The author fundamentally misunderstands how the industry functions. Money wise, the ARM acquisition for this amount of money makes no sense if they want to sav

The author fundamentally misunderstands how the industry functions. Money wise, the ARM acquisition for this amount of money makes no sense if they want to save on core, or architectural licenses. With the amount of money, and engineers Nvidia has, they can easily bootstrap a CPU ISA on their own. It was by far not their top priority. And in their function relevant to the "AI" trend as giant linear algebra crunchers, GPUs don't need much help from anything else. And in longterm, basically anything you do pales to logarithmic scaling of PPW/relative performance. A s..ttier, and cheaper chip few years down the line invariably beats current best. In practical terms, the winner in the GPU/CPU/WhateverPU is decided by who gets the better fab contract.