SanDisk Crashes 22%, Micron Plunges 17%: Why Are Memory Stocks Falling?

For months, the memory trade had a wonderfully simple story. The world wanted more AI. AI needed more memory and storage. And the companies making those essential ingredients were suddenly earning extraordinary amounts of money.

Then came a session that made the same story feel much less comfortable. On November 5, SanDisk stock plunged 22.26%, while Micron stock fell 17.46%. Measured against their 52-week intraday highs, the two were now down roughly 66% and 48%, respectively.

Yet their latest earnings still showed businesses growing rapidly. That is what makes this selloff worth understanding: investors were losing confidence in the future while the recent past still looked spectacular.

Why did the mood change so violently? To answer that, we need to follow the journey from a shortage of chips to a surplus of expectations.

How Bad Was the SanDisk and Micron Stock Crash?

Thursday’s losses were the latest stage of a retreat that had already been underway for months. Here is what the damage looks like when we separate one terrible day from the longer decline.

Sources: Barchart market data displayed by Raymond James. Drawdowns calculated using the closing price and 52-week intraday high. Year-to-date figures are rounded price returns.

That last row is the twist. Both stocks could suffer enormous falls from their peaks and still be big winners for someone who bought at the start of the year. Someone who bought near June’s highs would have had a very different experience.

Even before Thursday’s crash, SanDisk had already fallen about 56% from its high and Micron about 37%, calculated using their November 4 closes. The election-related headlines intensified an existing retreat; they could not explain all the damage that came before.

Why did SanDisk and Micron Stocks Crash on November 5?

The immediate backdrop was a broader retreat from companies supplying the AI infrastructure boom. The Nasdaq fell about 5.5%, while the Philadelphia Semiconductor Index dropped roughly 12%. Memory stocks were among the hardest hit.

A major concern was what the US midterm election results could mean for data-centre construction. Democratic victories, including in important states for the industry, increased investor fears of tighter scrutiny over projects, electricity costs and local approvals. Reuters identified those concerns as a driver of the selling.

The connection to memory suppliers is fairly straightforward. A planned data centre eventually becomes orders for servers, memory and storage. If permission, power connections or construction take longer, some of those orders may also arrive later.

For a company whose valuation assumes years of rapid expansion, delays can matter even when the eventual project still goes ahead. Investors may lower their estimates of near-term sales or demand a lower share price for bearing the uncertainty.

The market was reacting to the risk of slower expansion. An election result alone does not establish how many chip orders will actually be delayed or cancelled.

There was also fresh evidence of weakness elsewhere in chips. Qualcomm’s November 4 results showed handset-chip revenue falling 26% year on year to $5.14 billion. Its smartphone business has its own challenges, including lower Apple-related revenue, so this is not a direct forecast for Micron or SanDisk. It nevertheless added to an uncomfortable picture for technology demand.

Micron faced an additional company-specific headline. Bloomberg reported that it had suspended work on a $4.5 billion memory-chip plant in Taiwan amid a prolonged labour dispute. Reuters relayed the report but said it could not independently verify it, and Micron had not immediately responded to its request for comment.

If prolonged, a construction delay could complicate Micron’s future expansion. It should not be confused with a confirmed shutdown of existing chip production, and it cannot explain why SanDisk and the rest of the sector were falling together.

Why Were Both Stocks Falling Long Before this Session?

The earlier decline revolved around a difficult question: how long could the memory boom keep producing exceptional profits? Three developments made that question harder to dismiss.

First, investors were becoming more demanding about the economics of AI spending. A supplier can have strong orders today while shareholders worry about whether its customers will earn enough from AI to keep expanding at the same pace.

TrendForce’s November 4 NAND market update described that tension directly: heavy spending by cloud providers was raising questions about how AI would generate financial returns. Meanwhile, enterprise storage demand remained firm, but consumer demand and parts of the spot market were weak.

“AI needs more memory” was still a useful starting point. It was becoming less sufficient as a complete investment argument.

Second, the pace of growth in memory prices was drawing closer attention. Kioxia’s October 30 results offered an example. The NAND producer reported quarterly revenue of about ¥2.48 trillion, while its outlook implied approximately ¥2.5 trillion for the following quarter. That is only a small sequential increase.

Kioxia’s briefing also described a mid-20% increase in average selling prices in the reported quarter, while explaining that its next-quarter outlook involved shipment timing and currency effects. The figures did not demonstrate an industrywide price collapse. They did make it harder to assume that every quarter would resemble the extraordinary growth already delivered.

For shareholders, the difference matters. A business can remain highly profitable while the expectations supporting its stock price become less generous.

Third, investors had to consider where future supply would come from. China’s CXMT announced mass production of a new DRAM technology platform in September. Micron’s own annual report warns that investment by competitors, including Chinese manufacturers, could increase supply and pressure selling prices.

That is a credible competitive risk, but the products still matter. Progress in server DRAM is more directly relevant to Micron than to SanDisk’s NAND business. It also does not, by itself, prove that a competitor can immediately replace the most advanced HBM products.

