SpaceX Commercial Crew will cost 55M per seat for NASA. Add to that need need to amortize 2.3 billion in development financing that NASA has given out (includes 2 astronaut flights). SpaceX can sell the other seats on the capsule.
They payed almost 90M per seat for Soyuz at least in the last couple.
In comparison, the program cost of Shuttle per flight was 1.5 billion per flight. They would almost never fly more then 4 people on those flights (just as Dragon can also take 7 people, but NASA only takes 4). And the Shuttle was not getting cheaper and the flight rate was not getting faster.
NASA of course has 2 contractors the other was 4.2 billion and the per seat cost is somewhat higher, not sure exactly with the numbers, like 70M or so.
What the per seat cost overall cost are we will see over time. After the initial maximum flight Commercial Crew contracts run out, NASA can renegotiate after that initial contract and likely get a lower per seat price then.
We will see what the end-to-end lifetime per seat cost are for NASA.
I can't give you perfect numbers for perfect Cargo without a lot of research. It depends if we are talking CRS1 or CRS2 contract and there are 3 providers that are all somewhat different, different proportions of up-mass, down-mass and other capabilities, some that the Shuttle didn't have.
The Shuttle could theoretically transport a huge amount of unpressurized cargo but if you are not building a Space station its hard to actually utilize that space fully. So an analysis would have to include the actual payload and so on. That is getting to complex here.
What NASA values seems to be having much less risk on one system. Having different vehicles on different rockets and being able to deliver what they want when they want it and get it back when they want it back.
SpaceX charges about $46m per ton of upmass. So for the ~10 tons of upmass per mission, that's $460m. (Plus I think Dragon is pretty volume constrained compared to the MPLM that NASA used with Shuttle.)
SpaceX's price per seat is $55m, Boeing is $90m, same as Soyuz.
And it's false that "[t]hey would almost never fly more then 4 people on those flights", as you can see from the list of Space Shuttle missions: https://en.wikipedia.org/wiki/List_of_Space_Shuttle_missions (5th column lists number of crew... To ISS it's usually 7 crew but sometimes 6 and only one time 4). The reason why is they could do a "crew surge" where they could get a whole bunch of work done on Station for a couple weeks and then the extra folk would go back down.
And it's false that Dragon can now take 7 people... They switched to 4 because of the need to change the orientation of the seats to ensure safe splashdown and the 7 seat configuration no longer fit. There were some configurations of Shuttle that would've allowed like 74 additional passengers, too, with a passenger module, but they never flew that variant either.
So again, 7 seats at Boeing's $90 million would give an additional $560 million per mission (or $385m for SpaceX), for a total of $1.1 billion (or $845m), not counting the capaiblity of the extra airlock or the extra robotic arm for servicing capability, plus the ability to take up full ISPR racks or return large pieces of external cargo or...
...So again, even with the $1.5B figure (which includes a ton of overhead at KSC and JSC and elsewhere that would probably have to be covered in some other way), it's still pretty competitive with commercial crew. And that's not counting the NASA side of the commercial crew program, which is substantial.
So I'm with you on commercial crew and cargo being good. I think two or more providers is a more robust system, and I agree it would've been possible to build a big space station without Shuttle (the Russians did it). But the more you dig in to Shuttle's capabilities, the more expensive it looks to replicate. Shuttle really wasn't that bad considering how compromised the design was by zipcode engineering and inability to effectively cost share with commercial or military stuff (later on) and being essentially a traditional government contractor run system.
...the all-in price per launch of Shuttle may have been about $1.5billion, but the marginal price was around $500 million. If they had ever used that 74 passenger module (plus 6-9 crew), that would've been a marginal price per seat of just $6 million... Not too bad, really.
But it was never cheap enough or robust enough to launch often enough to ensure safety without launch abort. Hopefully Starship will change that, enabling a marginal price per seat of something like $100,000 for a trip to orbit...
They payed almost 90M per seat for Soyuz at least in the last couple.
In comparison, the program cost of Shuttle per flight was 1.5 billion per flight. They would almost never fly more then 4 people on those flights (just as Dragon can also take 7 people, but NASA only takes 4). And the Shuttle was not getting cheaper and the flight rate was not getting faster.
NASA of course has 2 contractors the other was 4.2 billion and the per seat cost is somewhat higher, not sure exactly with the numbers, like 70M or so.
What the per seat cost overall cost are we will see over time. After the initial maximum flight Commercial Crew contracts run out, NASA can renegotiate after that initial contract and likely get a lower per seat price then.
We will see what the end-to-end lifetime per seat cost are for NASA.
I can't give you perfect numbers for perfect Cargo without a lot of research. It depends if we are talking CRS1 or CRS2 contract and there are 3 providers that are all somewhat different, different proportions of up-mass, down-mass and other capabilities, some that the Shuttle didn't have.
The Shuttle could theoretically transport a huge amount of unpressurized cargo but if you are not building a Space station its hard to actually utilize that space fully. So an analysis would have to include the actual payload and so on. That is getting to complex here.
What NASA values seems to be having much less risk on one system. Having different vehicles on different rockets and being able to deliver what they want when they want it and get it back when they want it back.