The end of AI data centers coming? (From Chaikin Analytics)

When a stock sheds almost two-thirds of its value in a matter of months, the instinct is to go hunting for a bargain. Sandisk Corp. (NASDAQ: SNDK), the flash-memory and data-storage specialist, is dangling exactly that temptation in front of investors.

The stock finished down more than 22% on Nov. 5 alone, a single-session rout that dragged the shares to roughly 65% below their June peak. Its relative strength index has now sunk to its lowest reading since April 2025, the deeply oversold reading that usually has contrarians circling.

For a company once crowned among the biggest winners of the AI boom, that’s a staggering fall from grace. The pain hasn’t been Sandisk’s alone, either: the sell-off has spread across the entire AI trade, with the tech-heavy NASDAQ index suffering its worst day in 19 months on Nov. 5, down around 5.5%.

So what happened to the market’s favorite memory stock, and is this the moment brave investors have been waiting for? The answer starts with what didn’t cause the crash.

The Sell-Off Isn't About Earnings

Despite the latest drop coming just a week after the company’s Q1 fiscal year 2027 earnings, they can’t be blamed here. Sandisk’s late-October results were, by almost any measure, spectacular. Revenue came in above $11 billion, up 377.5% on a year earlier, with gross margins north of 85% and billions of dollars flowing back to shareholders through buybacks.

Management then guided the current quarter higher rather than lower, hardly the posture of a company bracing for a collapse. On the headline numbers, this was a business performing about as well as it ever has. In other words, the market looked straight past a blockbuster quarter and fixated on a very different question: how long can numbers this good possibly last?

Why the Market Stopped Believing the Margins

That question cuts to the heart of the bear case. Gross margins around 85% are extraordinary for a memory company, and they exist largely because NAND pricing has been running so hot. The worry is simple: margins like these tend to mark a peak, not a plateau.

The stock’s valuation only sharpens the anxiety. After the latest leg down in the crash, Sandisk trades at a price-to-earnings ratio of about 6.5x on trailing earnings. That sounds absurdly cheap—until you realize it’s only true if today’s bumper profits last. If those earnings are sitting near a cyclical top, that low multiple is a mirage rather than a margin of safety.

To be sure, the bulls have a real rebuttal. Sandisk’s data-center business is booming, with revenue up more than 60% sequentially, and a chunk of its future output is already locked in under long-term supply agreements covering roughly half of fiscal 2027 NAND bit output. That’s real visibility, rather than wishful thinking, but the market is obviously asking for more reassurance.

The NAND Cycle Is the Real Swing Factor

Underneath the stock’s potential sits the current memory cycle, and here the picture splits in two. Enterprise and data-center demand, powered by AI, remains strong enough to keep pushing specialist pricing higher. Consumer demand, by contrast, is buckling as those higher prices get passed through, with Sandisk’s consumer revenue falling nearly 30% from the previous quarter.

That divergence leaves the company leaning heavily on hyperscaler spending, which is a wonderful place to be until it isn’t. Any moderation in AI capital budgets, or a wave of fresh supply as rival NAND producers ramp up capacity, could unwind the very pricing that makes those eye-wateringly high margins possible.

This is why the sell-off is as much macro as micro. Sandisk slid alongside the broader AI trade as investors suddenly questioned whether the current boom is durable or merely cyclical, and a stock priced for flawless execution had the most to give back.

So Is Sandisk a Bargain or a Value Trap?

The honest answer is that it hinges almost entirely on what you think about one variable: whether Sandisk’s current earning power is a new structural reality or the high-water mark of a cycle. If enterprise demand holds and prices only drift lower gradually, the stock looks seriously mispriced at these levels.

If, on the other hand, margins normalize and consumer weakness spreads, then that tempting low multiple is exactly the trap it looks too good to be. Anyone betting on a bounce here is really making one call above all others: that this is the strongest phase of the cycle, not its final act. For now, the chart screams bargain while the cycle whispers caution, and only the next few quarters of NAND pricing data will settle which signal was right.

Read this article online ›