Things used to be a lot cheaper to build than they are now. There are a lot of contributing factors, but a big one is that we already have a lot of things, and new things have to work around the existing ones. For example, the early subways were built with cut-and-cover, where you dig up a road and then cover it up. These days people generally bore tunnels instead. Why? Because it's much less disruptive. With cut-and-cover, not only do you lose the use of the road while you're putting in the tunnel, you also have to figure out what to do about all the other underground infrastructure in your way. There's water, sewer, gas, electric, cable, and you can't just dig through them. While cut-and-cover is much cheaper if there isn't existing infrastructure, at this point there are a lot of things under our streets.

Here's Marohn of Strong Towns writing about Lafayette LA:

When we added up the replacement cost of all of the city's infrastructure—an expense we would anticipate them cumulatively experiencing roughly once a generation—it came to $32 billion. When we added up the entire tax base of the city, all of the private wealth sustained by that infrastructure, it came to just $16 billion. This is fatal.

This is the same problem: building things for the first time is much cheaper than rebuilding them while they're in use.

Can we replace infrastructure wholesale?

We take a small city, come in with a truly enormous amount of equipment and people, disconnect the existing water / gas / electric supplies, dig up the entire city, and put in new everything. Because we're replacing everything we don't have to care what's already under the street, which speeds things up enormously.

Run this on an incredibly short timeline, like a week. If you can dig up one block in a week then you an dig up all the blocks in a week. You just need a lot of labor and equipment.

What do the residents do? Give everyone $2,000 to go on vacation. You don't have to take a vacation, but you do have to leave. Compared to the overall cost of the project this is pretty small.

How do you have enough labor and equipment? You hire a lot of people, and go city-to-city. Maybe you do 20 cities a year.

It still makes sense to do maintenance, since some things will break much earlier than designed, but projects like "replace the water main" wouldn't make sense.