I think you are probably right, it is not a make or break factor. Especially when it comes to back-end flows, PPA/TAT/capacity/design closure are the key metric

I think you are probably right, it is not a make or break factor. Especially when it comes to back-end flows, PPA/TAT/capacity/design closure are the key metrics and the big 2.5 have large moats in these areas. On the front-end side though, areas like verification, debug, and various lint checks, I think there is more opportunity for smaller players to introduce point tools with better customizability & scripting interfaces. Ultimately I think the industry is stuck in a local maximum, where the frictional costs of rethinking overall methodology, e.g. from SystemVerilog to Chisel, is too high to justify the change, even if the end result would be marginally (or greatly) better. (Not an endorsement of Chisel, just an example.)