A lot of the time in the industry, developers are using money loaned by publishers. Things like getting more development time, which means asking for more money is a negotiation that the devs aren’t guaranteed to win.

Valve is one of the successful developer & publisher companies that managed to survive. The 90s were a much smaller time for video games, and a small startup like Valve could compete with the big names out there. They had more freedom in a sense, but they also were taking quite a gamble. Other companies tried the same and didn’t survive.

It’s easy to simply say “only release a game when it’s 100% done” but it’s a lot harder when you’re watching money that keeps your company afloat dwindle with each delay. Also, “100% done” is a very flexible concept. Games almost always have to cut content or make concessions in some way, so figuring out what a done version looks like while working on it can be difficult.

The modern version of a small Valve style startup would be something like a Kickstarter funded development. Again, unless you are (for some reason) a Star Citizen dev, people are going to stop giving you money and you have limited funds and thus limited development time.

And just because you delay to try and release a superior game doesn’t mean it will be a smash hit.

baldur's gate did that and other companies were complaining about the high standard it set

Art is never finished, only abandoned.

Also it's fucking expensive to market things so people are aware you just released it. Or at least it used to be, before wish lists, early access, and so on.