Well, anecdotally, I did. I built a minimum viable product as a 'side project'. I had money to run it for a year before I needed to monetize. It worked for me. It means I use 1/10th the number of machines/bandwidth of one of my competitors.

I agree though, in your circumstance it was probably the wrong approach.