BlockBeats News, June 28th, Critini Research analyst Jukan refuted the bearish view on Micron's future. Jukan stated that although Micron is not Nvidia, in the future it may be even more important than Nvidia. Currently, Inference (AI Inference) is directly linked to profitability, but the problem is that simply stacking Nvidia GPUs cannot improve the inference performance because GPUs are often idle due to memory bottlenecks. During the AI inference phase, adding memory is more valuable than adding GPUs, and the ultimate return on investment for inference depends more on memory than on GPUs. Therefore, everyone should not use the Nvidia framework to evaluate Micron, but should focus on a larger fact — Inference is memory.

Jukan believes that in the later stages of the AI era, memory, especially high-bandwidth memory (HBM), will become a more critical bottleneck and source of value than GPUs, so Micron's strategic position may exceed what everyone imagines at present.