The turbopumps in rocket engines may be 'similar' to turbine engines in general, heck, they're the same thing, but they operate in a much more demanding regime, as do all the other components of rocket engines. Check out the temperatures and pressures involved, the power outputs per unit volume, etc. Its many times more than jets, and every ounce is that much more critical. We ARE very good at building them, but they never get really cheap, no matter how many you build. Again, the fundamental HUGE challenge is in terms of materials and mostly design margin, there's just not a lot of room to make stuff cheaper and have it still fly every time you push the button.
Musk is a great manager, but actually his claims in terms of cost reductions haven't materialized yet. He's still more expensive than the Russians were in the 1980's (in adjusted prices). It remains to be seen if he can do better than that. He's a very capable business man and he's done a great job of breaking into an established market, but it was one that really had too little competition and fat complacent US providers that were protected from their cheaper foreign competition. He's just brought the cheaper prices on shore, its far from a huge revolution.
And yes, Mars is about science, but the 2050 generation of rovers will not need to be driven from Earth, will know how to look for new and different stuff (and can of course still be guided by humans), will have greatly increased sample handling and instrumentation, etc. Is there ever going to be a point where a human is 1000x better than a rover? Where it is impossible to design a specific mission to answer a particular question? The cheaper manned spaceflight were to get, unmanned spaceflight will get cheaper even faster. So, is science really served by sending men to Mars? I have to wonder.
But the Moon... Eh, mostly you don't even need more advanced systems than we have now, we can sit on the ground and run machines there today. Why wait? While launch costs will still be high for any manned missions they WILL still be significantly less risky and each mission can be a single self-contained launch, which is a LOT easier to plan and budget and subject to less overruns. Launch windows are not an issue, etc. Its much more appealing at this stage of the game. I get that sending someone to Mars has this emotional appeal, and so maybe it will happen regardless of ANY logic, but the logic is tough to find.