AMD's secret sauce is a smart cache using like a 1 or 2 layer NN or something.
We see NVIDIA's major insight long ago, to but generic compute cores on their chip (so as to no longer need to predict how many shader units, how many geometry units, etc, just have a pool of generic units)
So it follows that intel is doing something similar i.e. NN type units which may be used for ML, where the cache is not so important, or in your typical non-ML settings such as gaming or web browsing, these NN units will reduce cache misses at an incredible rate and thus (effectively) up memory i/o rates by 400-500%
