The idea is more interesting when you think about scrapers. Take any ecommerce website, there are several scrapers that download all pages every hours, it would be more efficient if a provider had a live copy of the website and then serve the requests to the scrapers or could even send webhook.

A website could handle tons of scrappers without having high bandwidth, only the provider will need high bandwidth.

The issue is that scrapers often play with cookies and dynamic websites and such solution wouldn't work on these cases