The bigger problem is that I casually listed average vehicles per minute, but in practice you likely have barley any traffic at night and four times the average during peak hours.
Also "finance over 50 years for free" is a giant stretch. Even at a theoretical 0% interest rate the opportunity cost makes this much more expensive in reality.
People might begin storing vehicles in buildings more often in places like Manhattan if any of this stuff works. You then leave Manhattan with your car by linking into the transport tunnels, driving wherever you please from there. It wouldn't be difficult to add floors to new buildings just for automated vehicle storage, with a tunnel linked right to a given building that has vehicle storage.
If I'm from Pennsylvania and I want to go to Manhattan for a weekend, I put my car on a sled, go into Manhattan via tunnel, park my sled/car in building storage linked to a tunnel, then do whatever in Manhattan. Then reverse the process for leaving. Dramatically more convenient than the way things are now.
Maybe it ends up making more sense to have the bulk of vehicle storage just outside of expensive real-estate zones, so the sleds into Manhattan are multi-person carriers instead. Then you take a multi-person carrier sled back to wherever you stored your vehicle. Either way.
Suppose they can do it for $75 million/mile. That gives us tunnelling costs of $250 billion, or about 1 million a day if financed for free over 50 years.
If we assume he wants to move sledded cars instead of passenger vehicles, a ticket price of $120 per transported car seems reasonable to me. That means he needs about 8300 cars/day, or on average 6 per minute. Transporting cars should also cuts down on station costs.
If he wants to transport passengers, at ten per hyperloop vehicle, $60 per person, that's one full vehicle per minute, not that far from some subway lines.
It's doable, but far fetched. I consider his plan to put humans on Mars within 7 years more realistic (but only very slightly so). But in any case people trying is not the worst thing to happen to tech, it's the essence of tech.
1: https://www.boringcompany.com/faq/ ($1 billion/mile reduced by factor >10)