Podcast Atlas
concept Updated 2026-08-07 Topics: Technology, Economics

Storage Industry Cyclicality

Meta and Microsoft report different AI earnings adds a global market-volatility update through SK Hynix and Samsung. Anita Ramaswamy says bullish investors argue AI memory demand could exceed supply for several years, but the market remains skeptical because memory suppliers have historically overbuilt after demand surges, creating later supply gluts.

Storage Industry Cyclicality is the boom-bust structure described in 存储三巨头破万亿市值，存储超级周期何时能见顶？｜ S10E13. The source says memory products are standardized, fabs require heavy upfront capital, and suppliers have strong incentives to run capacity hard once it exists, even when prices weaken.

AI is eating up the world’s computing memory adds Tom Minelli’s public-market explainer version. He says the memory industry has always cycled between overcapacity and shortage, but AI data centers have added a new demand source, while capital-intensive manufacturing means relief cannot arrive quickly from new entrants.

The episode’s cycle model is straightforward: demand rises, spot prices jump, customers stockpile, suppliers expand, supply arrives late, demand cools or customers destock, and prices fall sharply. AI changes this cycle by creating more structural demand and stronger Memory Capacity Lock-In, but the source explicitly says it only weakens cyclicality rather than abolishing it.

Key Claims

  • Capital intensity and delayed capacity make memory markets prone to overshooting.
  • Standardized products make supplier competition more price-sensitive when supply catches up.
  • Customer inventory behavior amplifies both shortages and downturns.
  • Supplier concentration and capital discipline can reduce destructive expansion, but cannot remove the lag between capex decisions and eventual supply.
  • AI data-center demand can extend the shortage phase of the cycle even when the underlying memory market remains cyclical.

Connections