Get rich (selling software) or die tryin' (anything else)

It seems to me there are two dominant startup models in tech today.

1. You can make money selling software.

2. You can lose money selling anything else.

- @DKThomp

blue = software, other colors = not "pure" software

I found this tweet and chart while reading about Casper last week. It’s a thought-provoking take on successful startups.

Basically, the author says that even though many startups are struggling in the public markets, all of the “real” technology companies (i.e., companies that sell software and have near-zero marginal costs) are doing well. The hardware + software, hardware-only, and marketplace businesses are the ones that are struggling.

Thoughts on this rule? What are the best exceptions, and is this trend going to change over the next five years?