I wrote this entire post and forgot that I didn’t post it.
It’s time for some more napkin math posting.
Chicago’s O’Hare, one of the busiest airports in the USA, handled 867,049 flights in 2017, with 79.8 million passengers. That’s about 92 passengers per flight, and 2,375 flights per day.
What scale of rail station would need to replace it? Well, that’s about 218,630 passengers per day. A random Japanese bullet train (a 16-car 700-series shinkansen) has a seating capacity of 1,323. (Equivalent to about 14 aircraft - not surprising, since train cars are similar to passenger aircraft fuselages in some ways.) So the station would need to support about 166 trains per day, or about one every 8-9 minutes.
If we give every train 30 minutes to unload, that’s about 3x 1.3k passenger platforms, probably 5-6 so we can play around with train arrival times. That’s honestly entirely manageable as far as construction projects go.
Let’s assume we want a similar speed to commercial aircraft, so we want the train to travel from NYC to Chicago in about 3 hours. That’s about 700 miles, so we’ll round up to 800 for various diversions the track has to make to handle geography. That makes for about 268 mph.
China built a maglev train in Shanghai. It moves at about 268 mph. It cost them about $1,200,000,000 for 18.95 mi of track (or about $63.3M/mi). Â
For an 800 mile track, that’s about $50.7 billion dollars. (Of course, New York City is apparently paying something ridiculous like $2.6 billion dollars per mile for subway construction. While not all American transit construction is quite that stunningly overpriced, we should probably assume Chinese magrail construction costs are around half of what it would cost to build the same length of track in America, or worse. For now, we’ll go with the optimistic estimate.)
Checking quickly shows a flight from NYC-Chicago costs about $100-110. The track is worth about 507M passenger flights. Assuming that 10% of O’Hare’s passengers comes from NYC, the construction costs of the track are worth 63 years of flights.
How much does a train ticket cost? Amtrak NYC-Chicago looks to be around $100-110 anyay, and a quick search for tickets in Europe suggested around $79-120 for a train ticket from Paris to Rome (also about 700 mi). Â
So here’s the issue. The cruising speed for a 787 Dreamliner is over 500 mph. In practice traversing ~720mi in about 2.5hr gives an average of 288 mph. Â
...but the range of a 787 Dreamliner is about twice the distance from Miami, Florida to Anchorage, Alaska. A flight to O’Hare could conceivably come from almost any medium-size airport in the continental United States.
That’s like having a maglev line between each US city with a population over 250,000. Â
A quick search suggests there are about 82 of those. If they’re typically about 200 miles apart... assuming we model it as a loop, we’re looking at something like 16,000+ miles of track and 1 trillion dollars in spending. In reality, it would probably be more like 4 to ten times that to build the whole network.
Okay, but the trains are higher capacity, right? Assume one train leaves every 30 minutes. You can move ~63,500 people per day. But this is a train we’re talking about, here. We could send 1 train every 10 minutes, for 190,512 passengers per day.
But can you do it for less than $100 per passenger? We already know the market demand for $100 3-hour intercity trips. If there were 3 times the number of people willing to pay $100 for that trip, the airline companies would probably have already purchased 3 times the number of aircraft.
So let’s say it now costs $50, or perhaps even $30. Well, now we need to cover food, lodging, opportunity cost of not working (unless this is a business trip), transit to and from local sites... and most of those are more time-sensitive resources that need to have excess capacity to handle peaks, so they’ll still be relatively costly, meaning that the overall cost of the trip is only reduced by maybe 10%.