The broad worry was that today’s shortage could eventually become a more competitive market. The timing, scale and product quality of that extra supply remain critical.

How Can Memory be Scarce While Memory Stocks are Falling?

Different buyers are experiencing different markets. An AI data centre may urgently need high-performance memory or enterprise storage, while a smartphone maker struggles to pass higher component costs on to consumers.

TrendForce’s September 30 outlook illustrated the split. It projected fourth-quarter contract-price increases of 10-15% for conventional DRAM and 15-20% for NAND flash, while highlighting stronger conditions for enterprise SSDs and pressure on consumer applications. These were forecasts, rather than completed-quarter price outcomes.

A shortage is helpful to suppliers only up to a point. If higher memory costs make a phone or laptop too expensive, buyers may postpone an upgrade. Device makers may then cut production or reduce the memory and storage included in each device.

That is one way pricing strength can eventually weaken demand. A chip supplier’s good fortune can become its customer’s problem, and that customer’s problem can return as smaller orders.

There is also a difference between spot prices, which reflect more immediate transactions, and contract prices negotiated with customers. Weakness in one cannot automatically be applied to the other, particularly across different products.

The useful question is therefore more specific than “Are memory prices falling?” It is: which prices, for which products, and what does that mean for the company’s actual sales?

What do SanDisk’s and Micron’s Latest Earnings Actually Show?

The latest releases showed substantial business growth. SanDisk’s October 29 results reported revenue of $11.03 billion, up 378% year on year. Data-centre revenue reached $4.80 billion, rising 61% from the previous quarter, and its non-GAAP gross margin was 85.1%. It forecast revenue of $12.2-12.6 billion for the following quarter.

Those numbers also show why the market has so much to debate. Profitability was exceptionally high, and data centres, though growing rapidly, still accounted for less than half of quarterly revenue.

Micron’s September 30 release reported fiscal fourth-quarter revenue of $54.23 billion, compared with $11.32 billion a year earlier. Its non-GAAP gross margin was 87%, and it forecast fiscal first-quarter revenue of $61.5 billion, plus or minus $1.5 billion.

The latest company outlooks therefore still anticipated further revenue growth. The unresolved issue is how much profit these businesses can sustain beyond the next quarter, and what shareholders should pay for it.

A stock market can begin marking down future earnings well before the current income statement looks weak. It can also overestimate how severe the eventual slowdown will be.

Could This be a Buy-the-Dip Opportunity?

There is a business case to examine, but a large percentage fall cannot establish it on its own. The optimistic case needs continued demand, durable customer commitments and profits that remain strong after the excitement fades.

SanDisk has been trying to make its business more predictable. At its August investor day, it said its New Business Model agreements included committed volumes, minimum financial guarantees and structured pricing. Agreements with eight customers were expected to cover about half of its fiscal 2027 bit volumes and roughly two-thirds in fiscal 2028.

That gives shareholders something more concrete to assess than a general prediction of rising AI demand. Micron has also pointed to strategic customer agreements as support for confidence in future performance.

There is support for the longer demand argument elsewhere in the industry. In its October 28 results, SK hynix said it expected demand to exceed supply through 2030, citing manufacturing complexity and the time required to build facilities. That is management’s outlook, rather than a guaranteed industry outcome.

If spending is mainly postponed, contracts hold and advanced memory remains difficult to supply, the recent falls could eventually look excessive. The more cautious case is that current profits depend on unusually favourable prices that will prove harder to sustain.

Consumer demand may stay weak, new supply may arrive, or AI customers may moderate expansion because their financial returns disappoint. Long-term agreements can reduce some uncertainty, but their protection still depends on the terms, the products covered and the customers’ ability to honour their commitments.

The distinction worth watching is between orders arriving later and orders disappearing, and between slower price increases and falling realised prices. Those outcomes would have very different consequences for earnings.

What Would tell us Whether the Memory Story is Improving?

Four developments would help separate a temporary scare from a lasting deterioration.

First, actual spending and construction. Are major customers continuing to fund projects, obtain power and bring data centres into service? Announced budgets become more useful evidence when they turn into operating capacity and supplier orders.

Second, pricing by product. Enterprise SSDs, conventional DRAM, HBM and consumer NAND need separate attention. Company disclosures about realised selling prices matter more than applying one weak spot-market quotation to every product.

Third, earnings and cash after investment. Revenue growth should translate into cash after factories, equipment and working capital are funded. Updates to margins, customer agreements and order commitments will help show how much of the current profitability can last.

Fourth, execution and competition. For Micron, that includes clarity on the reported Taiwan construction pause. For the wider industry, it includes whether new competitors can deliver reliable products at meaningful scale.

The latest figures still describe companies benefiting from strong demand. The falling shares describe investors becoming much less willing to assume that exceptional conditions will continue.

For someone worried about the crash, the useful response is to revisit what the investment depends on. For someone tempted by the fall, it is to work out what the business could earn after the boom becomes less extraordinary.

The next chapter will be written by orders, prices and cash flow. How far the stocks have already fallen cannot answer it.