$SNDK: 85% Margins, a 376% Revenue Jump, and a Stock That Keeps Falling. The Cycle Math Explains Both

Sandisk just reported one of the most extreme earnings quarters in semiconductor history. Revenue of $11.0 billion for its fiscal first quarter, up 23% from the prior quarter and nearly fivefold from the $2.31 billion it posted a year ago. A non-GAAP gross margin of 85.1%, a number that memory companies are not supposed to be able to print. Earnings of $49.51 per share against $1.22 in the same quarter last year. Guidance for the current quarter came in higher still: $12.2 to $12.6 billion in revenue and $55 to $57 in earnings per share.

The stock fell 5.3% on Thursday to close at $1,330.05, and as of Friday premarket it was indicated down roughly 9% more, near $1,209. From its 52-week high of $2,354, Sandisk has now given back around 45% of its value, and it is doing so while reporting the best numbers it has ever produced.

This is not a company missing expectations. Revenue beat consensus. Earnings beat consensus. The guide was above the Street. The market looked at all of it and sold anyway, hard, for a reason that every memory investor eventually learns the expensive way: in commodity semiconductors, the numbers are most beautiful at exactly the moment the cycle is most dangerous. The question the stream is arguing about today is whether this time the cycle logic is wrong.

What is actually happening here

Sandisk is the NAND flash business that Western Digital spun off in February 2025. For most of its first year as a standalone company it traded like what it was assumed to be: a cyclical commodity memory maker in an oversupplied market. The stock's 52-week low is $183.

Then the AI buildout collided with flash storage. Data centers training and serving large models need vast amounts of fast storage, and enterprise solid state drives started displacing hard drives in AI infrastructure at scale. At the same time, years of underinvestment in new NAND capacity across the industry meant supply could not respond quickly. Contract pricing for NAND went vertical through 2026, and Sandisk's income statement went vertical with it. Full fiscal year 2026 revenue came in at $20.25 billion, up 175% year over year, and the company has now done $11 billion in a single quarter with a gross margin that looks more like enterprise software than silicon.

The data center segment is the engine: $4.8 billion in the September quarter, up 61% sequentially, now roughly 44% of total revenue. Management attributed the growth to sustained demand from AI infrastructure customers and its shift toward long-term supply agreements. In its prior quarter, the company said sequential revenue growth was roughly one third volume and two thirds price. That ratio is the entire debate in one sentence. Volume growth is durable. Price growth in a commodity is a loan the cycle eventually calls in.

Briefings like this are free to read with a free AlphaBriefing membership. Join free →

Why the stock is falling on blowout numbers

Three things are driving the selloff, and none of them are about the quarter that was just reported.

First, deceleration inside the acceleration. Revenue grew 23% sequentially this quarter after larger sequential jumps earlier in the run, and the midpoint of guidance implies roughly 13% sequential growth next quarter. The business is still growing at rates almost no large company achieves, but the second derivative has turned, and momentum money trades the second derivative.

Second, the margin itself. An 85% gross margin in NAND flash is not a steady state anyone in the industry has ever sustained. Coverage of the print immediately framed the question as whether this marks the near-term peak of an unusually favorable pricing cycle. When a commodity producer earns margins like this, every competitor on the planet is being handed an engraved invitation to add supply.

Third, the stock had already discounted a great deal of perfection. Even after the drawdown, Sandisk at Thursday's close carried a market value around $200 billion on roughly 151 million diluted shares, against a company that was worth a small fraction of that eighteen months ago. Positioning was crowded, the run was a twelve-bagger from the lows, and crowded winners get sold on good news when the marginal buyer is exhausted.

There is also a detail from the release worth sitting with: Sandisk spent $5.0 billion buying back 3.2 million of its own shares during the quarter, an average price near $1,560. The stock now trades meaningfully below where the company itself was buying. Either management torched billions buying the top, or the market is offering the stock cheaper than the people with the best view of demand were willing to pay. Both readings have historical precedent in memory.

So the setup is this: the fastest-growing large-cap income statement in the market, a valuation that has compressed to a single-digit multiple on forward guidance, a dated mechanical catalyst three weeks out, and a cycle argument that has ended every previous memory boom without exception. Which force wins, and at what prices? That is the part worth being precise about.

A free membership gets you access to every free article at AlphaBriefing.com: Join AlphaBriefing.

The rest of this briefing is for paid members: the forward earnings math and what multiple the market is actually paying, the November 19 stock split mechanics and why they matter for flow, three scenario frameworks with price zones for the supercycle, plateau, and cycle-roll cases, and the bottom-line read on how to think about a commodity printing software margins.

AlphaBriefing Paid gets you every investment thesis, scenario framework, and catalyst brief we publish, the analysis private intel clients pay four figures for, at a fraction of that.

Unlock the full briefing →