Western Digital Corporation

▲ Bull Case

• ◆AI nearline demand is structurally durable, not cyclical. GPU-cluster pairings drive substantial HDD demand; global GPU deployments are forecast to compound 4–5x by FY2030, with resulting HDD demand floor structurally above pre-AI levels and 'capacity committed through 2027' commentary extending rolling-forward.
• ◆Duopoly pricing discipline holds through the FY2027–2028 supply-demand re-balancing. Both WDC and Seagate have publicly committed to margin over share; gross margins above 45% become a new normal (vs. pre-spin sub-30%), supporting a multi-year sustained 22–28x forward EV/EBITDA peer-parity multiple.
• ◆HAMR/EPMR roadmap execution plus buyback resumption catalyze a re-rating leg. WDC ships 40TB EPMR in volume by H2 FY2027 with 3 customers in qualification; HAMR follows; buybacks and dividend reinstatement resume by FY2028 alongside net leverage below 1x.

▼ Bear Case

• ◆Hyperscaler nearline order book digests in 2H FY2028 as the first wave of capacity catches up to the AI build. Inventory cycles in compute infrastructure pulse every 18–30 months; the 'committed through 2027' window has a clear end date. Revenue growth drops to mid-single digits; gross margin compresses 400–600 basis points; the trading multiple re-rates from ~25x toward the 10-year peer median of ~10x.
• ◆Trading multiple compression alone is the dominant risk. Even with FY2027 EPS at $16 (base case), a multiple revert from 34x to 18x P/E delivers ~45% downside; at 12x P/E (peer historical norm), downside is ~65%. Multiple compression in cyclicals historically arrives months before earnings deterioration—the leading edge can be sharp and pre-emptive.
• ◆SSD QLC pricing closes the bulk-storage economic gap faster than expected. Even a halving of the SSD-to-HDD $/TB ratio (from current 5–7x to 3x) would dramatically change nearline economics, particularly for warm-tier workloads where the latency premium becomes a usability advantage. NAND producers' aggressive pricing in a recovery cycle could pull this forward.

“Sell-side consensus is unanimously bullish (21 Buy / 3 Hold / 1 Sell; mean PT $520; range $360–660). The active debate is not whether to own the stock but how much further the AI-storage thesis can stretch the trading multiple. Three key questions dominate: (1) Is 'AI-infrastructure-grade' the correct re-rating framework deserving 18–25x EV/EBITDA, or do cyclical-duopoly economics support ~10x mid-cycle with current multiples capitalizing peak earnings?; (2) When does hyperscaler digestion arrive—the typical 18–24 month cycle pulse or an extended multi-cycle AI buildout?; (3) Which company wins the FY27–28 hyperscaler order book on areal density execution—WDC or Seagate? Contrarian bear-side consensus is essentially absent, a yellow flag for a cyclical at peak.”

• ◆Quarterly earnings with sustained gross margin (Q4 FY26 end-July 2026). Gross margin holding above 50% reinforces structural-durability thesis; the first quarter where GM ticks down >200 bps will be the signal that the cycle is turning.
• ◆40TB EPMR and HAMR volume ramp announcements (FY27). Customer qualification to commercial volume to market share gain vs. Seagate. Currently in qualification with 3 customers (EPMR) and 4 (HAMR) per Q3 FY26.
• ◆Capital return resumption (FY27–FY28). Buyback restart and dividend reinstatement each add 5–10% of total-return floor and broaden investor base; net leverage below 1x is the gating event.
• ◆Hyperscaler CapEx commentary (every quarter). AWS/Azure/GCP/Meta storage CapEx guidance is the primary leading indicator for the FY28 demand environment and durability of the 'committed through 2027' window.
• ◆NAND/SSD pricing surge from Sandisk peer set. If SSD prices spike on AI demand, the HDD bulk-storage moat widens; if they collapse on new capacity from YMTC/SK Hynix, the moat narrows.

• ◆Hyperscaler nearline digestion pause in FY2028 (40% 3-year probability): Very High impact (−50–60%). This is the critical cycle-pulse risk that would invalidate the structural-durability thesis.
• ◆Trading multiple compression to peer 10-year median (50% probability): High impact (−40%). Multiple revert from 34x to 18x P/E alone delivers −45% downside even with stable earnings.
• ◆Gross margin mean-reversion to 40% mid-cycle levels (50% probability): High impact (−25%). Erosion of duopoly pricing discipline would compress EBITDA and cut the valuation floor.
• ◆SSD QLC bulk-storage encroachment (15% 3-year, 35% 5-year probability): High long-term impact. If SSD $/TB approaches HDD parity faster than expected, warm-tier workloads migrate away.
• ◆AI CapEx reset on LLM ROI questions (15% probability): Very High impact (−50–70%). Macro slowdown in AI infrastructure spend would collapse the entire nearline demand thesis.