This, indeed, is where things get very dicey. Meritocracy tends to converge on winner-takes-all markets, which result in an extreme inequality of outcomes and all the socioeconomic harms that entails.
Physical limits and other regulations even things out but, particularly in computer and digital products and services, physical limitations are less and less important.
Could you explain this claim?
Presuming that A doesn't do something to demonstrate they don't deserve merit any longer, why would you ever go with a proven inferior (B) option, or an option that is unproven altogether (C)? Hence, a trend toward a single dominating entity.
In reality, I think things are more nuanced, of course. No one awards purely on merit, as you also have timing, capacity, and costs to balance. I'm not really aware of any market that you could consider a -true- meritocracy to the extent of the GP comment. And maybe I misread the comment altogether, but that is my understanding.