> This is a bit of a chicken-and-egg problem but we can point to the Falcon Heavy as a useful data point. There have only been ~13 launches thus far.

Falcon Heavy is only useful for specific missions as it doesn't improves cost economics to LEO over Falcon 9.

> "you can launch multiple payloads in one launch" but you really can't unless they're on pretty much the exact same orbit

This is not true. First of all, many things often go into the same orbit and many orbits are pretty standard.

Second, as you avg kg to orbit goes down, there is an ever greater intensive to use low mass high efficiency engines to do an orbital transfer after launch.

SpaceX just doesn't do that because they do actually need many sats in the same orbit. But its very possible.

> This is only a fraction of the total cost

Its not that small a fraction as you suggest. Internal cost at SpaceX for a Falcon 9 flight is somewhere between 15-30 million $. And of that at least 5-10 million $ are the upper stage. So its actually a huge fraction of the cost.

And the engine on the Upper Stage is also the longest lead most complex to manufacture part.

So you are right, the first stage was a bigger deal in absolute terms but the second stage re-usability is still a gigantic opportunity.

> but it's not really a suitable vehicle for that.

Seems like NASA disagree with that assessment.

> We're nowwhere near even testing that.

They did test launch and landing of the second stage on earth many times. Why do you suggest we are nowhere near testing that.

> And it's going to take a lot of testing for human-rating flight.

Seems like a good place to be for a company that does a lot of testing as part of its development.