expressions of interest for 50km+ projects, and an Australian politician has proposed a 50km Blue Mountains tunnel. The technology is scaling up.
Tesla and vehicle autonomy: Making road tunnels safer
Tesla and other manufacturers have introduced technologies that make it safer for vehicles to maintain consistent speed, safe following distance, and lane keeping. Adaptive cruise control, lane-keeping assist, and collision avoidance systems are rapidly becoming standard across the vehicle fleet.
By the time a Cook Strait tunnel could open (2050s), the vast majority of vehicles will have these features standard. This transforms the safety equation for long road tunnels. What once required rail shuttles (like the Channel Tunnel) could be achieved with managed road traffic in convoy, similar to Norway's long road tunnels but with enhanced safety through vehicle technology.
Modern tunnel safety systems
A Cook Strait tunnel would employ proven safety systems from international precedents:
Automatic licence plate recognition for toll collection and traffic monitoring (vehicles tracked entering and exiting)
Pilot vehicle systems managing traffic flow in controlled convoys, similar to Norway's Lærdal Tunnel (24.5km)
Deluge fire suppression systems with water mains and foam systems throughout
Emergency refuge chambers every 500 metres with direct communication to control centres
Advanced ventilation dedicated systems for vehicle exhaust management
Comprehensive CCTV and sensors monitoring every section continuously
These systems are proven in tunnels worldwide. The technology exists; the question is whether we'll plan to deploy it.
The questions holding us back: Myth vs Reality
Every transformational infrastructure project faces psychological barriers. Let's address them directly.
Are there fault lines?
New Zealand's longest rail tunnels—Ōtira (8.5km), Rimutaka (8.8km), and Kaimai (8.9km)—were built across continental faults and have withstood major earthquakes including the 2016 Kaikōura earthquake. The Seikan Tunnel in Japan crosses a major fault line and has withstood multiple megathrust earthquakes exceeding magnitude 9.0. Modern seismic design can handle fault zones. The proposed Cook Strait tunnel route avoids major fault lines and would be engineered to world-class seismic standards—precisely New Zealand's area of expertise.
How deep will it be?
The tunnel would reach approximately 300 metres below sea level at its deepest point, requiring only a 1% gradient—barely noticeable to drivers and actually advantageous for trains (gravity assists acceleration and deceleration). Norway's Ryfylke Tunnel, opened in December 2019, extends 14.4km and reaches 292 metres below sea level. The depth is well within proven engineering capability and provides adequate rock cover for structural integrity.
How long will it be?
Maximum length would be 67.4 kilometres between current ferry terminals (Wellington and Picton). The shortest viable route could be approximately 60 kilometres between Wellington Airport and Rarangi. For comparison, the Channel Tunnel is 50km (38km subsea), and the Gotthard Base Tunnel is 57km. A Cook Strait tunnel would be similar in length to proven international precedents.
What sort of rock would need to be removed?
The proposed route passes primarily through Torlesse Greywacke, an indurated sandstone/mudstone with Unconfined Compressive Strength (UCS) of 80–150 MPa. This is ideal for modern tunnel boring machines (rated to 200+ MPa) and very similar to rock successfully tunnelled in New Zealand's existing rail tunnels, Japan's undersea tunnels, and Norway's long road tunnels. The rock has very low permeability (<10⁻¹⁶ m²), minimising water ingress risks. It's actually excellent tunnelling geology.
Wouldn't a tunnel be dangerous?
Modern tunnels are statistically safer than ferry crossings. The Channel Tunnel has operated since 1994 with zero passenger fatalities despite carrying millions of vehicles annually. Norway operates multiple long road tunnels (Lærdal 24.5km, Ryfylke 14.4km) with exemplary safety records. Modern ventilation, fire suppression, emergency systems, and vehicle technologies make long tunnels remarkably safe. Compare this to Cook Strait ferries operating at only 87% reliability in “one of the most dangerous and unpredictable waters in the world” with ageing vessels approaching 30 years old suffering frequent mechanical failures.
Three tunnel options: Comparing the transformation
The business case evaluates three configurations, but let's focus on the transformational impact: how accessible does the South Island become?
