Isn't total cost per reactor more constraining than cost per Wh for these mega projects? Say you spend $10B on a reactor that produces essentially free energy. You're still limited by plant lifetime, maintenance and crucially transportation of that energy – just breaking even would be a challenge in many locations. Thus, we have to have many plants distributed, similar to today's fission, at somewhere in the 500MW range per reactor (give or take an OOM).

In short, we have diminishing returns for giant reactors, and instead need to have plants that can be mass produced, fast.