> So both are not very good examples, because they build up experience over long periods. They are examples of something they could similarly do for the Apple

> So both are not very good examples, because they build up experience over long periods. They are examples of something they could similarly do for the Apple Neural Engine but in a bigger scale in the future. They have experience deploying it in a smaller scale/different versions, they would just have to apply it in bigger scale in order to be able to compete with NVIDIA.