Recently, Samsung, SK Hynix, Micron, and SanDisk have all surged 🛫. The core long/short debate now is whether AI's explosive growth has transformed the cyclical nature of the storage industry into secular growth.

I had GPT organize my perspective on the storage sector, which is: AI hasn't eliminated cycles, it's just transformed storage stocks from "pure cyclical stocks" into "tiered cyclical stocks."

Among them, HBM/AI Server DRAM behaves more like structural growth assets, because it's directly tied to AI compute expansion and faces tighter supply constraints; but regular DRAM/NAND remain highly exposed to smartphones, PCs, consumer electronics, and price volatility—essentially still cyclical assets.

From an investment ranking perspective, I'd prioritize: SK Hynix > MU > Samsung > SNDK.

SK Hynix is the HBM leader with the most certain AI upside; MU is the most "AI-transformed memory stock"; Samsung looks more like a catch-up recovery play with upside elasticity, but with greater headwinds from legacy businesses; SNDK's earnings this cycle are strong, but it's more exposed to NAND cyclicality beta rather than pure AI storage exposure.

The conclusion: medium-term bullish, but I reject the long-term narrative that "storage is no longer cyclical." AI has shifted cycles upward, extended them, and created divergence, but hasn't eliminated cycles entirely. SK Hynix leads in HBM share, Micron is expanding advanced DRAM/HBM capacity, Samsung is promoting HBM4 and emphasizing sustained strong chip demand driven by AI, while SanDisk benefits more from AI-driven data storage demand and NAND strength.