Because they are just going by whats trendy I suppose. If they wanted to build something serious that can actually scale then they probably shouldn't be using a dymanic language at all.
The odds of your product ever having to scale are vanishingly small. Until you prove your product will actually need to grow to web scales, you should use whatever language allows you to deliver a product for the lowest possible cost in the shortest possible time. That usually means dynamic languages.
If you don't end up needing to scale then your application is most likely a failure, I guess I just don't quite get this approach of "we'll build it as cheaply as possible and then re-work it if needed" because I'm looking at this from the perspective of having worked for companies that did this (before I was working there) and then needing to deal with cleaning up the mess since their hastily built app did end up needing to scale. Web apps, by default, usually have the intention of attracting a wide user base so I can't say I agree with this line of thinking