Yeah, I'm not buying that. If your goal is to have a cache of humanity that survives x-risk scenarios, a dependent Mars colony is by definition not MVP because it is not V. For the enormous cost of building a human habitat in a distant gravity well, you've succeeded in protecting exactly 0 people against 0% of x-risk scenarios.

That sounds like exactly the sort of overinvestment in unproductive technology that MVP theory is designed to prevent.