Sure, and once you have those in place you have a startup that eats money and doesn't necessarily make any money. The vast majority of software startups end up building something that nobody wants anyway, because if there's something that lots of people want, somebody has already built it.

Whereas if you open a restaurant, you can make solid projections where "If we fill every table, we make $Y. If we're 1/3 full, we make $X. We're unlikely to be less than 1/3 full", and these are typically completely reasonable because you can see how other similar restaurants have done. For a restaurant, having a similar restaurant be successful is a very positive indicator. For a software company, having a similar software company be successful is an indicator that the market niche is already filled.

Thats a simplistic view of restaurants. From a purely market perspective, people already do this with personal loans, credit card or otherwise. I do agree that evaluating the final software's value is difficult but for a loan lender, it's only a matter of credibility rather than success.

As long as market actors don't do anti-competitive practices, I still don't see why a successful software can't be replicated and you can't compete in the same market niche. The user interfaces are one area which can obviously be different. Enterprise software is full of replicas.