Philosophically we don't _need_ anything of course :)
But fusion as a power source: abundant fuel source, cleanest (no emissions, minimal raw materials required, minimal waste to bury), safest (no proliferation risk, safest to install and operate, no need to climb towers), no intermittency problem.
All of those are desirable, but in no sense do any imply the necessity of fusion. They don't even imply the overall desirability of fusion, since one must also weigh the negatives, like cost.
The cost of development is amortized over infinity. (Sure, on the other hand the NPV of a perpetuity bond is finite.)
Of course (economically) the optimal allocation of current resources/funds probably would not involve fusion research, but instead instant (ASAP) replacement of coal, oil, gas power plants. Then working on fusion. But if we look at what fusion researchers should do, either they could be retrained or we can simply go ahead and print some money to pay them to work on what they want to do anyway (fusion research).
