Re: And the answer to the question is
>So even if you can get a future launcher which is better / similar for 1 million dollars, it will still cost about 10 billion for the whole telescope.
Not necessarily. (Speaking from Hubble era experience)
If a Shuttle launch is $1bn, and only one slot is available in 5-10years time then you need to be really sure the payload is going to work, so you test the living-Belgium out of it, for years and years. This costs a fortune and means you are limited to now obsolete but proven space-qualified technology.
If it is going on a manned mission - then all this 10x
If a launch is a $1M and is available every day, then you just use cheap off-the-shelf technology and figure you would rather have 50*$10M craft + $50M launch costs - than a single $500M payload