Like Oracle sells software to companies making software. It's called B2B.
The quality of those 'startups for startups' thingies is another thing, obviously some of them will fall more into a 'wannabe for wannabies' category, but it doesn't mean the model is broken.
When it comes to 'preaching to the choir', I feel like all those 'why you should start a company / learn to code' are exactly that.
A startup with no money is not in business. Yes they have limited funding, and they have to use that funding wisely to build their product and customer base. But if you sell them a tool that will clearly help them to do those things, they will spend money. Heroku is the canonical example of this. They extract a premium price from many small startups who value a hassle-free production environment more than a few hundred dollars a month they could save by hosting themselves.
I don't know about most of their money, but almost all of the companies listed on their (fairly extensive) success page are start-ups or small businesses.
To me, Heroku makes the most sense when you are building a business that may scale very quickly, and especially one that does not have a full-time sysadmin. It also makes sense for web apps that will stay very small and unlikely to grow that much past the free tier, but those are pretty much a loss-leader for Heroku.
The quality of those 'startups for startups' thingies is another thing, obviously some of them will fall more into a 'wannabe for wannabies' category, but it doesn't mean the model is broken.
When it comes to 'preaching to the choir', I feel like all those 'why you should start a company / learn to code' are exactly that.