I think you have vastly underestimated the size of some projects. For example the Manhattan project. Remember that ITER is 10billion over 20 years, not each year. Fusion does not get that much in the bigger scheme of things.