The explosive growth of AI is causing an unexpected supply shock in the market for traditional Hard Disk Drives (HDDs).
This phenomenon is primarily fueled by the massive capital expenditures, or CAPEX, from hyperscale data center operators like Google, Microsoft, and Meta. These companies are investing record amounts—potentially exceeding $600 billion in 2026 according to TrendForce—to build the infrastructure needed for AI. While cutting-edge GPUs get the spotlight, this expansion also requires an enormous amount of storage, and that's where the humble HDD makes its comeback.
The reason lies in the data itself. AI models generate vast quantities of 'cold data'—information like training logs and model checkpoints that doesn't need to be accessed instantly. For this purpose, HDDs remain far more economical than faster Solid-State Drives (SSDs). Projections suggest that even in 2028, the cost per terabyte for HDDs will be about one-fifth that of SSDs. This cost advantage makes them the default choice for bulk data storage, which still accounts for over 80% of data center capacity.
This surge in demand has collided with a period of limited supply. The causal chain began in late 2025, with industry reports warning of lead times for high-capacity drives stretching to over 52 weeks. By early 2026, the situation came to a head when Western Digital announced in an earnings call that its entire 2026 supply of nearline HDDs was already sold out to its top cloud clients, with long-term agreements extending into 2027 and 2028.
Finally, the market structure acts as a price multiplier. The HDD industry is a classic oligopoly, with Western Digital, Seagate, and Toshiba controlling nearly the entire market. With few competitors, these manufacturers have significant pricing power. When demand outstrips supply so dramatically, they can pass on price increases with little resistance, a dynamic now clearly visible in the market.
- Nearline HDD: A type of high-capacity hard drive designed for enterprise data centers where data is accessed less frequently but must remain readily available. It balances cost, capacity, and performance for 'warm' or 'cold' data storage.
- CAPEX (Capital Expenditure): Funds used by a company to acquire, upgrade, and maintain physical assets such as property, buildings, and equipment, like servers and data centers.
- Cold Data: Information that is infrequently accessed by an organization and does not require real-time access. It is typically stored on lower-cost, higher-capacity storage media like HDDs or magnetic tapes.