I think with SpaceX we have the situation similar to computing - ie. whether to start a 20 year computation today on a $1B hardware or wait 5 years and complete

I think with SpaceX we have the situation similar to computing - ie. whether to start a 20 year computation today on a $1B hardware or wait 5 years and complete it in 5 years then on a $100K hardware - to proceed to fund the $450M+ for the many years ahead or just wait a couple years until SpaceX gets to it and has it done for $10M in half a year.