SpaceX runs basically as a giant R&D program that also happens to accomplish useful things as byproducts of its approach (with the miracle of insurance to cover costs if things get explody). Every flight is data for incremental improvement.

If you think of all SpaceX has ever done was try to get to this point, with somewhere around 80 flights on the Falcon 9, but for a variety of mission buyers (NASA, USAF, Commercial Sat companies, SpaceX itself!), somewhere around $5b (~$60m a launch) has been spent getting an entirely new, ground-up manned space program up and running to include launch vehicle with a majority re-usability. But along the way, numerous satellites and other stuff ended up in space as well. This allowed NASA to also split the R&D bill with multiple "partners".

To put this into context, the Mercury program ran about $2.25b in today's dollars, Gemini ran $7.3b ($723m per mission!).

The Space Shuttle program had a lifetime cost of $209b ($1.5b per flight!!!) If you think the Shuttle program is not comparable, the F9Heavy program, a derivative of this other work, can put up more mass than the Shuttle, managed to split much of the development costs with the Falcon 9, can land most of its mass, and has a per flight cost of $90m.

We now have the capability to put up another ISS today, resupply it, and staff it, at a fraction of the cost, all with launch equipment made by a single company that started a decade ago as a side project for a guy who was trying to make electric cars after making money on the internet.

This is a ridiculous bargain and Spacex isn't even close to done yet.