The CPUs aren’t the problem. It’s an architectural problem.
A “normal” computer contains a multitude of separate components which communicates with each other. This makes the design very flexible and customizable (CPU/RAM/GFX/IOs etc.pp.), but also slow (already just by the physical distance between the chips).
What Apple has done with the M1 is to put almost all of these normally separated components on a single chip. They treated flexibility for speed and less power consumption. With the M1, for example, the amount of RAM is fixed and cannot be extended, but the CPU which is on the same chip as the RAM can read and write to it with speeds not possible otherwise. The same holds true for the graphics subsystem and the io controllers. Everything’s fixed, but fast like Speedy Gonzales.
A company which main focus is to produce CPUs and only some of the other separate components simply cannot compete unless they start to produce also such SOCs.