AMD is gaining ground against Intel in datacenter CPUs but is not taking market share from Nvidia in datacenter GPUs, according to the article. AMD’s Instinct GPU output is constrained because it cannot get more HBM allocations from Samsung, SK Hynix, and Micron, and AMD is selling even fewer GPUs than it otherwise might because it is loading up the HBM memory on the Instinct MI400 series a lot more than Nvidia does with its “Blackwell” B200s and B300s and its “Rubin” R200s to get a more fair balance of compute and memory capacity. The article argues that the Instinct MI500 series should have optical interposers and put disaggregated and composable HBM memory on trays adjacent to and interconnected with GPU accelerators, and expresses the same wish for the future “Florence” Zen7 series of Epyc datacenter CPUs. AMD is taking share from Intel in datacenter CPUs because it has designed a superior X86 product and has taken care of its process nodes and supply chain of packaging, allowing it to grow X86 shipments faster than Intel at a higher average selling price. The article says third quarter market share numbers are not yet available but that AMD has a little more than a third of X86 server CPU shipments and a little less than half of revenues for the first half of 2026. The gap between Nvidia and AMD is growing, not shrinking