I feel like the trick must be if you can amortize some of the maintenance costs across multiple projects.

Maintaining DNS records and SSL keys is not 3x as hard for 3 domains as for one. Probably neither are taxes. Or keeping on top of security fixes for projects using the same tech stacks.

There’s a couple groups out there who do something like this for public sector projects, sort of taking over the devops aspects and keeping the lights on. And in that case creating project generators could be feasible as well, because it’s easier to run a project that’s built the way you prefer them to be built.

Yeah and you can leverage learning across projects also. Find a marketing strategy that works well for one of the projects, then deploy to the rest. And one can share the cost of outsourced labor down the line. For example one customer support agent, software engineer, marketing person could work across all products/businesses, which might together fill near a full time role. It is often terribly hard to hire good part-time workers for many roles. And with a diverse portfolio the risk of high impact business failures should go down also. More tickets for potential buy-outs from bigger companies also, those can often be quite random, so more chances is better chances.

Not to mention you can leverage skills learned from earlier projects, greatly accelerating your progress. For example, if your project relies on content marketing, perhaps you learned to write quickly and produce enticing YouTube videos, and outsource effectively to remote writers (all skills that take dedicated time and effort to get good at). You can definitely leverage that and speed up your progress.

If I were to guess, the money is actually in writing about the experience and parlaying it to an entry into Angel or VC.

Putting in a lot of work to your own projects without answering to a boss or freelance clients wouldn't be too grim.