Comparing Energy Technologies in terms of: Cost per Power Capacity  and Research Spendings Cost per Power Capacity
For a number of energy technologies I'd like to know what the minimum costs are to install a given power capacity.  Are there any such comparisons available?
Research Spendings
Technologies often own much to publicly funded research efforts so I'm also interested in estimations regarding development cost of the technologies themselves.  Of course foundations to a given technology may have been readily available, so let's limit this to spendings on larger research projects dedicated to developing a technology and motivated by the need of energy.  I'm aware it's unlikely any exhaustive comparison in full detail exists but maybe something related?
(Maybe someone should tag this question 'energy-economics' -- I'm lacking the reputation)
 A: I have found this article in Wikipedia on relative energy costs. It contains a lot of data and some discussion.
To summarise parts of it we note that the official UN bodies have developed a formula called "Levelised Energy Cost" (LEC) which is a lifetime accounting formula which includes the following cost factors:


*

*Investment (per year) - the totals are
over the n years of the system's
life.

*Maintenance and Operations (per year)

*Fuel Use (per year) [minimal for some
sources like Nuclear, obviously high 
for anything Coal/oil based.]
With the benefit factor being the Electricity generation itself (per year).
As the other Answers have reminded us, with Nuclear in particular we need to be careful about including the "hidden subsidy" and dual use (civilian/military) of these technologies, so maybe one should use the Nuclear columns in the data as a reference point, and investigate it in more detail if required. For the less military fuels the data may be more accurate (at least from that perspective).
So the data tables in these lists give the LEC values, for the different energy types as estimated by various national bodies: US, UK , Australia.
Here is the US DOE Table
Here is a recent (commercial) study for the UK:-
Technology --------------------------------------                                 Cost range (£/MWh)


*

*New nuclear --------------------------------                                   80-105

*Onshore wind ------------------------------                                        80-110

*Biomass -------------------------------------                                  60-120

*Natural gas turbines with CO2 capture---                                           60-130

*Coal with CO2 capture --------------------                                    100-155

*Solar farms        -------------------------------                                125-180

*Offshore wind    -----------------------------                                150-210

*Natural gas turbine, no CO2 capture------                                          55-110

*Tidal power     -------------------------------                                   155-390


In all these data the "research costs" have not been separated out, but seem to be included in the "investment costs" associated with the given estimates. As the articles also point out there are other costs associated with each type in addition to those considered in this question including any transmission and connection costs which any "geographically remote" technologies might incur.
A few extra points about reading the data from Commercial organisations. If company A has expertise in technology a then obviously their assessment (no matter how genuine) of the costs of technologies a, b, c will price the R&D and other costs for b and c much higher than a. (An oil company estimating solar versus oil costs, might be an example.) Whereas a specialist company B in technology b (which might be much smaller than A if dealing with newer technologies) will have a lower cost figure for b (versus anything else for them, and lower than A's estimate of b too if their product is good.) Of such are commercial and political arguments made.
A: The question is complicated and related to something called ERoEI (Energy Return on Energy Investments).  This is the general question on how much energy one must put into getting a source of energy balanced with how much energy you get out.  The direct case can be seen with fusion and the break even point where you get net energy out. 
As Georg indicates the nuclear power industry has been hugely subsidized in connection with nuclear weaponry.  The industry is largely a spin off from plutonium production for weaponry.  The stable Pu isotope for weapons is Pu-239, while the other isotopes tend to spontaneously fission.  These isotopes are farmed out to the nuclear power industry.  So the nuclear power industry is a spin off from the weapons industry.  The net financial investment by the United States in nuclear technology is over $10 trillion.  We might of course question whether we have recovered this cost, either in financial terms or energy terms.  Nuclear accidents are also seriously costly and dangerous, and the industry has some appearances of being like the dirigible industry before 1936.  At some point there may be the big nuclear “Hindenburg” incident which might put the kibosh on the whole thing.  Events of late come somewhat close to that.
The industry, in part due to its connections with weaponry, has a huge political connection.  Unfortunately this has been politics of the far ends of the spectrum, where in the USSR this was with the extreme left communists and in the US this has been associated with the far right that is moving further to the right.  The further on either spectrum politics becomes the more it degenerates into a filthy pile of lies.  Unfortuantely the economics and politics surrounding nuclear power, mainly with those who run the system, have a long history of being surrounded and mired in mendacities.
The issue with research costs is of course important.  The fusion research programs, such as ITER, seem to be a long process of getting closer to a goal, but where the goal also seems itself to recede further away.  Renewable energy research tends to get at best second billing, for it is not as “sexy” as “big science” programs with impressive expensive facilities.  The funding and political interest is highly skewed in some ways.  For instance there is the whole thing about “clean coal,” which is a way of subsidizing the future of the coal industry.  However, clean coal makes about as much sense as saying “partially pregnant.”  In spite of what the mendacious right wing in the US says, we do have a little problem with $CO_2$ in the atmosphere and the climate perturbation it is generating.  
The ERoEI issues then comes into play, where there are people who spend considerable time on this problem.  As petroleum becomes less accessible and approaches this peak situation the ERoEI declines.  This is a ratio of energy out/energy in, and for a large oil field that is “virgin” the ERoEI is about 25.  Currently ERoEI is estimated for oil production globally to be around 10 to 15.  Nuclear power is estimated to be a number of values, where about the highest I have seen is 5.  The solar, wind, and renewable power systems have a similar ERoEI.
