The truth is that many of these companies can be run with fewer and fewer people as abstractions in technology become higher-level. Software takes a lot of effort to write, and much less to maintain.

We just went through an era of effectively no software, to one with a large saturation of software that fills many niches. In the 2010's there was no competition for the majority of these businesses as VC capital tended to move towards new ideas, rather than directly competing with existing ventures.

All of that is changing. The world will continue to need new and improving software, but the value provided by a single engineer continues to trend higher, and fewer and fewer developers will be needed to create these systems.

Some SaaS has high costs of switching and a defensible moat, but many do not and are easily replicated. Existing players may even be at a disadvantage to competitors due to existing architectures using old patterns that require more manual labor/cost to maintain. e.g. Something like DocuSign is a good example of a low moat SaaS. Database systems are a good example of large moat (high cost of switching, even if better tech comes out)