I would say that every project chooses the standards that fit best the philosophy of their tools and the people that build it.
And that philosophy and the tools change over time.
Regarding your question, mass is no proof of quality. And I am not arguing about the quality of xml here, I think any decision can be justified.
Probably it was chosen and pushed forward because it was there at that time when it started to gain momentum.
Many alternatives were invented later or their qualities were understood later. In software design theory there is a rapid development of guiding principles.
Principles like maintainability, readability by humans, simplicity, don't repeat yourself, they become more and more clear, but some years ago they did not seem so important.
But xml is now baked in so many applications, brains of developers and managers and into their habits, so it will probably take much time until alternative standards can spread to a similar amount.
Maybe it can only be swapped out in the process of swapping out generations of applications and developers. There is no judgement in this, but you asked for it :)