- Light rail transit construction — urban transit build pace in international context.
- Pipeline construction — how Canada builds major energy linear infrastructure.
- Electricity generation capacity — whether the grid is being built fast enough for demand.
All cost figures on this page are capital construction cost, converted to 2026 CAD at May 2026 FX rates (1 EUR = 1.48 CAD; 1 USD = 1.38 CAD; 100 JPY = 0.90 CAD) and inflation-adjusted to 2026 where the construction midpoint precedes 2024. Rolling stock, depot, and long-term maintenance are excluded where they can be separated from the construction figure.
Five recent dedicated high-speed rail projects, plotted four ways: capital cost per kilometre, capital cost per resident in the served corridor, calendar years per kilometre, and capital cost against corridor population density. Toggle between the views below.
Five recent dedicated high-speed rail projects, sorted ascending. Alto highlighted in red at $83.5 M/km — the infrastructure capital figure in the 2023 internal briefing; published envelope $60–90 M/km.
At $83.5 M/km, Alto is priced well above the mature European lines restated in 2026 dollars (Madrid–Barcelona $29 M/km, LGV Est $36 M/km) and below modern Shinkansen ($120 M/km). The honest peer set is Hokuriku and California ($219 M/km).
All five projects sorted with Alto first, then by ascending capital cost per kilometre within each tier.
| Project | Country | Tier | Length (km) | Top km/h | Build period | Capital cost | Capital/km (CAD) | Pop. served | CAD/person | Yrs/km |
|---|
Capital construction cost only; rolling stock, depot, and long-term maintenance excluded where they can be separated. Foreign currencies converted to CAD at May 2026 rates (1 EUR = 1.48 CAD; 1 USD = 1.38 CAD; 100 JPY = 0.90 CAD) and then inflation-adjusted to 2026 using local CPI. Population served is the sum of metropolitan-area populations of stations on the line, using each country's standard metropolitan definition. Alto's 18 M figure is the official Transport Canada estimate; Hokuriku's ~3 M is the regional corridor only and rises to ~38 M with through-service to Tokyo (noted in the table).
The published $60–90 B figure is a Class 5 estimate (Alto's August 2026 document says so in terms), which is the band set before design, not after it. The one itemized number, $83.5 B for infrastructure in the 2023 internal briefing, lands at the top of that band. Even so, $83.5 M/km is below Hokuriku ($120 M/km) and California ($219 M/km), and more than double Madrid–Barcelona (€14.4 M/km nominal in 2008 ≈ $29 M/km in 2026 CAD) and LGV Est (€18.7 M/km nominal in 2007–2016 ≈ $36 M/km in 2026 CAD). Both of those were built by sovereign HSR primes with five or more prior lines in operation, mostly through low-density farmland, with no comparable expropriation overhead. None of that applies to the Quebec–Windsor corridor.
Hokuriku Kanazawa–Tsuruga delivered $120 M/km in 2024 dollars through mountain terrain under Japan's disciplined JRTT model. California Phase 1 is tracking at $219 M/km through flat-to-rolling Central Valley under a public-corporation model that closely resembles Alto. Alto's Ottawa–Montréal first segment must expropriate across ~1,700 properties. That is closer to California than to France.
At ~18,000 people per km of line, Alto serves roughly half the catchment density of LGV Est or California Phase 1 (both ~32–33k/km). Lower density means weaker fare-box recovery, which puts upward pressure on the public capital subsidy required per rider. At the $83.5 B infrastructure figure, that works out to ~$4,640 CAD of public capital per resident, more than triple the French and Spanish precedents.
Alto's published aspiration of 1,000 km in 12–14 years, or 0.013 yrs/km, would be the fastest per-km HSR build in the comparator set. Madrid–Barcelona, which is the closest precedent in scale and was delivered approximately on plan, took 0.021 yrs/km with an established AVE supply chain. The first Ottawa–Montréal segment is the testable promise: if it opens in 2037 at $25–30 B CAD ($125–150 M/km), Alto is on a California trajectory, not a French one.
Alto's August 2026 economic case states that operating revenues are expected to fully cover all operating and maintenance costs once the line is running. The 2023 internal briefing obtained by The Globe and Mail in August 2026 is the first document to put numbers behind that claim. The numbers support it, but only for the claim exactly as written. Over forty years the briefing projects $105 B of revenue against $62.6–67.0 B of operating and maintenance cost. The fare box clears O&M with roughly $38–42 B to spare. It does not come close to clearing the build.
