I don't think that rules out success. I have code in production at work and I'm not even a dev. I plotted the structure and vibed it out 1 function at a time. Full build from nothing. It gave me the balls to go nuts trying to build things for myself. Not a server guy but I vibed a proxmox server into existence for my self hosting, trying hard to infrastructure as code the whole thing so I can add new services and fix issues easily. I've also got an arduino portenta on my desk slowly coding up a home automation project, brand new to C++ and learning as I go function by function. I've got a lot to add but what I built so far took an afternoon instead of a month and it works.
I assume major players are seeing similar internal adoption styles. No one is shitting out a brand new piece of commercial software in a brand new market overnight. Individual PRs, bug fixes, and new features for existing big players are being built by LLMs though. Also lots of tiny productivity hack tools.
The major players and the startups have access to the same toys here so it makes sense to not see massive market disruption.