I think SpinLaunch is great example. They were actually able to fund and make a working prototype, tested it, but still decided it is not practical after all, before pivoting to production of knockoff Starlink-like satellites.