Peter Van Dijck’s Guide to Ease
has an interesting discussion that considers the relative merits of using Drupal for startup businesses. The assumption one must make is that this is instead of creating a solution from scratch. He general thesis is :

"… Drupal keeps evolving to solve a lot of different problems, it tries to be a swiss army knife. You’ll probably use 20% of Drupal, which means you have 80% cruft (which 20% can be different for everyone), and you’ll probably only have 20% of your needs addressed by Drupal, which means you’ll have to hack around the 80% cruft to get your 80% needs addressed. It’ll just keep frustrating you. .."

http://poorbuthappy.com/ease/archives/2006/12/09/3382/drupal-considered-...
I would agree but think it is still better than many alternatives.
I do think the point could be made by saying something like this: The start ups don’t fully understand that it can provide a basic framework but should not see it as more than that. If you start from scratch 80 percent of your time and energy will be spent creating the core system or the systems represented by Drupal Core. Its the other 20% that makes their start up a unique business and this 20% is 100% of what they should be focusing on.
Another way to look at this would be Drupal VS. Scratch. I contend that except in rare circumstances starting from scratch is a competitive disadvantage and can’t think of one inherent benefit
The big assumption here is that they want to use the functions and not creating a software distribution for actual resale. That analysis is slightly different but not completely.