But its more nuanced than that. Ubunut (or rather Canonical) is not just an ad-supported service. It has "real" enterprise support customers, it has OEM engineering services, certification services and other things that puts Ubuntu as at least part of the product, if not front and centre. In a multivariate business model like this, the pulls in each direction tend to even out for the betterment of the product as a whole (or pull it apart completely, but with Ubuntu that hasn't happened yet).

I think the simplistic view of Canonicals business model is the problem with the article.