I’ve seen companies that made poor technical decisions up front spend years paying for those mistakes. One could argue that those mistakes are what even put them in the position to where they could pay em down... but there is definitely a way to measure twice/cut once and save yourself from a world of hurt while still moving quickly and making sales.
Step 1-1000 of building a startup is selling.
But even with that being said, prototypes and v1s have a nasty habit of continuing along far past their originally intended expiration date, and if your only criteria is what gets sales today with no regard for future maintainability, you're taking on a very significant debt that is going to be hard to repay when you do need your tech to be easy to work with and performant.