Responding for the parent. They are basically making predictions with S curves. Technology often starts out as really expensive but after awhile gets super cheap, following what appears to be an exponential curve (bottom of the S), before it levels off (top of the S). Yeah, look more like an integral sign or sigmoid.

But with fusion the endless claims of "too cheap to meter" are because how much energy there is in a fusion reaction. [0] We know that fission produces a lot of energy (but is expensive) but fusion produces significantly more. It also doesn't have the radiation drawbacks and so it is expected to follow the S curve (fission did initially but things changed. This is part of why France has so much nuclear).

So if (big if) fusion does follow this S curve (which there are good reasons to expect it to) then it could provide a very cheap and sustainable energy source. Yes, it is a bet, but every technology is. We won't know until we spend significant time and money into researching it. But honestly, a few billion dollars isn't that crazy for the potential upsides. We've spent that money on far greater risks with lower payout. Despite what the OP said, the money for ITER does not require international collaboration. Any rich country could do it themselves.

[0] (Fission and fusion can yield energy graph) http://hyperphysics.phy-astr.gsu.edu/hbase/NucEne/nucbin.htm...