Not 10-20 years maybe, but 50 years? Pretty high chance.

We've been saying it's 50 years away for about 50 years...

https://www.engineering.com/story/why-is-fusion-power-is-alw...

15 years ago AI was pretty much a dead end.

At our stage of world development, mathematical and algorithmic breakthroughs happen at a much faster pace than physical/mechanical breakthroughs.

It's still not that impressive, more a result of computational power than anything fundamental.

Fusion milestones have been newly achieved in the past 10 years. You may be right that it'll remain "50 years away" for generations still, but it's not evident to me that this will be the case. We barely had fission energy stable 60 years ago, we now know a lot more about the requirements and limitations for fusion, and we're allegedly still saying 50 years? That's a lot more reliable, and note that n=2 here (a bit more than one set of fifty years having gone by, a second fifty being underway and it still seeming like that). If it were "always 10 years away" since the 50s, that'd be quite a different story.

We would still need to go from 0W to 2-3TW production once we get actual generation; that's a lot of work planet-wide.