Great UI has diminishing returns that, for most orgs, are not worth the talent and effort it takes to create them. For that reason investment in these frameworks, or new frameworks at all, is discouraged. Web based approaches have completely supplanted native || crossplatform frameworks because its good enough and established enough to diminish the many development liabilities that other frameworks pose.
But the point of cross-platform frameworks is for non-tech or non-mobile focused companies to quickly churn out LOB apps with whatever developers they have on hand. (There was even one for PHP - KikApp - which appears to be dead.) So in theory, ignoring great UI/UX in favor of quick development time would drive investment in and adoption of some of these frameworks.
Your point on web based approaches also doesn't seem to match reality yet as PWA are still the once and future destiny of mobile apps, much like \(currentYear) is the year of Linux on the desktop. afaik, it would seem like most mobile apps are still primarily native, and if crossplatform, possibly React Native or Xamarin, with some older ones using Cordova/PhoneGap/Titanium. Web based approaches are far from replacing native and cross-platform; most companies that support mobile web also have their own apps.
Not everyone is trying to display a widget, with data from a remote database. If you work with local/edge data, then having access to numpy/pandas/scikit-learn/pytorch can be quite handy
