The Boring Company tunnel costs: what is demonstrated, and what 5–10 years might look like
Brief compiled 27 September 2026.
Source notes at the end. This is a research briefing assembled from public company statements, completed-project prices, contracted bids, and published transit-cost studies. It is not an official xAI, Tesla, or The Boring Company document.
Bottom line
The Boring Company has already cut the price of a small-diameter vehicle tunnel by roughly an order of magnitude versus U.S. subway construction. It has not shown the same cut for a conventional rail subway.
- Company 5–10 year target: $3–4 million per mile
- More realistic range for their 12-foot Loop product, if current machines keep improving and they get volume: $8–15 million per bored mile all-in
What “traditional” actually costs
The Boring Company’s talking point has long been that U.S. tunnels cost $100 million to $1 billion-plus per mile, and that they need a greater-than-10× cut. That range is directionally right for American urban rail, and too high as a global or apples-to-apples figure.
Documented U.S. subway projects:
- Second Avenue Subway Phase 1: about $2.5 billion per mile (full project, stations included)
- Phase 2: budgeted near $4 billion per mile
- East Side Access: several billion per mile of new track
A large international database of rapid-rail projects puts the global average closer to ~$350 million per mile, with Spain, Italy, and Korea often well under $250 million and China cheaper still. New York is an extreme outlier because of labor rules, consultants, station caverns, utilities, and project structure — not just the tunnel-boring machine.
A U.S. highway lane is often cited around $3–10 million per mile, which is the benchmark The Boring Company keeps pointing at.
What The Boring Company has actually built and priced
The only completed, publicly priced system is the original Las Vegas Convention Center Loop: about $47 million for 1.7 miles of tunnel and three stations (firm-fixed price, opened 2021). That is about $28 million per route-mile all-in, or about $17 million per kilometer.
| Project | What’s in the number | Implied cost |
|---|---|---|
| LVCC Loop (built, 2021) | 1.7 miles + 3 stations, firm-fixed | ~$47M total ≈ $28M / route-mile |
| Nashville Music City Loop (under construction) | Twin tunnels, ~10–13 route miles, privately funded | $240–300M total ≈ $15–25M / route-mile, or ~$9–12M per bored mile if both tubes are counted |
| Dubai Loop pilot (contracted, not yet dug) | 6.4 km + 4 stations | ~$154M ≈ $24M/km (~$38M/mile) all-in |
| Dubai full route (estimate) | 22.5 km + 19 stations | ~$545M ≈ same $24M/km all-in |
Nashville is the first hard-rock job (limestone). Dubai is the first export price. Both still look cheap versus a U.S. subway and expensive versus the original “$10 million a mile” slogan.
Company language has been consistent for years: Loop tunnels at “order of magnitude $10 million a mile,” with stations, exits, and safety roughly doubling that to about $20 million all-in. President Steve Davis said in late 2025 they are 2–5 years from hitting $10 million, and that the end state is $3–4 million per mile — comparable to a surface highway lane. The Prufrock product page has targeted all-in Loop tunnel cost under $5–8 million per mile.
Why the reduction is real — and why it is not a 100× subway discount
The cost cut is mostly product redesign, not magic dirt removal.
- Diameter. The Boring Company bores ~12-foot tunnels for Teslas. Subway tunnels are often 18–28+ feet, and stations are huge caverns. Excavated volume and lining scale with the square of diameter. Halving diameter is a 3–4× cost cut by itself — the single biggest item.
- Continuous mining instead of “dig 5 feet, stop, erect rings.”
- Porpoising / surface launch — no launch pit, no retrieval shaft; the machine drives onto a trailer.
- Vertical integration — they build the TBM, liners, and electric support vehicles in-house.
- All-electric + “zero people in tunnel” — less ventilation, less labor in the hole.
- One standardized section reused on every job.
Those are real. They do not make a Second Avenue–style station-and-rail project cost $10 million a mile. A 12-foot car tube is not a high-capacity metro. Capacity, fire/life-safety, ventilation, and station cost scale with what you put in the hole.
Cancelled or inflated earlier bids (for example Ontario, California) illustrate the same point: once urban utilities, environmental review, and stations are added, the headline number rises.
A Houston flood-tunnel comparison is a useful illustration: The Boring Company’s narrower tubes were far cheaper than the county’s 30–40-foot design, but they also moved far less water. Price tracked capacity.
Realistic 5–10 year numbers
Company target (roughly 2027–2031): $10 million per mile soon; $3–4 million per mile after the next Prufrock generations and factory scale. Davis tied that to automation and machine cadence.
What the evidence supports:
| Horizon | Estimate |
|---|---|
| Already demonstrated | ~$15–30 million per route-mile for a Loop-class system with small stations |
| Current generation boring | Appearing in the ~$8–15 million per bored mile band (Nashville rock quotes sit near the low end if twin tubes are counted) |
| 5–10 years, if volume arrives | $8–15 million per bored mile all-in for the 12-foot Loop product in ordinary geology. That is a ~10–50× cut versus a U.S. subway mile, and maybe 2–5× versus a cheap European small-diameter utility tunnel |
| Company target of $3–4M/mile | Possible as a raw mining cost in soft ground with a fully utilized remote fleet and almost no station work. Unlikely as a fully loaded urban system cost. Treat as aspiration, not forecast |
This will not apply to conventional large-diameter rail, highway, or water tunnels without redesigning those products down to The Boring Company’s cross-section.
The binding constraints over the next decade are less “can the machine spin faster” and more geology, urban utilities, permitting, and station cost. Digging is the part they have actually industrialized. Everything around the hole still looks like construction.
How to use this comparison without getting fooled
- Always ask whether the number is tunneling only or all-in (stations, exits, vehicles, operations fit-out).
- Always ask route-mile versus bored-mile. Twin tubes double the hole.
- A 12-foot Tesla tube is a different product from a 20-foot rail subway. Cost per passenger-capacity-mile is the fairer metric if the question is “replace the subway.”
- The best published evidence is still Las Vegas plus contracted Nashville and Dubai prices — not the $3–4 million end-state.
Principal public sources
- The Boring Company project pages for Las Vegas Convention Center Loop / Vegas Loop; Prufrock product page (speed and all-in cost targets)
- Steve Davis remarks reported November 2025 (including via Sawyer Merritt): ~$10M/mile in 2–5 years, $3–4M/mile longer term; traditional projects $500M–$3B/mile depending on size and location
- LVCC Loop firm-fixed price of approximately $47 million for 1.7 miles and three stations
- Nashville Music City Loop published figures in the $240–300 million range for the privately funded twin-tunnel airport–downtown system; some briefings cited ~$15 million per mile for the machine to dig a mile
- Dubai Loop agreement figures: ~$154 million for a 6.4 km / 4-station pilot and ~$545 million for a 22.5 km / 19-station full route
- NYU Marron Institute / Transit Costs Project work on Second Avenue Subway and global rail construction costs
- Contemporaneous reporting on Second Avenue Subway Phase 2 budgets (~$7.7 billion for ~1.8 miles)
Figures move. Treat every per-mile number as an order-of-magnitude comparison unless the source states what is included.