I'm betting the amp is integrated into a multi function chip or codec where it was selected because of integrated features rather then matched to the required speaker power.
It's sort of like the software industry dynamic of having a QA department. In some places, you have QA and they catch the occasional bug. The engineers take that seriously, work hard to improve, and avoid having the same kind of bug again. But in other places with QA, QA is expected to catch the bugs, so engineers just throw stuff over the wall. In places without QA at all, most engineering groups learn to self-manage quality, because when they don't they feel the pain.
In effect, QA breaks the feedback loop between making a bug and organizational pain, meaning that the upstream organization loses its incentive to get and keep its shit together. This can happen in Hollywood, too, with "fix it in post"; people get sloppy.
So my concern is that yielding to the short-term economic imperative could lead to an undisciplined hardware organization. Then you get long-term economic harm because you waste a lot of resources (and worse, lengthen your cycle time) due to having to fix a bunch of stuff in software. And that lack of discipline eventually leads to quality issues that you can't fix in post, or don't fix in post because they're too busy chasing obvious bugs to catch the subtle ones.
But in the bad old days before we could do that, engineers and companies learned lessons when they made something that precluded the hardware from ever reaching the market. Some organizations are still going to learn those lessons. Many won't. I'd expect the latter category to still have significant catastrophic failures. It's just that they'll come later and be part of an amorphous "normalization of deviance" culture, rather than a clear learnable lesson of "before we order 100k of these, make sure to X".