On August 5th after the market close, Western Digital and SanDisk Corp. simultaneously released their fiscal Q4 2026 results. Both exceeded expectations, and SanDisk Corp. also authorized a massive $140 billion share buyback. The immediate market reaction was a sharp decline, with Western Digital shares falling 11% after hours, while SanDisk Corp. dropped 7%.
This scenario has become a recurring theme this year: record-breaking financial performance paired with declining stock prices. The memory of SK Hynix plummeting 30% intraday in July is still fresh, and the sell-off in the memory sector continues. This is not an isolated issue for one or two companies—it signals a broader market shift in the pricing logic for the entire memory super-cycle.
SanDisk Corp.: A Business with 84.6% Gross Margins
The financial report shows that SanDisk Corp. achieved Q4 revenue of $89.65 billion, a 372% year-over-year increase and a 51% sequential rise. GAAP net profit was $69.03 billion (diluted EPS of $43.97), compared to a net loss of $23 million in the same period last year. Non-GAAP EPS of $39.25 significantly surpassed the Wall Street consensus estimate of $34.37. The gross margin stood at an impressive 84.6%.
The primary driver of this performance is singular: NAND flash pricing. According to SanDisk Corp. management, about one-third of the sequential revenue growth came from increased shipment volumes, with the remaining two-thirds attributed to price increases. A more critical change is occurring in the customer mix. Data center revenue in Q4 reached $29.77 billion, doubling sequentially and surging 1298% year-over-year. The proportion of data center shipment bits relative to total output skyrocketed from 12% a year ago to 38%. Revenue from edge devices (including smartphones, PCs, and automotive) was $54.32 billion, up 48% sequentially. Consumer revenue fell 32% sequentially to $556 million, as the company deliberately shifted production capacity away from retail channels to serve enterprise clients.
CEO David Goeckeler succinctly summarized the situation on the earnings call: "The consumer business simply cannot keep pace with the trajectory of the transactional market." A key indicator of this strategic shift is the "New Business Model" (NBM) frequently mentioned by SanDisk Corp.—where cloud providers pre-commit to volumes and pricing, granting SanDisk Corp. revenue visibility and financial guarantees. The Q4 report details the current scale: 10 agreements signed with 8 customers, covering a minimum contract revenue of $939 billion, backed by $165 billion in financial guarantees (including cash deposits and financial instruments). The weighted average contract term exceeds four years. The CFO disclosed that approximately 50% of shipment bits for FY2027 are already locked in via NBM, with this figure expected to rise to about two-thirds by FY2028. "Customer demand growth is outpacing our supply capacity, and bit allocation will continue beyond 2027."
If these agreements are honored, even during a market downturn, SanDisk Corp. could transform from a highly cyclical NAND wholesaler into an infrastructure provider with long-term revenue visibility.
Western Digital: The "Contingency" in GAAP Profit and 89% Cloud Dependence
Western Digital reported Q4 revenue of $37.47 billion, a 44% year-over-year increase. Non-GAAP gross margin was 54.4%, operating margin was 44.2%, and diluted EPS was $3.56, up 109% year-over-year, all exceeding expectations. However, a crucial detail needs highlighting. Under GAAP, Q4 net profit was $31.95 billion, a 1215% year-over-year increase. This figure is often cited in media headlines but significantly overstates the actual operational performance. The reason is that after the spin-off, Western Digital retained a portion of SanDisk Corp. shares. Due to the sharp rise in SanDisk Corp. stock in Q4, the market value revaluation of these holdings generated an unrealized book gain of approximately $20.5 billion. Excluding this, non-GAAP net profit was only $13.82 billion. The "12-fold surge in net profit" requires a substantial discount.
