💾

WDC

💡
Thesis. A pure-play HDD duopolist growing 44% with a 34% free cash flow margin, net cash, exabytes up 25% and a 40TB-to-44TB roadmap — ~90% dependent on hyperscaler capex and trading as a memory-complex proxy whether it wants to or not.
Reviewed 19 Sep 2026, after fiscal Q4 2026 results on 5 Aug. Next review after fiscal Q1 2027 earnings on 28 Oct 2026.

What it does

Western Digital has been a pure-play hard disk drive company since spinning off its flash business as
🧠
SNDK in February 2025. It makes nearline HDDs for datacentre bulk storage, effectively a duopoly with
💾
STX. About 90% of revenue now comes from the cloud segment; client and consumer are roughly 5–6% each.
The technology names matter because they are the whole competitive position. ePMR and UltraSMR are the current recording methods; HAMR is the next step. Each generation raises terabytes per drive, and that is what keeps HDD cost-per-bit below flash for bulk storage.
LTAs — long-term agreements with pricing structures — are how hyperscalers now buy this capacity. Coverage percentages are not disclosed.
Two companies make this product and neither can make it faster than the other. That is the whole position, and it lasts exactly as long as the discipline does.

Bull case

  • Fiscal Q4 2026 revenue $3.75B, +44% YoY and +12% QoQ. FY26 revenue $12.92B, +36%
  • Free cash flow $1.3B — a 34% FCF margin on a hardware business. Net cash $500M ($1.6B cash against $1.1B debt)
  • Exabytes shipped +22% YoY in Q4, +25% for FY26, guided 25%+ ongoing, with pricing structures locked via LTAs
  • Q1 FY27 guided ~$4.1B, +42% to +49% YoY — acceleration, not deceleration
  • 40TB ePMR began shipping in the June quarter and is entering volume with two customers, on track for 50% of nearline bits by Q3 FY27. 44TB HAMR on track for H1 CY2027 with "very positive" qualification feedback. UltraSMR ramping with a third major customer, expected to be ~60% of nearline exabytes exiting FY27
  • Capital return: 2.3M shares repurchased for $1B in the quarter, including $328M to settle convert premiums in cash and avoid ~773K new shares. $0.15/share dividend
  • The Sandisk separation now looks decisively right — Sandisk did $20.25B of revenue against WD's $12.92B, and each is focused

Bear case

  • ~90% cloud concentration. The business is entirely a function of hyperscaler capex, and management acknowledges the dependence
  • It trades as part of the memory complex regardless of its own fundamentals — it shed 13% in a single session in late August with no company-specific headline, and went from ~$589 on 13 August to ~$461 on 4 September, a 22% drawdown in three weeks
  • Long-term SSD substitution as flash cost-per-bit falls each generation
  • Quarter-to-quarter exabyte and margin variability by product mix, per management
  • LTA coverage percentages are not disclosed, so the visibility is asserted rather than quantified
  • Duopoly pricing depends on both players holding capacity discipline, and current margins are the incentive to break it

Major customers

  • Hyperscalers — the cloud segment at ~90% of revenue. Individual names are not disclosed, but this is Microsoft, Amazon,
    🏛️
    GOOG     and
    🏛️
    META    . Two customers in volume on 40TB ePMR; a third ramping UltraSMR
  • Enterprise storage OEMs — Dell, HPE, NetApp
  • Client and consumer — ~5–6% each, shrinking
🔗
Read-through. WD and
💾
STX are one trade on the mechanical storage tier and should be sized as such. WD's exabyte growth of +25% is a direct read on hyperscaler cold and warm storage demand — the same capex that drives
🧠
SNDK on the flash tier and
❄️
VRT on the facility side. The 40TB/44TB roadmap is precisely what keeps HDD cost-per-bit ahead of Sandisk's QLC — the two are partners in the buildout and competitors on the substitution curve, which is why they can both be right for several years and only one of them right eventually. Note also that
💾
P sits on the other side of this trade as a NAND buyer, so a Storage-category average blending all four is measuring opposite exposures.

What would change the view

  1. Q1 FY27 landing at ~$4.1B (+42–49%)
  2. 40TB ePMR reaching 50% of nearline bits by Q3 FY27, and 44TB HAMR shipping in H1 CY2027
  3. Exabyte growth holding above 25%
  4. LTA coverage and pricing structure disclosure — currently the biggest gap in what is known
  5. Whether the 22% August drawdown was positioning or something HDD-specific
  6. Any capacity discipline break by either duopolist

Update log

19 Sep 2026 — Converted to the standard template. Current through fiscal Q4 2026; no new disclosures since.
7 Sep 2026 — Correction carried forward: WD completed the monetisation of its remaining 1.7M Sandisk shares during Q4, exchanging them for 4.8M WD shares. It no longer holds any SNDK. Earlier pages in this database describing WDC as a 10%+ Sandisk holder are out of date — that overhang is gone.
5 Aug 2026 — Fiscal Q4 and FY2026 reported. Q4 revenue $3.75B (+44% YoY, +12% QoQ); FY26 $12.92B (+36%); free cash flow $1.3B at a 34% margin; exabytes +22% in Q4 and +25% for FY26; net cash $500M. Q1 FY27 guided ~$4.1B (+42–49%). 2.3M shares repurchased for $1B. Received $2.05B from Sandisk in Q4 via the equity stake, $6.5B for the full year. 40TB ePMR in volume with two customers; 44TB HAMR on track for H1 CY2027.
Research and education only — not investment advice.