> Sure, at this stage of any project it's easy to be "looking at" very low costs. They haven't done anything yet. The nature of these kinds of projects is there are a whole bunch of costs, technical an regulatory, that aren't apparent until you actually start building something.
I don't agree. For example, the Space Shuttle estimates were about on the money. As in they said, the cost estimate is $X, but we'll need $1.2 X, to allow for obvious contingencies. President Nixon went: we don't budget contingencies, we'll give you X and then fund the overrun later. NASA: OK boss.
So actually, it cost what they said, but it looked to the rest of the world like an overrun.
And the Space Shuttle main engine was about as complicated as SABRE looks like it will be, maybe more so, it was an unreasonably complicated design.
And Reaction Engines actually have a careful design, with computer modelling of everything. That bodes well for a relatively straightforward detailed design and build. The X-33 had none of that, and when they got around to it, they found the horizontal stabilisation was total shit.
And the engine is particularly clever in that it works almost the same at all speeds; the precooler means that it doesn't care whether it's at ground level or Mach 5, the air behind the cooler is at the same temperature. That means, like a rocket engine, they can do almost complete testing when stationary. And the precooler also, they've already tested the precooler; it works fine. And the precooler was the most challenging bit of the whole system; it's something like half a gigawatt per tonne of cooling.
The take-home message is not that it's not a clever design, it's that most of the clever bits are easy to ground test. About the only bit they can't totally test on the ground is the aerodynamics of the aeroshell- but that was basically the same problem that the Space Shuttle faced and dealt with.