SanDisk and Western Digital: The Memory Recovery Is Not a Demand Story. It's a Supply Constraint Story.

The Consensus View

Western Digital is set to report its fiscal fourth quarter and full-year 2026 results on August 5, 2026. The prevailing narrative framing these earnings is straightforward: AI demand has created unprecedented appetite for storage, and both companies are riding a demand-driven supercycle. Analyst notes and headlines point to data center spending, generative AI workloads, and the expanding addressable market for persistent storage.

That explanation is incomplete. It treats supply as a passive variable - the idea being that demand pulled up, and manufacturers responded by shipping more. The data shows something different. The recovery is being driven by manufacturers' willingness to constrain supply, then by structural allocation exhaustion.

Western Digital and SanDisk are no longer one company. They became two after Western Digital spun off its NAND flash business into SanDisk in February 2025. That separation matters because what was once a single storage company is now two structurally different businesses in two structurally different markets - one in hard disk drives, one in NAND flash - and each is being driven by supply discipline, not demand discovery.

Western Digital: Sold Out Before Earnings

Western Digital's HDD business is in a position that would have been inconceivable a few years ago. CEO Irving Tan confirmed that the company's 2026 HDD production capacity is fully sold out, with only 5 percent allocated to the consumer market. The remaining 95 percent is committed to enterprise and cloud data center customers through firm purchase orders and long-term agreements extending beyond 2026.

This is not a demand signal in isolation. It is a supply constraint where the supply side has no room to respond. Western Digital cannot build more drives to meet demand because its manufacturing lines are already booked. The result is pricing power - and that is what has been driving the margin trajectory.

Western Digital's gross margin expanded from 39.8 percent in Q3 FY25 to 50.2 percent in Q3 FY26 - a 1,040 basis-point increase year over year. That is not the profile of a company simply shipping more units. That is the profile of a company whose customers are absorbing price increases because there is no alternative capacity. The competitive landscape in HDDs has shrunk to three players - Western Digital, Seagate, and Toshiba - and none of them can meaningfully expand capacity on short notice. HDD fabs have multi-year lead times, and the existing players have chosen to allocate capacity to the highest-value enterprise contracts rather than flood the consumer channel.

The capital expenditure story reinforces the supply-side frame. Western Digital's trailing twelve-month capex stands at $381 million against $2.9 billion of free cash flow. That is not the spending pattern of a company racing to build capacity. It is the pattern of a company collecting cash from a constrained market. The company raised its quarterly dividend 20 percent to $0.15 per share in the Q3 quarter - a move consistent with excess cash in a structurally tight market.

SanDisk: Margin Expansion From Pricing, Not Volume

SanDisk's trajectory since its February 2025 separation has been even more extreme. The stock is up approximately 500 percent year-to-date. Revenue grew 251 percent year over year in Q3 FY26 to $5.95 billion. But the number that carries the structural thesis is the gross margin: 78.4 percent on a non-GAAP basis, up from 22.7 percent a year earlier.

A near 55.7-percentage-point margin expansion in one year cannot be explained by unit growth. SanDisk's margin explosion is a pricing and mix phenomenon. The company shifted toward datacenter customers - which grew 645 percent year over year - while consumer revenue declined 10 percent sequentially. Simultaneously, SanDisk moved to what it calls a "New Business Model" of multi-year customer agreements with firm financial commitments. That is a contract structure that locks in pricing ahead of delivery, which is functionally a supply-commitment model. The company signed three such agreements in Q3 and two more in Q4.

The NAND supply picture confirms why SanDisk has this leverage. Samsung, SK Hynix, Kioxia, and Micron collectively scaled back NAND production in the second half of 2025. SK Hynix cut NAND output by approximately 10 percent. Micron held production at its main Fab 7 plant in Singapore at low levels. This deliberate supply discipline, layered on top of AI-driven enterprise SSD demand, has eliminated any buffer in the NAND market. NAND contract prices were projected to rise 70 to 75 percent quarter over quarter in Q2 2026 - outpacing DRAM for the first time in the current cycle.

SanDisk's Q4 FY26 guidance reinforces the supply-constraint thesis. The company expects revenue of $7.75 billion to $8.25 billion and non-GAAP EPS of $30 to $33. Those numbers assume the pricing environment holds and the company continues to shift mix toward datacenter.

The NAND Market Has Split Into Two

The NAND flash market is not a single market right now. It has bifurcated along two fault lines.

The first split is between enterprise SSDs and consumer flash. Enterprise customers are locked into long-term agreements with multi-quarter commitments. Consumer buyers face spot-market pricing with no contract protection. Suppliers are allocating production to the former and rationing the latter. SanDisk's end-market breakdown - datacenter up 233 percent sequentially, consumer down 10 percent - is the operational manifestation of this split.

The second split is between players with structural capacity and those without. Samsung, SK Hynix, and Micron control approximately 63 percent of the NAND market. SanDisk and Kioxia sit in the second tier, and Western Digital has reportedly re-entered merger discussions with Kioxia to combine their NAND operations. Whether that deal proceeds depends on valuation terms and regulatory clearance - but the underlying logic is the same supply-constraint thesis: smaller NAND players need scale to survive a market where pricing power belongs to capacity.

What the Earnings Call Will Tell Us

The market's reaction will depend on whether the supply-constraint thesis holds through the full year and into the forward guide. Here is what matters.

For Western Digital: The question is whether Q4 revenue hits the midpoint of $3.65 billion and whether gross margin reaches the guided range of 51 to 52 percent. More important is the commentary on 2027 capacity and whether long-term agreements are extending further out. If management says visibility stretches to 2028 or beyond, that confirms the structural constraint is durable rather than cyclical.

For SanDisk: The Q4 EPS guidance of $30 to $33 is the anchor. At those levels, the market is implicitly pricing in that the 78-plus percent margin environment continues. Any crack in ASP growth or datacenter mix would be a signal that the pricing power is softening. The number of New Business Model agreements signed and the duration of those contracts will indicate how much of the margin expansion is locked in versus exposed to spot-market volatility.

Both companies should be watched for capex guidance. If either announces a meaningful increase in capital expenditure, that would be the earliest signal of capacity response - and the beginning of the end of the current pricing environment. The implication is fairly straightforward: as long as capex stays disciplined relative to cash flow, pricing power persists.

Investor Takeaway

The key issue for Western Digital's August 5 earnings is not whether demand remains healthy. Demand is the background condition. The more important question is whether supply discipline holds.

For Western Digital, the structural story is intact as long as the company continues to allocate its constrained HDD capacity to enterprise contracts at elevated pricing and does not announce aggressive capacity expansion. The stock has risen approximately 218 percent year-to-date, which reflects the current thesis. The risk is not in the near term - the risk is in the cycle turn, which comes when someone decides to build more capacity.

For SanDisk, the margin expansion from 22.7 percent to 78.4 percent is extraordinary but structurally dependent on continued NAND supply constraint and datacenter mix. The $7.75 billion to $8.25 billion Q4 guidance assumes this environment persists. The stock's 500-plus percent year-to-date return has front-run a significant amount of this optimism. The risk here is the same: the cycle turns when capacity catches up, and the Kioxia merger talks - while interesting - are a consolidation play, not a capacity play.

The forward condition that determines whether both stories hold is the same. Watch capex trajectories, watch new capacity announcements from the larger players, and watch for the first signal that ASP growth slows while unit growth accelerates. That would be the inflection - the moment supply discipline gives way to supply response. Until then, the pricing power belongs to whoever controls the bottleneck.