Max Koh

HBM/DRAM, NAND, and HDD all benefit from AI data center growth, but the three AI storage routes have different elasticity, cycles, and risks. HBM/DRAM is closer to AI accelerators and server memory bottlenecks, NAND is driven by enterprise SSDs and hot data, while HDD supports low-cost mass capacity, nearline storage, and cloud archiving demand.

Compare how Shanghai Fudan, ASMPT, SMIC, and Hua Hong Semiconductor are connected to the AI memory supply chain. This article breaks down HBM, advanced packaging, NOR Flash, eNVM, wafer foundry services, and mature-node processes to help you judge which company is closer to the AI memory theme.

Is Hua Hong Semiconductor 1347.HK a memory-related stock? This article analyzes Hua Hong’s real connection with the memory supply chain from the perspectives of eNVM, sNVM, NOR Flash, MCU, mature process nodes, capacity utilization, and the semiconductor cycle.

Cloud provider capital expenditure is a key signal for the AI infrastructure cycle, but the total spending number is not enough. This guide explains how Microsoft, Amazon, Alphabet, and Meta’s CapEx can flow through to Micron, Western Digital, Seagate, and enterprise storage stocks, and what investors should track across orders, margins, inventory, cash flow, and trading costs.

Server DRAM price increases are being driven by AI inference demand, DDR5 RDIMM upgrades, early capacity locking by cloud providers, and memory vendors shifting capacity toward high-end products. Understanding these drivers can help you evaluate Micron, SK hynix, Samsung, and the broader AI server supply chain.

Seagate STX stock is gaining renewed attention because of AI data centers, cloud customer demand for nearline HDDs, the HAMR and Mozaic technology roadmap, and improving free cash flow. This article analyzes STX earnings, Nearline HDD supply and demand, shareholder returns, risks, and key factors investors should watch.

HBM is High Bandwidth Memory, a critical memory system placed close to AI GPUs and AI accelerators. Learn what HBM, HBM3E, and HBM4 are, how they differ from DDR and GDDR, and why large-model training, inference, and the AI chip supply chain all depend on high bandwidth memory.

DRAM is volatile working memory used when computers, servers, smartphones, and AI systems run programs. Learn the differences between DRAM, DDR5, LPDDR, and HBM, and understand the investment logic, pricing cycles, and risks behind server memory, PC memory, and mobile memory.

Micron MU and SanDisk SNDK both benefit from the AI storage cycle, but through different paths. MU is more exposed to DRAM, HBM, and server memory, while SNDK is more focused on NAND Flash, enterprise SSDs, and data center storage. Compare revenue mix, gross margin, cycle sensitivity, and risk indicators to understand which company benefits more from AI storage demand.