The node size hardly explains everything. And I don't see people rushing to diminish Intel lead on the competition ever since the 90s primarily due to smaller node size. That is ofter all how Intel beat superior RISC chips from Sun, SGI and others in the 90s.
Funny how the tune sounds so different when Apple is doing to Intel, what Intel has been doing to its competitors for decades.
And node size is not purely down to the wallet. When a new manufacturing process has not yet matured sufficiently it is costly to put large chips on that node, because the failure rate will be too costly.
Apple makes smaller chips than Intel and AMD which makes them more suitable to be produced on 5nm. So there is a strategic choice here you are ignoring.
Basically Apple has made a number of strategic choices and prioritizations which allow them to be first on 5nm. By the time AMD's Zen4 is put on 5nm, the 3nm node size is actually already scheduled to be available.
Thus whatever catchup AMD does, may be very short lived.
And it is not like any of this matters. Apple isn't doing magic. They are able to do a series of strategic choices which are not easily availble to others. E.g. they can go into heteregenous computing much sooner as they control the software that will speak to specialized co-processor. Other vendors are sitting on the fence waiting for standarizations to happen before putting more co-processors on their SoCs.
Basically the whole development of the industry towards SoCs has simply played naturally into the advantages Apple has long enjoyed by doing full vertical integration.