The problem with Option A is the same that the English Channel Tunnel experiences: That the only difference in time savings is the time to travel through the tunnel. Marshalling road vehicles onto and off wagons takes about the same time as for a ferry. The operational costs are also higher than for a road tunnel. A single tunnel is problematic as only one train will be able to use it per hour. Ferries may be required to cater for excess demand.
Option B is would require two tunnels of small bore. As much as The Boring Company uses this method, it uses it in a loop configuration, which doesn't require twin tunnels. Boring two tunnels doubles the cost and there is no rail option for freight and hazardous goods.
Option C is the most practical option. Single tunnel, two decks, and can accommodate utilities.
Here is the overall time savings for the road tunnel options:
North Island | Distance | Time | South Island | Distance | Current | Tunnel |
Paraparaumu | 40km | 0:40 | Picton | 67 km | 4:00 | 0:40 |
Masterton | 100 km | 1:15 | Blenheim | 95 km | 4:30 | 1:05 |
Whanganui | 190 km | 2:30 | Nelson | 180 km | 5:45 | 2:25 |
Napier | 320 km | 4:10 | Kaikoura | 225 km | 6:00 | 2:40 |
New Plymouth | 355 km | 4:30 | Christchurch | 400 km | 8:30 | 5:10 |
Taupo | 370 km | 4:30 | Greymouth | 410 km | 8:30 | 5:10 |
Rotorua | 450 km | 5:20 | Ashburton | 485 km | 9:10 | 5:30 |
Gisborne | 480 km | 6:20 | Tekapo | 505 km | 8:45 | 6:15 |
Hamilton | 520 km | 6:00 | Franz Josef | 515 km | 9:00 | 6:30 |
Tauranga | 530 km | 6:00 | Timaru | 560 km | 9:45 | 6:25 |
Auckland | 645 km | 7:30 | Oamaru | 645 km | 11:00 | 7:40 |
Whangarei | 805 km | 9:30 | Dunedin | 760 km | 13:00 | 9:30 |
Kerikeri | 890 km | 11:00 | Queenstown | 830 km | 14:00 | 10:30 |
Kaitaia | 960 km | 12:00 | Invercargill | 917 km | 15:00 | 11:30 |
Notice the pattern: Blenheim becomes closer than Masterton. Nelson becomes as accessible as Napier. Oamaru becomes the same distance as Auckland. The entire Top of the South and east coast South Island integrate into the Wellington regional economy.
The economics: Induced demand and multiplier effects
The psychological transformation
The most profound impact isn't about those who currently use ferries—it's about those who don't. A tunnel transforms the psychological barrier from “must book a ferry weeks in advance” to “just drive there.” This shift unlocks massive induced demand.
The precedent is Canada's Confederation Bridge, which connected Prince Edward Island (population 179,000) to the mainland in 1997. Despite tolls of approximately NZ$55, traffic grew from 1 million ferry crossings to 1.5 million bridge crossings annually—a 60% increase. Cook Strait serves a combined catchment of 1.4 million people (Wellington region 550,000; Marlborough/Nelson/Tasman 160,000; Canterbury 650,000). Equivalent per-capita usage suggests 4–5 million crossings annually—eight to ten times current ferry volumes.
The demand curve: Finding the optimal price
Traffic demand responds to price through three distinct user segments:
Essential/Freight traffic (~300,000 vehicles annually): Price inelastic, will pay up to $150 as there's no alternative for moving vehicles
Discretionary travel (tourism, family visits): Moderately elastic, demand halves approximately every $42 toll increase
Commuter traffic (weekly/monthly users): Highly elastic, demand halves every $17 increase and disappears above $100
At a $150 toll (ferry parity), the tunnel carries only existing ferry traffic (~500,000 vehicles, generating $150M revenue). At a $50 toll, the tunnel unlocks the commuter market entirely, generating ~4.5 million crossings and $900M+ revenue. The optimal price point isn't what maximises revenue per vehicle—it's what maximises total revenue through volume.
The multiplier effect: Where savings get invested
When households and businesses save money on cross-strait travel, those savings don't disappear—they're redirected into the broader economy. A family saving $300 on their annual holiday (tunnel $100 vs ferry $400) spends that on accommodation, dining, or activities. A business saving $50,000 annually on freight costs invests in equipment, staff, or expansion.