On an operational level, Western Digital is now a pure-play HDD manufacturer after the separation. Cloud customer revenue in Q4 accounted for 89% of total revenue, reaching about $33 billion, a 43% increase year-over-year. Client revenue contributed 6%, and consumer revenue 5%. This highly concentrated customer base is a double-edged sword: demand is predictable, but pricing power is almost entirely in the hands of those customers. CEO Irving Tan repeatedly argued on the call that AI is structurally driving HDD demand. He explained that training generates data, inference generates even more data, and AI agents executing multi-step tasks create intermediate data that needs persistent storage at every step. "Computing power can be reused, but data only accumulates." On the product front, the 40TB ePMR hard drive has begun shipping and is expected to contribute over half of nearline bit shipments by FY2027 Q3. The 44TB HAMR drive is slated for mass production in the first half of 2027, with a 50TB version targeting the second half. The company is negotiating long-term agreements covering 2029 to 2031. On pricing, the blended average price per TB in Q4 increased about 17% year-over-year (up from high single digits last quarter), while cost per TB decreased 8% year-over-year. This combination of volume and price increases, coupled with declining costs, pushed gross margins from 41.3% a year ago to 54.4%.
What is the Market Really Afraid of?
In the calendar Q2 2026 (corresponding to the Q4 FY2026 quarter for Western Digital and SanDisk Corp.), the memory industry delivered perhaps its most spectacular results ever. Samsung Electronics reported quarterly revenue of 171.5 trillion won, up 130% year-over-year, with operating profit surging 1814% to 89.5 trillion won. The semiconductor division contributed 99% of the company's total profit. SK Hynix posted quarterly revenue of 79.3 trillion won, up 257% year-over-year, operating profit of 60.5 trillion won (up 557%), and net profit of 93.9 trillion won. HBM (High Bandwidth Memory) shipments continued to soar, and the company has signed long-term supply agreements (LTAs) with over a dozen clients, including a core customer. Micron Technology reported Q3 FY2026 (March-May) revenue of $414.6 billion, a 345.7% year-over-year increase, and GAAP net profit of $282.4 billion, up 1398.3%. Kioxia reported Q1 FY2026 (April-June) revenue of 1.77 trillion yen, up 415.5% year-over-year, and net profit of 842.2 billion yen, a staggering increase of over 45 times. Enterprise SSD revenue for servers grew over 440% year-over-year. Seagate Technology reported Q4 FY2026 revenue of $36.29 billion, with a non-GAAP gross margin of 52.7% and non-GAAP EPS of $5.71. Full-year revenue was $121.95 billion, up 34%.
All five companies reported record highs, driven by a highly consistent force: global cloud capital expenditure in 2026 is projected to exceed $800 billion (from Microsoft, Google, Amazon, Meta, etc.), with AI training and inference clusters consuming storage far more than anticipated. TrendForce data shows that in Q1 2026, DRAM contract prices surged 93% to 98% quarter-over-quarter, and NAND Flash contract prices rose 85% to 90%. Server DRAM demand share exceeded 50% for the first time, marking a shift where data centers, not smartphones, are the primary driver of storage demand. Yet, amid this collective euphoria, the capital market sent a disturbing signal. In July, SK Hynix shares saw a maximum drawdown of 54% on the Korean stock market, Samsung Electronics 42%, Micron Technology 33%, and SanDisk Corp. plunged 47% in a single month (wiping out over $150 billion in market value). The Philadelphia Semiconductor Index recorded its worst monthly decline since 2008.
This reflects a vote of no confidence from the market regarding the sustainability of the entire memory super-cycle. There are two layers of explanation. The shallow explanation is "buy the rumor, sell the news." SanDisk Corp. provided Q1 FY2027 revenue guidance of $103 billion to $108 billion, but the Wall Street consensus was $111.6 billion – a mid-point miss of about 6%. For a stock that had risen over 400% year-to-date, exceeding expectations was merely a passing grade; falling short triggered a stampede. Western Digital’s guidance actually topped expectations – a revenue midpoint of $41 billion versus the expected $40.2 billion, and EPS of $3.85 to $4.15 versus the expected $3.83 – but even that wasn't enough. After a nearly 200% gain for the year, the market demands not just a slight beat, but a signal strong enough to support another 200% rally. The deeper explanation relates to the memory cycle itself. TrendForce predicts that Q3 2026 DRAM contract prices will rise 13% to 18% quarter-over-quarter, and NAND Flash prices will rise 10% to 15%. If confirmed, this would mean the pace of price increases has decelerated from the doubling rhythms of previous quarters to mid-to-high single digits. A July 28th report from Jefferies was even more pessimistic, suggesting actual Q3 price increases might only be 15% to 20%, well below the previously expected 25% to 30%, and warning that "price peaks may be arriving earlier than expected." The consumer segment is also pushing back. OPPO and vivo have reportedly refused to accept Samsung’s Q3 price quotes, and SanDisk Corp.’s own Q4 consumer revenue fell 32% sequentially. Goeckeler acknowledged the company is "finding the balance between price and volume" – at a certain price point, demand disappears.
