The point is that they specialize in the theoretical side, not that they don't try to validate it. The difference is that their validation focuses on math rather than physical experimentation.
Eg there's no way to verify Hawking radiation in person (because well, good luck getting near a black hole to do so). But it's still supported by existing models of physics, thus it is science and not mere speculation.
Another recent example would be that the idea of time crystals was a theoretical physics result, backed by existing physics. That backing made it more than just speculation, leading to the investment of resources that eventually led to them actually being created.
Of course probably the most famous example of a theoretical physicst would be Einstein.
Lots of stuff typically goes through theoretical physicists before experimentalists determine how to test the idea. For example, all the particle accelerators, synchrotrons etc are initially designed by theoretical physicists via simulations etc to meet certain goals before putting in the billions of dollars it costs to actually translate those designs into something that can be built and used.
Okay, first of all, no, that's not the fundamental principle of science. That's the fundamental principle of science according to Karl Popper. Karl Popper is probably the single most influential philosopher of science, but his perspective is not the whole story. I know, I know, your high school science teacher told you otherwise, but I promise you there's more to philosophy of science than that. There's a reason they teach Kuhn and Feyerabend and so on.
But even if you ignore that, or don't believe me, or whatever...
You're conflating two different meanings of "speculative" here, which I think is probably whence your confusion stems.
On the one hand, you're (reasonably) saying that theoretical physics is speculative, in the sense that, once you finish your modelling work, your output could be said to be a speculation about what would happen in an experiment.
On the other hand, you're (again, reasonably!) describing certain proposed engineering projects as speculative, in the sense that if someone says "we'll have economically viable fusion within N years", that's a speculation.
My point is that those are just very different types of speculation, they really don't have much to do with one another. It's perfectly reasonable to be someone who wants to use theoretical (a.k.a. model-based, a.k.a. speculative) techniques to try to advance human knowledge about the nature of reality, while also being someone who is skeptical about the practicality of various (speculative) proposed exploitations of that knowledge. "I'd like to advance physics in areas P and Q; I don't think engineering is likely to advance in areas X, Y, and Z".
For example, Einstein was a theoretical physicist (very much not an experimentalist). He wasn't (initially) particularly bullish on the plausibility of doing engineering based on his work (and when it became clear that atomic bombs were gonna happen, he was horrified). For a later famous example of theoretical physics, Peter Higgs (and others who did the same work) were not concerned with whether or not it would be possible to design an experiment to validation the existence of the Higgs Boson, while still being vaguely hopeful that someone might one day figure out that experiment (using equipment invented decades after the proposal of the Higgs mechanism).
ALSO, btw, just because someone is a theoretical physicist doesn't mean they don't think that they're uninterested in experimental validation. It means they're not working on experimental validation. Maybe someone else will figure out how to do experiments, maybe not, it's just not the theorist's problem.
