Public services (or even non-profit private organizations) still need to manage their resources efficiently. If the output of the service (number of meals fed to people, kilowatt hours produced, houses built, lives saved, passenger kilometers transported) is much smaller on a per dollar basis than comparable services, without good reason, then that organization is mismanaging resources.

Plenty of public services can and do break even, or bring in more than they cost.

Money is, in many ways, fake. But its accounting represents something real, in the real resources involved: land, human labor, physical materials like concrete and steel and copper that need to be turned into very complex equipment that require highly trained workers to operate and maintain.

So when a project doesn't break even, it's worth asking whether the project is still worthwhile, whether the external benefits outweigh the localized costs. I don't think modern nuclear passes that test (even if I believe that already-constructed nuclear should be extended as long as possible and operated at as close to full capacity as possible, because those up-front costs are already sunk).