Treasury fiscal multipliers for New Zealand infrastructure typically range from 1.3 to 1.8. Applied conservatively at 1.5, the $400–500 million in annual travel cost savings generates an additional $600–750 million in GDP contribution. Over a 50-year asset life, this multiplier effect alone contributes $30+ billion in present value—completely separate from the direct transport benefits.
The tunnel doesn't just move people more efficiently; it removes a tax on inter-island economic activity, allowing resources to flow to higher-productivity uses.
At what cost does this become feasible?
Here's where we flip the traditional question. Instead of asking “what does this cost and can we justify it?” we ask: “given achievable revenue, what's the maximum viable capital cost?”
The revenue calculation:
• Vehicle tolls (4.5M crossings @ $50 average): $900M annually
• Utilities revenue (gas, electricity, fibre transmission): $45–65M annually
• Gross revenue: $945–965M annually
• Operating costs: ~$150M annually (far below ferry costs of $270M)
• Net operating income: $750M+ annually
Infrastructure investors typically require 5–7% returns depending on risk. For a tunnel (low risk once operational, 100+ year asset life), a 5% return is reasonable. The viability threshold calculation is straightforward:
Maximum Capital Cost = Net Operating Income ÷ Target Return
$750M ÷ 5% = $15 billion maximum viable capital cost
At a 4% return (patient capital from sovereign wealth funds), the threshold rises to $18.8 billion. This becomes our target: can tunnelling technology deliver a 67km tunnel plus terminals and systems for $15–19 billion?
The cost trajectory: Why planning now pays off
Historical context: What tunnels have cost
Let's examine comparable projects:
• Channel Tunnel (50km, completed 1994): NZ$12 billion at 1994 rates = ~NZ$24B in 2026 dollars
• Gotthard Base Tunnel (57km, completed 2016): NZ$18 billion = ~NZ$280M per kilometre
• Traditional tunnel costs in 2026: NZ$150–280M per kilometre
At traditional costs of $200M per kilometre, a 67km Cook Strait tunnel would cost $13–19 billion for the tunnel itself, plus terminals and systems bringing total costs to $20–28 billion—above the viability threshold.
The cost revolution: Why waiting makes sense
But here's the transformation: The Boring Company has demonstrated costs of $29 million per kilometre with Prufrock-4 (2025) and targets $11 million per kilometre with next-generation machines. Even accounting for the larger diameter required for a Cook Strait tunnel (11–12 metres vs 4 metres for current Boring Company tunnels), if only partial cost reductions are achieved through technology partnership and process innovation, costs could reach $50–100 million per kilometre.
At $75M per kilometre—halfway between Boring Company breakthrough and traditional costs—a 67km tunnel costs $5 billion for tunnelling, perhaps $8–10 billion total with terminals and systems. This sits well below the $15 billion viability threshold and delivers excellent returns to investors.
Technology costs follow predictable decline curves. Consider solar photovoltaics (90% cost reduction 2010–2020), lithium batteries (97% reduction 1991–2018), or LED displays (50%+ reduction during Wellington stadium construction). Tunnelling technology is on a similar trajectory.
The critical insight: A 10–15 year planning horizon allows technology to mature. Beginning serious investigation now positions New Zealand to capture cost reductions rather than locking in today's assumptions.
The time to plan is now
Here's what prudent infrastructure planning looks like:
2026–2028: Commission detailed geological surveys ($10–15M). Engage tunnelling consultants for preliminary engineering.
2028–2032: Monitor tunnelling technology developments globally. Track Boring Company progress on larger-diameter tunnels. Update cost estimates as technology matures.
2032–2035: Make go/no-go decision based on updated technology costs and geological findings. If costs are within viability threshold ($12–18B), proceed to detailed design. If costs remain prohibitive, plan for next ferry generation.
2035–2055: If proceeding, complete design, consenting, and construction.
2055–2060: Tunnel opens, coinciding with ferry end-of-life.
This phased approach requires only $15–25 million over the next 5 years to establish whether this transformational project could become viable. That's trivial compared to the $1.86 billion being spent on ferries that will need replacing in 30 years anyway.
The opportunity cost of not investigating is potentially enormous: dismissing a nation-building project because we locked in outdated cost assumptions instead of planning intelligently for technological change.
Every generation has dismissed this project as too expensive. This generation has the chance to be different—not by committing to build today, but by planning wisely for a future where the technology makes it inevitable.
