At least on the Mac, there‘s much better system framework support for annotation of PDFs, which may partly explain why PDF annotation is much more common among tools.
FWIW, I have long-standing plans to implement annotation of web archives. But the required tool chains aren’t as straightforward as with PDF annotation. And since the latter is still the current status quo, I have to focus on that first.
Both formats have their advantages & disadvantages, and PDFs have their use cases, so I don‘t see them as mutually exclusive. That said, I am really looking forward to working with a text+markup-based format such as HTML, the ability to easily read & modify the DOM, to parse or embed additional metadata, and to facilitate better semantic relationships.