AI Data Explosion: Can WDC Become a Bigger Storage Winner?

Western Digital is positioned as a long-term beneficiary of AI-driven data-storage demand, with roughly 80% of hyperscale data-center data still stored on HDDs. The company began shipping 40TB ePMR drives in June, expects 44TB HAMR products in 1H 2027, and targets UltraSMR for about 60% of nearline exabyte shipments by fiscal 2027. Fiscal 2027 and 2028 EPS consensus estimates have risen 7.5% to $20.03 and 7.6% to $34.74, respectively, although WDC trades at 20.86x forward earnings versus 10.67x for the industry after a 394.5% one-year share gain.

Analysis

The investable issue is not aggregate data growth but whether the HDD duopoly can sustain disciplined supply while hyperscalers qualify higher-capacity drives. WDC and STX have unusually high operating leverage: a modest improvement in nearline pricing and factory utilization can drive earnings materially faster than exabyte growth. The second-order beneficiary is STX, whose HAMR transition could command a technology premium if it reaches volume qualification ahead of WDC; conversely, any qualification delay makes the capacity roadmap a cost burden rather than a margin catalyst.

WDC’s rerating leaves limited room for roadmap execution risk. The valuation premium implies investors are already capitalizing several years of mix improvement, while hyperscaler procurement remains lumpy and can shift between suppliers after qualification. Over the next 1-3 months, quarterly nearline pricing, customer qualification commentary, and gross-margin guidance matter more than AI-storage narratives; over 6-18 months, the key question is whether higher-capacity drives reduce customers’ total cost of ownership enough to preserve HDD share versus falling QLC NAND economics. NTAP is a less direct beneficiary: its upside depends on enterprise AI-storage budgets converting into broader infrastructure refreshes, which typically lag hyperscale spending.

Consensus appears to treat all AI data as incremental archival demand. Much inference data is disposable, compressed, deduplicated, or retained in lower-cost cloud tiers, so raw token growth need not translate proportionally into purchased HDD capacity. The thesis is falsified by sequential nearline exabyte weakness, renewed HDD price concessions, gross-margin guidance below expectations, or a meaningful decline in NAND $/TB that accelerates flash substitution for warm-data workloads.

Key Decisions for Investors

• Prefer a 3-6 month long STX / short WDC pair rather than outright HDD exposure. STX offers cleaner upside if HAMR qualification and nearline pricing remain tight, while WDC carries greater multiple-compression risk after its sharp rerating; target a 10-15% relative return, with a stop if WDC reports superior capacity qualification or STX cuts gross-margin guidance.
• Do not add WDC outright ahead of the next earnings print unless channel checks confirm hyperscaler volume commitments and pricing above prior-quarter levels. A long becomes attractive only if forward earnings estimates rise faster than the current valuation premium; otherwise roadmap milestones are largely narrative, not incremental cash flow.
• Maintain NTAP as a watch-item rather than an AI-storage substitute. Initiate only after management demonstrates acceleration in all-flash/Keystone bookings and raised recurring-revenue guidance; enterprise AI projects can be delayed by data-governance and application-readiness constraints even when GPU spending remains strong.
• Set alerts around NAND price declines and hyperscaler capex guidance. A sustained flash cost decline or a cloud-capex moderation would pressure HDD total-cost-of-ownership assumptions and should trigger profit-taking on any HDD longs within days, not quarters.