The core point makes sense, although I don't think there is a clear line between 'rule' and 'exception'; instead, the influence of companies probably follows so

The core point makes sense, although I don't think there is a clear line between 'rule' and 'exception'; instead, the influence of companies probably follows something like a power law on metrics like 'how many people recognise the brand' - there are a few heavy hitters like Facebook, and a long tail of much less recognised companies. I don't agree with their analysis that there is no bubble in Facebook's valuation because Facebook has revenue; you can't look at one company in isolation and say 'they aren't part of a bubble', you need to look at the big picture. Facebook has value because it has a lot of users and information on those users, and can use the information to target ads at those users. As pg has pointed out before, the 2000 dot-com bubble also had players like Yahoo bringing in high advertising revenues. Advertising revenues are much lower now compared to then on a CPC or CPM basis, but, especially for some targeted users / keywords, are still very high, and companies using them probably make a loss paid for from capital unless they have very high LPVs for customers, or from advertising revenue. In other words, advertising revenue is largely hot air circulating inside the bubble, and VC capital and cross-subsidisation from other areas pays to inject into the bubble. Some companies operate entirely in the bubble - they receive large amounts of bubble advertising revenue and reinject into the bubble by spending it on advertising, to cannibalise someone else's bubble revenue. Bubbles aren't stable - as soon as investors lose confidence in the market, investment will slow down, so there will be less advertising spend, which will reduce both direct and indirect advertising revenue and reduce confidence further.