Kioxia's 74% operating margin is a shortage premium with an expiration date set by its own investor deck. AI inference storage demand is real, but management's supply-demand chart shows NAND returning toward balance by late 2027, and its long-term model targets mid-20% operating margins. The stock is about 12x the annualized peak quarter but roughly triple that on the company's normalized plan. The bull case: hyperscaler contracts and the inference-storage tier keep margins structurally higher. … Show more

Kioxia's need to do a buyback is proof that the flash-memory boom is over. x.com 

Kioxia is gaining NAND share as AI eats through storage, while Samsung and SK hynix shift resources to HBM. Staying focused on NAND has given it production costs 20-30% below competitors. Its quarterly operating profit hit $3.8B, passing Toyota as one of Japan's most profitable businesses, with about $15B of net income expected in 2027. Even after rising 2,000% in a year it trades at 8.9x forward earnings and is preparing a US listing. The key question is whether NAND pricing holds. x.com  x.com 

A supply response could compress Western Digital's pricing and margins, eroding its apparent cheapness. Its 2013-24 normalized earnings merit a look: even huge demand growth in other industries eventually drew supply and mediocre returns. A digestion/reset would hurt buy-and-hold investors, though the outcome is uncertain. Pre-AI economics were mediocre and unlikely to become amazing post-AI; TSMC's stronger pre-AI economics make its post-AI case better. x.com  x.com  x.com 

SK hynix's signed long-term cloud contracts give its AI-memory demand greater visibility and may support sustained pricing. Fourteen agreements cover over 50% of 2027-28 production capacity; HBM4 shipments are ramping, with the company expecting HBM4 to exceed HBM3E in volume and revenue by Q4 2026. AI-memory cash generation and Kioxia investment gains also underpin planned 2026 shareholder returns of about 100 trillion won, including dividends and buybacks. x.com  x.com 

Kioxia's P/E and growth are about as good as investing gets, so anyone worried about holding it here simply owns too much. Fear around Chinese AI models looks misplaced: they cost just as much to serve as US models, which points to an even bigger compute shortage. Short term, momentum has turned and selling across the compute trade is likely to continue, but the plan is to ride it out; there is some risk, as always. x.com 

Inflation, rising rates, Chinese competition and rising AI-related financial costs are pressuring the memory market. Micron, Samsung and SK hynix look exposed; waiting for the cycle to turn. x.com  x.com 

Semiconductors could stay undersupplied for a long time as more capable AI per dollar creates more uses and drives compute demand. Micron and NVIDIA trade at low forward P/Es because investors fear peak cyclical earnings, yet current earnings show no slowdown. Shortages may worsen as AI adoption spreads, though the industry could remain cyclical; a reversal of the forced rollover trade is hoped for, not assured. x.com  x.com  x.com 

Kioxia and SanDisk are the purer bets on NAND and enterprise SSDs, with more torque to QLC, KV-cache and offload storage, and Kioxia trades at a far lower multiple. Enterprise SSD is where the memory supercycle is being priced, as US cloud providers raised eSSD demand forecasts and QLC emerges as the cost-efficient growth driver, per a post citing TrendForce. x.com 

TSMC may lose pricing power as performance gains shift from shrinking transistors to stacking, interconnects and optics, while logic folding could make older processes competitive. Despite strong leading-edge demand, investors may overestimate the profit share its node leadership can retain. Its cautious expansion and past reluctance toward advanced packaging suggest a risk-averse culture that protects returns but could leave it late to disruptive technologies and key customers. x.com