The case for a railway of this size rests, in the end, on a forecast. The number of riders sets the benefit, drives the ticket revenue, and sits underneath every cost-per-passenger figure anyone quotes. Choose the ridership and you have largely chosen the economics.

So we answered the question twice, on purpose, using two methods that were free to disagree.

The first asks what comparable railways actually carry, and where this corridor sits among them. The second builds a number up from the ground: how many people live along the line, how often they travel between its cities, and what share of those journeys a train would win. The first gives a floor grounded in what people really do. The second gives a ceiling based on how travel choices work. Where they differ is itself worth knowing.

The measure is simple: intercity rail trips in a year, divided by the population the railway serves. Trips per person per year. It lets you compare countries of very different size.

Across twelve operating systems, two tiers appear. Networks that serve a whole country — France, Germany, Japan, Taiwan, Korea, Spain, Italy, Sweden — sit between 1.0 and 3.4 trips per person. Single corridors — one dominant city pair, or a cross-border link — sit near 0.3, whatever technology they use. Eurostar is 0.35. Brightline in Florida is 0.31. The Boston–Washington Northeast Corridor is 0.30.

Toronto to Québec City is long, but it is a corridor: a single chain of cities. Today VIA carries 0.23 trips per person along it.

Why Taiwan carries ten times what the Northeast Corridor does

Not speed, and not the price of a ticket. The thing that predicts intercity rail use better than anything else we tested is how committed the surrounding society is to the private car — how many cars per thousand people, how cheap the fuel, how few the tolls, and how poor the transit at the far end. We combine those four into a single score.

Measured that way, this corridor scores 71. That puts it beside VIA today (71), Brightline (73) and the Northeast Corridor (64) — and a long way from Sweden or Taiwan. The fitted relationship predicts about 0.29 trips per person, which is roughly four to five million trips a year. Barely above the 0.23 that conventional service already achieves.

The case for building fast leans on three things: the train is quick, it can be made cheap to ride, and it runs through a populous corridor. Tested against real railways, none of the three works the way it is assumed to.

Speed explains almost nothing. At around 300 km/h, ridership across the set ranges from 0.35 to 3.4 — a tenfold spread at the same speed. At 200 km/h, Sweden carries 2.9 and Brightline carries 0.31. Once car dependence is accounted for, speed adds nothing measurable. The world’s fast railways are mostly its least car-dependent ones, which is a different fact.

Density does not rescue it either. The Northeast Corridor has roughly three times this corridor’s population per kilometre of route. That threefold advantage buys it about 0.07 more trips per person — the difference between 0.23 and 0.30. Not the two- or three-fold gain a high-speed business case needs.

The fare does move it. The same physical railway carries about 36 per cent of corridor travel under heavy subsidy and about 19 per cent under minimal subsidy — close to a doubling, decided by policy rather than engineering. That is the real lever, and it is a recurring cost rather than a capital one.

Measured centre to centre, a 177–240 km/h railway already beats flying between Toronto and Ottawa — about 2h55 against roughly 3h30 once you count getting to the airport, security and the trip in at the other end. Toronto to Montréal comes out at about 3h32 against 3h45 by air. The extra speed of a 300 km/h line buys perhaps twenty-five minutes on that journey, at several times the capital cost per kilometre.

This is the part that surprises people. The faster, more expensive railway does carry more — but not by much, and not for the reason usually assumed.

At maturity, for every 100 riders the greenfield high-speed line carries, the brownfield route carries about 84. Part of that gap is speed. Part of it is simply that the high-speed line also runs on to Québec City, which the brownfield route does not.

Over the whole horizon the gap is much smaller. Counting every passenger carried between opening and 2080, the two are 5.7 per cent apart — about 321 million against 341 million. Both open in 2040, so neither gets a head start. The brownfield route opens as a single stage and is running at full strength by about 2050, while the high-speed line phases in over three stages to 2050 and only passes the brownfield route on the running total in 2069.

And the remaining gap can be closed with the fare. Price the brownfield route 11 to 18 per cent below the high-speed line and the two carry the same number of people. Price it twenty per cent below and it carries more at every point across the fifty years — a discount that sits inside its subsidy budget precisely because its costs are so much lower. Matching that discount would oblige the high-speed line to give up revenue it can far less afford.

Our central figures are about 9.2 million a year for the high-speed corridor and 8.2 million for the brownfield route, around 2055. Those sit just below the published forecasts — and well below the twenty-four-million target.

There is also something worth noticing in the fourth. The Joint Project Office — the federal body that has since become Alto — put corridor ridership at 11.0 million by 2034 and 16.8 million by 2059 in its 2021 business case. That works out at roughly 15.5 million around 2055. The twenty-four-million target now advanced for the same corridor is about half as much again, and the increase arrived with the change from high-frequency to high-speed rather than with any new measurement.

Read all the methods together, the honest range for 2055 is somewhere between four or five million and twelve or thirteen million. The published forecasts cluster in the upper half of that. Alto’s target sits above all of it.

The discipline this report asks of Alto is that a forecast be tested against what comparable completed projects actually achieved, rather than defended from inside its own assumptions. A standard applied only to the other side is not a standard.

The estimate is bounded rather than precise, and the bounds are wide. The comparison set is twelve systems and predicts its extremes better than its middle. How often people travel between these cities is the least settled input, and ridership scales directly with it. Only one leg of the car market is measured at a roadside count; the others are sized from travel demand. And Alto’s own demand model is unpublished, so we can measure the size of the disagreement but not say which assumption causes it. That is why our first recommendation is that the model be released for independent audit.

So here is what would overturn ours. The empty quadrant rests on an absence: one car-committed society carrying intercity rail ridership well above one trip per person would remove our strongest argument at a stroke. None appears in the record we examined. Statistics Canada’s National Travel Survey showing corridor travel materially higher than we assume would raise every figure here in proportion — though not the ratio between the two railways, which is why we report that ratio separately. A simulated journey time showing a larger speed advantage than the twenty-five minutes we estimate would narrow the gap.