Western Digital and Seagate have the clearest long setups in storage because they combine higher margins, low competition risk and no concerns about capex or production expansion. That fits a regime where a sharply higher long end rewards free cash flow and punishes capex without an immediate payoff, after a Warsh press conference that raised doubts over Fed inflation control. The risk is that both keep getting dragged down by the broader memory sector, where DRAM and NAND look stretched. x.com 

Rising memory contract prices strengthen the case for memory stocks still far below their highs. JPMorgan forecasts Q4 DRAM prices up 13% and NAND up 18%; Micron is 28% below its 52-week high, Seagate 44%, Sandisk 47% and Western Digital 55%. x.com 

AI data and content creation could lift Seagate's and Western Digital's cross-cycle earnings even if supply catches up and pressures prices and margins. Western Digital reported hyperscaler demand signals through 2030/31, making 2013-2024 normalized earnings a poor baseline. Still not a huge fan of their stocks given their large premium to memory makers and Toshiba's production ramp. x.com  x.com 

Western Digital's margins are riding price increases that management sees lasting three to four more quarters. Price per terabyte rose 28% year on year while cost per terabyte falls less than 10%; December gross margin is guided to 60-61%. At about 10x fiscal 2028 consensus earnings, the stock prices in an early end to hikes despite scarce heads and media that WD mostly makes itself. Buy, $406. Long-term contracts cover less volume further out, leaving spot sales exposed if pricing turns. x.com  … Show more

Western Digital benefits from HDDs' low storage cost and growing cloud demand, with higher NAND prices adding pricing power. Nearline bit demand normally grows 20-25% with some deflation, while AI data growth now exceeds 30%; HDD costs ~$0.02-$0.03/GB versus ~$0.25 for NAND. Before AI, the consolidated, ~95% cloud-exposed market looked worth a low-to-mid-teens multiple of normalized EPS. Entry barriers are extremely high, but irrational competition among the three players is a risk. x.com  x.com  … Show more

Seagate and Western Digital trade at a huge multiple premium to other storage and memory, driven by vibes rather than cash flow, and Toshiba's expansion undermines the stable duopoly that justified it. Per @wallstengine, Toshiba plans about $380M to double HDD capacity for AI data centers by FY2027. Toshiba is a generation behind, with ramp issues, and is not a material near-term threat to profits, but the market will price the broken duopoly. The charts are awful. x.com  x.com  x.com  x.com  x.com 

Western Digital's hard drives capture AI-generated data, supporting stronger margins and guidance. Most data from models, agents and robots ends up on hard drives. Revenue rose 55% to $4.38B, gross margin hit 58.1%, and next-quarter revenue guidance rose to $5.0-5.2B versus $4.57B expected; EPS guidance midpoint is 21% above consensus. Before earnings, at ~15x earnings it looked priced as a cyclical, not an AI winner. Overweight, $613 target and 62% indicated upside in the pre-earnings report. … Show more

Western Digital may offer the best chance to get on board, because in a market this concentrated, duopoly economics don't care who gets there first. It just posted the best quarter in its history, though Seagate's HAMR drives are already in hyperscale racks while Western Digital's HAMR ramp starts in 2027. x.com 

The valuation gap favors SK hynix and Samsung over Corning, Seagate and Western Digital. Companies trading at 20x are still called cheap while the memory makers trade at 2x; Corning is not cheap on a $4 run rate even with growth, and neither are Seagate or Western Digital. The gap should narrow as memory rises while HDD stalls. x.com  x.com 

Samsung Electronics and SK hynix can invest heavily and still generate free cash flow that makes memory stocks look strikingly cheap. In 2027, they will spend $150B combined on capex yet generate $500-600B of free cash flow, against just $1.8T of combined enterprise value, or less than 3.5x 2027 FCF. Long memory. x.com