Earnings Analysis · Sandisk (SNDK), Fiscal Q4 2026
By Manish T. · BreakoutBulletin
Sandisk just posted one of the most extraordinary quarters you will ever see from a hardware company. Revenue grew 372% from a year earlier. Gross margins roughly quadrupled to nearly 85%. The board authorized a $14 billion buyback. By any normal reading, this is a company firing on every cylinder. And yet the stock trades about 50% below its summer high and barely moved on the news. That gap – between spectacular results and a skeptical market – is the whole story. Underneath it sits a single question: how long does this last?
The Numbers Were Staggering
Fiscal fourth-quarter revenue hit $8.97 billion, up 372% year-over-year and 51% sequentially. Non-GAAP gross margin reached 84.6%, up from 78.4% a quarter earlier and just 26.4% a year ago – a swing almost unheard of in hardware. Non-GAAP earnings of $39.25 per share handily beat expectations. The company generated $5.04 billion of adjusted free cash flow at a 56% margin and swung to full-year net income of $11.4 billion from a loss the year before. For the full fiscal year, revenue rose 175% to $20.25 billion, with datacenter revenue alone up 437%. These are not the numbers of a cyclical business having a good quarter. They are the numbers of a company caught in a genuine supply shock.
The Engine Was Pricing, Not Just Volume
The single most important detail sits beneath the headline growth. Management disclosed on the earnings call that roughly two-thirds of the sequential revenue jump came from higher average selling prices, with bit shipments driving the rest. That is the unmistakable signature of a real shortage. When demand outruns supply, sellers raise prices, and because those increases carry almost no additional cost, they flow straight to the bottom line – exactly why gross margins went from 26% to 85% in twelve months. As CEO David Goeckeler framed it on the call, “AI is fundamentally a memory-centric, storage-intensive problem,” and the company is seeing that demand translate directly into pricing power. A 58-percentage-point jump in gross margin is not a demand story; it’s a scarcity story, and scarcity is what buys pricing power.
Why AI Is Pulling Storage Along With Everything Else
Sandisk makes NAND flash, the memory that stores data, which is distinct from the DRAM and HBM that feed the GPU directly, but it is caught in the same squeeze. Management described an “era of inference” where every AI interaction generates content that must be stored, retrieved, and served at scale. Datacenter revenue’s 437% full-year surge confirms that the AI build-out is not only a compute and high-end memory story; it is a storage story too. The same wave that lifted GPUs and HBM is now lifting the humble flash drive underneath them.
The $14 Billion Vote of Confidence
On top of the results, the board authorized an additional $14 billion share repurchase, lifting the total remaining authorization to $15.5 billion, after $4.5 billion of buybacks were already executed during the quarter. Management pointed to more than $42 billion in long-term contracts and over four years of demand visibility as the basis for that confidence. The capital-allocation signal is unambiguous: the people running the company are betting the upcycle has staying power.
So Why Is the Stock Down 50%?
If the quarter was this good, why does the stock sit roughly 50% below its summer high, and why did it barely rise? Because the market is not debating whether this quarter was strong. It is debating whether it marks the peak.
Memory is one of the most brutally cyclical industries in technology, and pricing-driven booms have always been followed by downturns when supply catches up and prices roll over. With margins near 85% and revenue up nearly fourfold, the bar for the future is extraordinary. A stock priced for continuation is acutely vulnerable to any hint that pricing is topping out. The 50% drawdown is the market pricing the eventual downturn in advance, even while current numbers remain glorious.
The Skepticism Has Multiple Layers
The cautious reaction isn’t just memory-cycle muscle memory. It reflects several specific concerns the bull case must overcome.
First, supply is never static. Samsung, SK Hynix, Kioxia, and Micron are all ramping high-layer-count NAND capacity. NAND has historically been more commoditized than DRAM, and today’s prices are powerful incentives for competitors to bring supply online faster. Any signs of accelerating capacity additions could start to erode the scarcity premium.
Second, those $42 billion in long-term contracts – which management calls its New Business Model agreements – provide demand visibility, but the market questions whether they lock in price. Memory contracts commonly include annual price renegotiations or volume flexibility clauses. Management has not detailed specific floor-price protections, so the market remains uncertain whether these agreements will truly insulate Sandisk if spot prices begin to decline.
Third, valuation tells its own story. At the current stock price, Sandisk trades at a low-single-digit multiple of trailing net income – roughly 4 times the $11.4 billion in full-year earnings. That kind of multiple implies the market expects profits to collapse, not drift lower. The stock isn’t just pricing a mild correction; it’s pricing a severe cyclical mean-reversion, the kind memory markets have delivered many times before.
What Would Change the View
The entire bull case now rests on whether the cycle has genuinely changed shape. The critical signposts are:
- Gross margin trajectory. This is where a memory cycle turns first. If margins hold above 80% through the next two quarters, the structural-change argument gains credibility.
- Pricing trends. Management’s guidance for next quarter – revenue of $10.3–$10.8 billion and gross margin of 83–85% – implies pricing is still firm. Any slippage would be a warning.
- Competitor capacity announcements. If Samsung, SK Hynix, or Kioxia signal aggressive NAND expansions, the market will price the coming supply wave immediately.
- Contract renewal terms. Details on whether those long-term agreements are being renewed at current pricing or with escalators will signal whether the demand model truly dampens cyclicality.
- Datacenter storage demand. As long as AI inference workloads keep growing and generating data that needs to be stored, the demand side of the equation remains supportive.
If those signposts hold, the case that this cycle is structurally different strengthens. If any begin to turn, the market’s instinct to fade the peak will look prescient.
The Bigger Picture
Sandisk is the clearest single-company evidence of a theme running through this entire AI cycle: the build-out is a memory-and-storage story as much as a compute story, and the companies that control scarce supply capture enormous, pricing-driven profits. It is the same dynamic visible from advanced packaging to the T-glass bottleneck – now expressed as an 85% gross margin.
The lasting lesson is in how the market prices a boom it believes is cyclical. Spectacular results confirm the supercycle is real and immensely profitable right now; the muted stock says investors have seen memory cycles before and are already hunting for the top. Both can be true at once. The interesting question was never whether Sandisk had a good quarter. It was whether the contracted-demand model has actually broken memory’s brutal cycle, or merely delayed it – and that is what the next several quarters will decide.
Note: Figures are drawn from Sandisk’s fiscal Q4 2026 earnings release, earnings call commentary, and public reporting as of the publication date, and may be revised. Valuation multiples are based on the approximately 50% stock price decline from the summer high referenced in the article; exact multiples will vary with daily price changes.
Related Reading
• The Memory Shortage Behind the AI Boom: How HBM Demand Is Draining Conventional DRAM
• It's Not the Wafer Anymore: How Advanced Packaging Became the Binding Constraint on AI Chips
• Follow the Cargo: The AI Boom Is Hiding in Plain Sight in the U.S. Trade Deficit
Disclaimer
BreakoutBulletin publishes educational and analytical content only. Nothing here is investment, financial, legal, or tax advice, or a recommendation or solicitation to buy, sell, or hold any security. Past performance does not indicate future results. Readers should conduct their own research and consult a qualified, registered financial adviser before making any decision.