Undiscounted nominal totals as set out in the 2023 internal briefing. Operating and maintenance plotted at the $64.8 B midpoint of its $62.6–67.0 B range; capital is $83.5 B infrastructure plus $2.1 B rolling stock.
Revenue covers operating and maintenance roughly 1.6× over. Set against the whole programme, the same revenue leaves about $43–48 B of capital never recovered from riders, some $2,400–2,650 per corridor resident, undiscounted. That residual is the number a benefit–cost case has to justify on wider economic grounds.
The $148.1–152.7 B in the briefing is the sum of three things: infrastructure capital ($83.5 B), rolling stock ($2.0–2.2 B), and forty years of operating and maintenance ($62.6–67.0 B). It is not a revised capital estimate, and it is not comparable to the per-kilometre figures above, which are capital only. Read fairly, the $60–90 B public envelope is not contradicted by it. The two are answering different questions. The criticism that survives is narrower and harder to answer: the operating half has never been published alongside the capital half, so the only number in public circulation describes rather less than half the forty-year commitment.
Alto's cost-recovery claim is true and narrow. On the briefing's own projections, $105 B of revenue against a $64.8 B midpoint for O&M leaves roughly $40 B, enough to clear operations with margin, and roughly half of the $85.6 B capital programme. The remaining $43–48 B is carried by the public and is never returned through the fare box. Alto offers $49.5 B of traveller and community benefits to justify it. Those two figures should not be netted against each other: the benefits are welfare estimates over sixty years of operation, the capital residual is nominal cash over forty, and neither is discounted to the other's basis.
Alto's own August 2026 document forecasts up to 24 million riders a year by 2055, rising to up to 43 million by 2084. The federal Major Projects Office project page states, without qualification, that Alto "will serve up to 43 million passengers annually". No year is attached to it anywhere on that page. The higher number is a fifty-eight-year forecast presented on a government project page as a characteristic of the service. The near-term forecast that matters for the revenue arithmetic above is the 2055 figure.
At $83.5 M per km, Alto ranks third of five peer projects. It is well above the mature European lines and below Hokuriku and California, on a pre-design estimate.
- Compared with
- Four peer projects the page benchmarks Alto against: LGV Est, the Madrid–Barcelona AVE, the Hokuriku Shinkansen Kanazawa–Tsuruga extension, and California HSR Phase 1.
- Where Canada sits
- At $83.5 M per km, Alto ranks third of five, behind Madrid–Barcelona ($29 M/km) and LGV Est ($36 M/km) and ahead of Hokuriku ($120 M/km) and California ($219 M/km). It also has the thinnest corridor in the set, at about 18,000 people per km.
- Which way it is moving
- Up. The one itemized figure, $83.5 B for infrastructure, sits at the top of the $60–90 B Class 5 envelope rather than its midpoint, so the cost reading has moved toward the dear end. Construction has not started, so there is no delivered trend.
- What the verdict follows
- Position in the peer set. The page argues the fair peers are Hokuriku and California, and it says the $83.5 M figure is, if anything, understated against 2026-dollar comparators. Alto sits at the lower edge of Watch. One place lower would be Weak.
- Data basis
- Transport Canada (Dec 2025 Ottawa–Montréal selection), Alto's economic case (20 Aug 2026), and the 2023 internal briefing reported by The Globe and Mail (31 Aug 2026), with Adif Alta Velocidad, Railway Gazette / EIB project files, JRTT, the California High-Speed Rail Authority Draft 2026 Business Plan, and the California Legislative Analyst's Office for the peers. Costs are in 2026 CAD at May 2026 FX. Page last reviewed September 2026 · Data current to August 2026.
Methodology notes
Cost basis — capital construction only. LGV Est: combined Phase 1 (€5.5 B, RFF/national-government/EU/local-authority funding, 2002–2007) and Phase 2 (€2.01 B, 2009–2016) per Railway Gazette and EIB project files; rolling stock excluded. Madrid–Barcelona AVE: ~€9.0 B total construction investment per Adif Alta Velocidad's historical record and Global Railway Review (some sources report €9.5 B excl. land and stations). Hokuriku Kanazawa–Tsuruga: ¥1.67 trillion per JRTT and IHRA. California HSR Phase 1: $126.3 B 2026 Draft Business Plan figure per the California High-Speed Rail Authority and the Legislative Analyst's Office (April 2026 briefing). Alto: $60–90 B preliminary federal envelope as announced by Transport Canada (Dec 2025 Ottawa–Montréal segment selection). Foreign-currency totals are first converted at May 2026 FX (1 EUR = 1.48 CAD; 1 USD = 1.38 CAD; 100 JPY = 0.90 CAD), then inflation-adjusted to 2026 using local CPI for projects whose construction midpoint is more than two years before 2026.
