I think Ryzen proves this isn't true. AMD can win marketshare, but they have to push a similar product (performance, features, quality) and sell at a lower price. AMD we're nearly dead when Ryzen was released, lower prices attracted share, 2nd and 3rd gens just got better and better. Prices rose and they now dominate market share.
Performance is doable, features are difficult, quality is good IMO.
Once they have market, they can raise prices. But there is no doubt, catching up with features is going to be tricky with the immense resource NV have.
In terms of home consumers, CPUs offer two types of performance: desktop applications and multithread workloads, secondly, gaming.
GPUs also offer two types of performance for home users: rasterisation and Ray/Path tracing
The reason Ryzen succeeded is because it was able to beat Intel in at least one of these performance types while doing it at a lower price at every tier. AMD does not have a complete GPU stack that does that, when they do, they will catch up to Nvidia quickly
Ryzen 1000, at every tier in the stack, beat Intel in multithread performance by a large chunk and did so at a cheaper price, that's why it worked. It didn't beat Intel in gaming but it didn't matter you only have to beat the competition in one thing to get your niche.
AMD doesn't have that with RDNA: There is no RDNA generation so far that easily beats Nvidia in Rasterisation or RT at every tier in the stack and does it at a cheaper price. Thats why RDNA desktop gpu is a commercial failure if we're being honest