Why assign a price to a spacecraft in, say, 2237? Whatever it costs will be quoted in rupees, florins, or some new unit, and the raw figure wouldn't mean a thing to us if we knew it. Which is just the point. We want to know how much a spaceship costs in context - how many people can afford a ticket, how many schools can be built for the cost of a gigawatt laser star.
In this ongoing discussion we're considering a recognizable future economy and technology. No Singularity, and not enough time for society to have evolved out of recognition. Perhaps the Industrial Revolution will reach maturity, in which further refinements are gradual. The first decades of aviation had Moore's Law style progress, but jetliners have had much the same performance and appearance for the last 50 years. A lot of low hanging fruit has already been picked.
In any case, we're dealing with a fairly recognizable future economy and technology, able to do a lot of things we can't, but doing many familiar things in familiar ways. In 2237, long distance air transport (the only kind likely to be common) will probably be at high subsonic speeds, in jetliners that mostly just look and fly like jetliners. They may have nanomaterial strips instead of flap actuators, but only geeks would notice.
Spacecraft will also be broadly recognizable, because we already build them. We've built many small robotic ones, and are starting to build big ones. The International Space Station is not unlike the forward end of an interplanetary ship, minus the main drive with its big propellant tanks and radiator fins or solar wings.
The ISS has a 'dry' mass - structure and equipment, minus all consumables - of 304 tons, about 10 percent more than the super jumbo Airbus A380 (277 tons). They are also roughly the same size, in overall dimensions (both having 'wings') and in bulk, with pressurized volume about 1200 cubic meters. The ISS has a crew of six as an exploratory craft, and about a third of the volume is living space, so if configured as a transport, so to speak, it might carry 20.
An improved and mature technology will surely improve on the ISS, but the most desirable improvement, better toilets aside, is making it cheaper. The ISS has cost something like $50 billion, about $170 million per ton. This is too much. We can't be getting around in spaceships that cost $50 billion just for the forward section, minus the parts that scoot it around.
As a comparison, the slightly smaller Airbus A380 can be yours for about $325 million, depending on options, or about $1.2 million per ton. 747 variants are comparably priced. In the broadest structural engineering sense they are all similar constructions: lightweight pressure tanks and trusses (e.g. wing spars), fitted with complex, lightweight equipment and electronics. So what accounts for the difference in cost?
The notorious Washington 'iron triangle' bears a share of the blame, from outright fraud to shameless overcharging, to subcontracting work out into 400 congressional districts. Throw in policy changes that can throw out years of design work.
Also the ISS is a prototype, in a fact a sort of meta-prototype. Skylab and Mir were caravels; the ISS is our first try at a galleon. The builders not only had to handbuild it, which makes any prototype costly; they had to work out all the design requirements. My rule of thumb is that a prototype costs about 10 times as much to build as a production item. It doesn't take too much squint to imagine that a comprehensive development program costs 10 times as much as handbuilding the prototype. Which pretty much gets you to the price differential between an Airbus A380 and the ISS.
A bit of bad news is that building satellites is an established, mature industry, and they also cost on order of $100 million per ton dry mass. One reason is a launch cost to LEO of $10 million/ton, with most satellites going to higher orbits, a strong motivation to cut weight to the extreme. Satellites squeeze a lot of sophisticated gadgetry into a very small mass. They are equivalent to the most expensive parts of a large spacecraft, minus the (relatively) cheaper large tanks and structures.
So I'll ignore satellites and repeat my round number guess, that future production spacecraft might, with some good luck, cost the equivalent of $1 million per ton of dry mass, more or less what commercial jets cost today. (Allowing for inflation, big jets of the 1950s and 1960s cost somewhat less, but still on the order of $1 million/ton.)
I think it is highly optimistic, given that it is a 100 to 1 cost reduction over current space practice. A lot of people think it is very conservative. A million a ton - $1000 per kilogram, $450/lb - is pricey, and it makes really big spacecraft horribly expensive, like the $50 billion laser star I outlined for Ferrell.
Modern naval ships cost only a tenth as much, ton for ton (carriers even less). But naval ships use thousands of tons of cheap shipbuilding steel. Lightweight materials are used extensively in superstructures, but the hull, where most of the mass is, is built of good old steel, and everything on all decks is massively braced against the heave of the sea. Seagoing ships are very strong, but they are not lightweight.
There is a loose analogy here to laptop versus desktop computers; it costs more to pack power into a lightweight container. Spacecraft don't need to be compact, and crew cabins and propellant tanks are inherently bulky, but they do need to be as lightly built as practical. This naturally pushes up cost per mass: you're hanging the expensive parts on a much lighter (and thus more expensive) chassis.
And after all, the cost of jetliners hasn't kept jet travel from becoming pervasive. For the adventure minded, third-hand deep space ships will in time be available for far less than their original sticker price, perhaps $10 million for Serenity.
Progress in space has been and will be slow, because up front costs are enormous, tens of billions, even hundreds of billions over a couple of decades, say to establish the first base on Mars.
But it we do it at all, eventually someone will come up with the DC-3 of deep space, or more likely a combination of basic drive, support, and habitat pods that can be configured to a variety of missions and turned out in production runs of scores, eventually hundreds.
Models might be in production for 50 years and in service for a hundred years, still serving outposts that don't rate modern ships. When they are finally gone they'll be remembered as the ships that opened up space.
Related link: On my old website I wrote about (FTL) interstellar trade.