Micron (MU) is more important than Nvidia! Citrini analyst says AI model inference performance relies more on memory than GPU

Odaily reported that in response to the community debate over "Micron VS Nvidia," Citrini Research analyst Jukan posted on X, stating, "MU (Micron) may not be Nvidia, but its future importance could surpass Nvidia. Inference is now directly linked to revenue, but improving inference performance is not achieved solely by increasing Nvidia GPUs. GPUs are often underutilized during inference due to memory bottlenecks. For inference, increasing memory has greater value. The investment return of inference ultimately depends on memory, not GPU. Therefore, why do people still focus on acquiring Micron through NVIDIA's framework? A more comprehensive approach should be considered. Inference is all about memory."