What the Alto envelope does and does not contain. Alto's August 2026 economic case establishes two things the December 2025 announcement did not: the $60–90 B figure is a preliminary Class 5 estimate (the band set before design, the widest on the estimating ladder), and it is stated in 2024 dollars, not 2026. Transport Canada has still never itemized it: whether it includes land acquisition, rolling stock, or contingency has not been published, and it is an envelope, not a let contract. The 2023 internal briefing reported by The Globe and Mail in August 2026 is the only itemization in the public record (infrastructure $83.5 B; rolling stock $2.0–2.2 B), and this page now plots Alto at $83.5 M/km on that basis, retaining the $60–90 B band in the table and tooltips. Two cautions attach to that figure: it is a September 2023 document, so it predates the February 2025 announcement and is not a revision of it; and it is stated in the dollars of its own day, so against 2026-dollar comparators it is if anything understated. If land acquisition is excluded, and the Ottawa–Montréal segment alone crosses roughly 1,700 properties, Alto's like-for-like per-kilometre figure is understated relative to the comparators, which strengthens rather than weakens this page's conclusion. The same treatment issue is flagged for Madrid–Barcelona above (€9.0 B used; some sources report €9.5 B excluding land and stations). The overrun-sensitivity table in the companion workbook (multipliers from 1.7× to 3.8×) is how this page bounds the envelope's uncertainty rather than pretending it away.
Why the forty-year figures are not on the comparator charts. The $148.1–152.7 B total, the $105 B revenue projection and the $62.6–67.0 B operating cost are lifecycle figures for one project. No equivalent forty-year operating series has been published for LGV Est, Madrid–Barcelona, Hokuriku or California Phase 1 on a basis that could be held constant, so putting Alto's lifecycle number on a chart of capital-only outturns would be exactly the mismatch this project's comparator rule forbids. They are therefore treated in their own section, against Alto's own claims, and never ranked against the peer set.
Why California costs $219 M/km. The anchor of the "honest peer" argument deserves its own explanation, because the causes are process failures, not geography. The Central Valley is flat-to-rolling farmland. Per the California Legislative Analyst's Office and CHSRA's own business-plan reviews: construction began before design and right-of-way acquisition were complete, generating thousands of change orders; there has never been committed full funding, so work is sequenced stop-start around bond tranches and cap-and-trade receipts, capitalizing 23 years of inflation; CEQA litigation and piecemeal parcel-by-parcel expropriation slowed the Central Valley segments; and politically driven alignment choices plus community-demanded viaducts and grade separations added scope. Cumulative cost growth since the 2008 baseline is ≈3.8×, the top multiplier in the Alto sensitivity table. The relevant point for Alto is that it shares the public-corporation delivery model and the expropriation burden, not the funding chaos — yet.
Population served. Sum of metropolitan-area populations of stations on each line, using each country's standard metropolitan definition (INSEE aire urbaine; INE Spain functional urban area; Japan e-Stat metropolitan employment area; US Census Combined Statistical Area; Statistics Canada Census Metropolitan Area). Alto's 18 M figure is the official Transport Canada estimate covering Toronto, Peterborough, Ottawa–Gatineau, Montréal, Laval, Trois-Rivières, and Québec City. Hokuriku Kanazawa–Tsuruga's ~3 M is the regional corridor only (Kanazawa, Toyama, Fukui, Tsuruga, and intermediate stops) and rises to ~38 M with the existing through-service to Tokyo via the rest of the Hokuriku Shinkansen. The regional figure is used as the like-for-like measure of population the new infrastructure was built to serve.
Schedule. Build-period years are calendar years from major civil works start to revenue service. For California Phase 1, end-year uses the 2038–39 projected Phase 1 completion in the 2026 Draft Business Plan (the Initial Operating Segment Merced–Bakersfield is on the 2032 schedule). For Alto, planning numbers are placeholders. Actual yrs/km will only be measurable once Ottawa–Montréal Phase 1 opens, which is why that row is flagged "planned" in the years-per-km view.
Primary sources by project
