It's quite likely that we live in an infinite universe. Also, of anything we typically say is morally valuable, it's likely that the universe contains an infinite amount. The standard ΛCDM model, our current 'best guess' theory of cosmology, predicts that our universe will persist forever and contain infinitely many beings like ourselves who experience happiness, pleasure, high well being, etc. Indeed, it will continue to exist 'happily ever after'. Given this, and a few basic assumptions, the total moral value in the universe will be infinite, and our present actions cannot change that fact.
Suppose that we accept this physical view and, in addition, are standard consequentialists - that we hold the view that we ought to do whatever maximises the total amount of moral value in the world (whatever might constitute value). There is infinite total value no matter our actions, so all actions are equally preferable. In practice then, standard consequentialism permits everything - it makes the same practical judgements as nihilism.
But not all hope is lost - several promising solutions have been proposed. Perhaps we might restrict the domain of value, or perhaps we might discount remote value and reach a finite total. Perhaps we might even compare cumulative totals as we aggregate value, as is done by expansionism, the leading solution so far.
I argue that, unfortunately, none of these solutions work. They each run into another problem. Our universe is relativistic, and locational information changes with velocity. There exist scenarios in which these approaches make judgements which change with velocity. They hence entail a difficult trilemma, whereby we must accept either: a bizarre form of moral 'relativism'; quite radical incompleteness; or intransitive preferences, which expose us to hypothetical value pumps. None of these options make for a viable moral theory, so none of these proposed solutions are satisfactory. The problem remains.