The Bright Data mentioned in the article, as well as other similarly malicious but even harder to identify parties, most certainly do fire up residential proxies, no matter what the site, no matter how useless or duplicate the data which they're trying to get. Not at first - they start by trying to get your content from cheap data center connections - but as soon as some kind of bot-mitigation appears, they move to residential proxies to try and evade that, with a first tier coming from "global south" residential proxies, and then scaling up to (presumably more expensive / less widely available) proxies from the USA (I've seen some from Europe, but very few in relative terms). Each tier also appears to have the option to run JavaScript.