Western Digital's Long-Term Contracts Booked Through 2031 as AI Drives Structural Growth in HDD Demand
Claire Weston
Western Digital's CEO disclosed that customers are negotiating long-term purchase agreements stretching to 2031; full-year revenue grew 36% to $12.9 billion as AI inference and physical AI create a dual engine turning HDD demand from cyclical to structural.
Orders booked through 2031 — what changed in the HDD industry?
CEO Irving Tan revealed the company is negotiating long-term agreements (LTAs) for calendar years 2029 through 2031. Customer willingness to extend "remains very strong."
This is one of the longest forward-commitment windows ever publicly disclosed in the HDD industry. This means → large customers are not just restocking short-term; they are locking HDDs into multi-year infrastructure plans.
Cloud revenue now accounts for 89% of total revenue ($3.3 billion), up 43% year-over-year. In plain terms = Western Digital has transformed from a consumer-electronics HDD company into a cloud-infrastructure supplier.
How strong are the financials?
FY2026 full-year revenue grew 36% to $12.9 billion; EPS doubled. Q4 revenue hit $3.75 billion, up 44% year-over-year, at the high end of guidance.
Blended average selling price per TB expanded from high-single-digit growth last quarter to "high-teens" growth. This means → the gain is not just volume — each TB sells at a higher price. Both volume and pricing are rising.
Q1 FY2027 guidance: $4.1 billion, roughly 45% year-over-year growth. Free-cash-flow margin reached 34%, with $1.3 billion in quarterly free cash flow.
Why does AI keep generating HDD demand?
Tan offered a core thesis: "Compute cycles can be reused, but data compounds." In plain terms = processing power can be shared and recycled, but the data each training and inference run produces — inputs, outputs, logs — only accumulates and must be stored.
AI agents — programs that autonomously execute multi-step tasks — amplify the effect. Agents generate data at every step of a workflow, "increasing both the data created and the data that must be stored over time."
This reflects a structural shift: AI's storage demand is not a one-time purchase but a compounding growth curve tied to usage volume.
How much incremental demand does physical AI add?
Tan cited an autonomous-vehicle customer whose EB-scale demand "grew several-fold" within this calendar year.
The driver: autonomous driving, humanoid robots, and similar systems lack sufficient real-world training data. They must generate and store massive synthetic datasets — training data simulated by algorithms.
This means → physical AI does not just consume compute; it manufactures vast volumes of "data that did not previously exist," creating another incremental demand source for HDDs.
Tight supply plus restrained expansion — how long can pricing hold?
In a very tight supply environment, the LTA model locks in future EB-scale demand (expected growth of 25%+) and gives the company "an opportunity to add pricing leverage."
Roughly 80% of hyperscale data-center data sits on HDDs, a structure that "will very likely persist." This reflects a counter-narrative: in the AI era, HDDs are not being displaced by SSDs — cold-data explosion is reinforcing their role.
The CFO stated explicitly that meeting demand growth "does not require incremental capacity spending." Investment goes to heads, media operations, and automation to raise productivity. Put simply = no new fabs, no new lines — technology upgrades absorb demand, and supply discipline is the source of pricing power.
Can the product roadmap deliver?
40TB EPMR drives began shipping in the June quarter and are expected to reach roughly 50% of nearline shipments by Q3 FY2027.
44TB HAMR drives — heat-assisted magnetic recording, using a laser to heat the disk surface for denser data writing — are planned for H1 calendar 2027; 50TB drives are planned for H2.
Management committed to reducing cost per TB by "approximately 10% year-over-year" over the medium term. This means → whether high-capacity products ship on schedule is the key milestone for validating whether this cycle's pricing and profitability can be sustained.
Content is for reference only, not financial advice.