Higher clocks means more heat is produced when running the chip, which means tighter parameter ranges for what you target when fabricating it. That results in more chips not hitting those parameter targets and thus not being usable in XSS machines, hence more cost. You have to balance that out against the reality that almost nobody who will ever buy an XSS will notice or care about the downsides of not having that extra grunt.
Also, proper next gen game engines are gonna run just fine on XSS anyhow. During the cross gen phase where new consoles are just brute forcing everything and not leaning into any new gen features you won’t see XSS keep up as well, but with actual next gen engines it will likely look just like XSX version at 1080p instead of 4k (yes, devs will dial back draw distance and texture res and and AA but only bc ya can’t resolve all that detail on a 1080p screen anyhow so you can dial that stuff down w/o it being noticeable).
The issue of optimization mostly stems from trying to find ways to brute force old gen titles into new gen expectations but with next gen engines that won’t be much of an issue.