On the web there is a lot of confusion about "MVC" in that the heart of it is not the model, it is rather having the controller (router) which is able to look at the URL, form parameters and such and decide which view to show you.

This is in contrast to the 1990s model of web programming where you wrote an HTML page with a <form>, pointed the action to some URL, and that URL was a cgi-script that couldn't redraw the form so error handling was difficult.

In a lot of cases you could say the data fetching is a dependency of the view and not the other way around, for instance if it is a blog post you might have a model object for the actual blog post but then want to put arbitrary widgets into the view which in turn requires fetching whatever model objects are necessary to draw the widgets. From the viewpoint of a CMS user, for instance, they want to drop the widget into the template and have it "just work" (have the framework figure out the fetching.)

The first exposure a lot of people had to this paradigm was Ruby-on-Rails and since it had a rich model system people thought the model system was the important bit but I'd say the router is the most important bit and how you fetch the data and format it is secondary, in fact it's totally fair to use different fetching and templating paradigms for different pages that live under the same router.