This is already demonstrated by your own examples provided, just at a smaller scale.
If there were ever such a project to push against the capacity of our ability to do truly enormous, complex engineering, I’d say a massive, cutting-edge fusion reactor is as good a candidate as one could propose.
Moreover, the economic and educational stimulus these projects provide cannot be ignored when accounting for the indirect, long-tail returns this project, and those similar, provide.
Put another way: ostensibly, achieving net energy gain from fusion is the end of our near-term (energy) needs, conveniently breezing over the evolution and refinements of any system, as well as delivery and storage, but these are paths that are comparatively well mapped out. It then follows that, short of catastrophic losses prior to succeeding (which while not a given per se, seems more a function of time than of ability outright), any reasonable cost is worthwhile. Reasonable, in this context, meaning one that doesn’t bankrupt, or otherwise severely impact the participants in a negative fashion. Given the scale of these budgets vs. that of social welfare programs, military spending, etc. I don’t see that as an issue worth being concerned over. One can know the budget has been exceeded, without that also bringing down the house.
Ultimately, to an extent you’re asserting a false dichotomy. It can be true that there’s continued, substantial progress towards the stated goals of these projects, even if the budgets, horizons, and timelines aren’t to your taste. It can also be true that there’s waste, inefficiencies, and even (both legalised and otherwise) corruption. One does not preclude the other.
