Skills don't go from unknown to crowded overnight. They move through fairly predictable stages, and knowing which stage something is in tells you a lot more than "is this in demand" ever will.
Early on, the people doing the work are mostly self-taught, pricing varies wildly because there's no shared benchmark, and there's no single agreed-upon job title. This is the highest opportunity, highest uncertainty stage. Nobody is teaching it yet because nobody has fully figured it out yet either.
A dedicated tool, certification, or framework appears. People start using the same handful of terms to describe the work. Agencies begin offering it as an add-on service rather than a specialty. This is usually still a strong window, the skill is provable and nameable now, but the field hasn't filled up yet.
The skill shows up as a line item on job postings, bootcamps and courses multiply, and pricing compresses toward a stable market rate. This is the mainstream stage. It's still a perfectly good skill to have, but it's no longer a differentiator on its own, it's now table stakes.
The practical takeaway: the biggest gap between effort and reward tends to sit in stage one and early stage two, not stage three. That doesn't mean skip stage three skills entirely, it means go in knowing you're competing on execution and speed rather than novelty.