Think of it this way: how hard is it to estimate the time required for a software project? How successful are most such estimates? We don't know how long it wi

Think of it this way: how hard is it to estimate the time required for a software project? How successful are most such estimates? We don't know how long it will take, but at the same time there is visible progress occurring on many fronts: understanding plasma behavior and how to control it, better superconducting magnets, better control systems, solutions to the neutron embrittlement problem, etc. There's not going to be a single breakthrough... or rather the fusion breakthrough already occurred. We've already learned about fusion and how to trigger it. That happened in the early 20th century. Instead of a single breakthrough it will be continued progress on all fronts until at some point everything matures enough that someone manages to build a proof of concept reactor. At that point investment will flood into the space because it will have been sufficiently de-risked.