More fundamentally, although this cycle is driven by AI data center demand, pricing has already run far ahead of actual operational performance. SanDisk Corp.’s full-year non-GAAP net profit was $109.9 billion. Against a market cap of roughly $2.1 trillion, its P/E ratio is around 20 times, which does not seem expensive. But this is based on NAND prices rising five-fold in a year. Any slight correction would lead to a very sharp contraction in profits. Goldman Sachs warned after the earnings that SanDisk Corp.’s weaker-than-expected guidance would likely spill over to Micron Technology, potentially triggering a cascading revaluation across the memory sector.
One Year Post-Split, Two Earnings Reports, One AI Story
On February 21, 2025, Western Digital completed the most significant restructuring in the storage industry in nearly a decade: spinning off its NAND flash memory business into an independent, publicly traded company, SanDisk Corp.. The logic behind the split is straightforward. HDDs and NAND are fundamentally different businesses. The former is a low-growth, high-cash-flow "rent-collecting" model, while the latter is a high-volatility, capital-intensive cyclical gamble. When Western Digital bought SanDisk Corp. for $19 billion in 2016, the goal was vertical integration. By 2023, under pressure from activist investor Elliott Management, the company finally acknowledged the marriage was not working. At the time of the split, the market assigned very different labels to the two entities: Western Digital was seen as a "relic of the past," an HDD manufacturer destined to be replaced by SSDs, while SanDisk Corp. was the "ticket to the AI era," holding NAND capacity and targeting data centers. In reality, both Western Digital and SanDisk Corp. play different roles within the same AI narrative. SanDisk Corp. is at the forefront, capturing the benefits of cyclical elasticity – it can net $69 billion in a single quarter during an upturn but could also swing to a loss during a downturn. Western Digital holds the rear, profiting from capacity barriers – its growth rate (44% vs 372%) is lower, but its profit predictability is higher. The common signal from both earnings reports is clear: AI demand for storage is far from its peak. But market pricing has stopped following fundamentals. When a stock has risen 400% in a year, "meeting expectations" is bad news, and "slightly below expectations" is a rout. The entire AI hardware sector is undergoing a paradigm shift in valuation, moving from "any connection to AI is valuable" to a phase where "growth must consistently prove it can exceed expectations."
On the technology front, the two companies are also diverging. SanDisk Corp. is betting on HBF (High Bandwidth Flash). On August 3rd, it jointly released an OCP specification with SK Hynix, aiming to use NAND to supplement HBM and penetrate the AI inference market. Western Digital is taking a more deterministic path, using HAMR to continuously push the HDD capacity ceiling: 40TB is shipping, and 44TB is on the way. The former offers higher potential but carries greater risk; the latter is more stable, but its ceiling is clearer. Key monitoring points for the coming quarters include: the actual magnitude of Q3 storage contract price changes (if TrendForce's predictions materialize, the narrative of peak pricing will strengthen), the execution quality of SanDisk Corp.’s NBM agreements, and the progress of Western Digital’s HAMR production ramp. In a cycle where "performance keeps hitting new highs, but stock prices keep hitting new lows," the memory industry is transitioning from a phase where "everyone can profit from the cyclical dividend" to one where only structural barriers can create lasting value.
Comments