Trigger: @demian_ai's 2026-05-29 Dell-upstream tweet → Dell-upstream investigation → a focused nearline-HDD chase the user greenlit. Demand claim ("nearline allocated through 2027") was corroborated by primary sources (Seagate/WDC Q3 FY26 calls + TrendForce); two parallel research agents mapped demand + supply chain; valuation pull (Massive TTM + yfinance forward P/E) sized durable-vs-bubble.

Why a separate perspective (not a memory-supercycle bullet): HDD nearline is density-led, not cycle-led, and the oligopoly explicitly refuses to add capacity — so unlike memory, supply does not flood and there is no 2027 capacity cliff breaking the thesis. Different mechanics → own frame. Captive heads/media/HAMR means STX/WDC themselves are the purest plays.

Set-up state: STX/WDC blown off (RSI 80, +136–165% 3m); the durable upstream pick Hoya (HOCPY) is the laggard (RSI 48, −3.6% 3m, −9.9% off high). Forward P/E flags memory (MU/SNDK ~10) as a peak-cycle trap, WFE oligopoly (AMAT ~31) as the trustworthy durable multiple.

Coverage built: nearline-storage watchlist (STX, WDC, HOCPY, SHWDF, NJDCY, 5991.T); security-families added to _securities.json (hoya, nhk-spring [JP primary — NHKSY ADR dead], resonac, nidec).

Status: DRAFT, not in ACTIVE.json. Awaiting promotion decision: (a) promote to active perspective, (b) fold into memory-supercycle as a storage split, or (c) hold. Filed for review in TASKS-PERSPECTIVES.md.

Open / watch: Q4 FY26 prints (~Aug) for pricing/lead-time/discipline; HAMR ramp milestones.