Western Digital and SanDisk Both Beat, and Both Fell Hard

One company beat estimates. The other beat estimates and guided the coming quarter above what Wall Street had modeled. On Thursday, August 6, both collapsed: Western Digital fell 16%, SanDisk 11%. In the same session, QQQ — the Nasdaq 100 tracker — slipped 0.2%, and SPY 0.2%. This was not a market day. Two earnings reports repriced an entire sector.

The part worth explaining is not how bad the numbers were. It is how good they were.

Gaining 58 Points of Margin in a Single Year

SanDisk makes NAND — the flash memory that holds data when the power goes off. It sits inside your phone and inside the enterprise SSDs stacked in a data center. Western Digital is the other half of the old company: hard disk drives, the classic spinning-platter storage. The two separated in 2025 and have traded independently since.

The same wind carried both. AI data centers are absorbing memory and disk at once, while the manufacturing capacity serving them was planned years ago, back when today's demand was unimaginable. The gap showed up in price. By TrendForce's count, contract prices for some flash categories rose more than 100% in the first half of 2026.

SanDisk's quarter shows what that does to a P&L. Revenue of $8.97 billion, up 372% from a year ago and 51% from the prior quarter. By the company's own account, roughly one-third of that sequential jump came from bits shipped and two-thirds from price. Non-GAAP earnings came in at $39.25 per share against the company's own guidance of $30 to $33.

Western Digital's numbers are calmer but point the same way: $3.75 billion in revenue, up 44% year over year, with gross margin at 54.4% — an improvement of 13.1 points in twelve months. Cloud customers now account for roughly nine-tenths of revenue.

What Happens When Cost Stays Still

This section is the core of the story. Without it, the August 6 selloff makes no sense.

Gross margin is what remains of revenue after production cost. At SanDisk that figure went from 26.4% to 84.6% in a year. Over the same span the company's bit output grew only in the mid-teens. No new fabs opened; the increase came mostly from node transitions — getting more bits out of the same wafer.

This is why buying a memory stock is less a bet on a company than a bet on a price curve. For how that logic feeds into a multiple, see the valuation guide.

Prices Did Not Stop Rising — They Stopped Accelerating

That is what was actually sold on August 6. SanDisk guided the new quarter to $10.3–$10.8 billion in revenue. The midpoint is $10.55 billion; consensus stood at $11.15 billion. The shortfall is 5.4%.

But the size of the miss matters less than where it comes from. Two-thirds of last quarter's 51% sequential growth was price. In the new guidance, growth falls to about 18%, and management expects it to come mainly from volume, with only a "modest" contribution from pricing. Price is still rising — its contribution to a quarter just went from roughly 34 percentage points to a handful.

Industry data says the same thing. TrendForce's July 3 outlook called for third-quarter contract prices to rise 10–15% for NAND and 13–18% for conventional DRAM: still up, but far below the pace of the first half. The firm also wrote that wafer contract price increases would moderate substantially.

SanDisk has built a defense against that turn. The agreements it calls new business models carry both floors and ceilings, with a weighted average duration of more than four years. Ten signed deals carry $93.9 billion in minimum revenue; remaining performance obligations stood at $59.8 billion at quarter-end, or $91.1 billion including deals closed after the quarter. Customers have posted $16.5 billion in cash deposits and financial instruments.

That is a real shield if prices fall. But a contract with a floor also has a ceiling: if prices keep climbing, the company no longer captures all of it. The unbounded upside the market grew used to last quarter narrows as these agreements take over the book.

August 6: How the Selling Spread

The damage was not confined to the two reporters.

Western Digital closed August 5 at $519.17 and traded near $437 the next day. Micron fell to $844, Seagate to $786. That companies which reported nothing fell alongside those that did is the tell: this was not disappointment with one forecast but a change of view about a sector.

To see how that landed in the broader tape, start with the index.

On the memory side, a name that reported nothing at all shows how far the selling traveled.

Western Digital's Different Problem

At SanDisk there was at least a reason: guidance came in light. At Western Digital there is not even that. The company guided the new quarter to $4.1 billion in revenue and $4.00 in earnings per share — both above consensus. It expects gross margin to expand to 55–56%. It has supply agreements with customers running out to 2031.

And the stock fell 16%.

The only explanation left is the price itself. Western Digital had risen 202% year to date and 625% over twelve months. When a stock climbs that fast, most of the good news to come is already inside the price. What remains is an argument about how good the good news is — and in that argument, "beat estimates" is not a sufficient answer. This is the most frequently skipped part of reading an earnings report: the number can beat the estimate while the price has beaten it by more.

The Other Side

Analysts looking at the same numbers did not reach the same conclusion.

Barclays' Tom O'Malley described SanDisk as attractive on a pullback. RBC leaned the other way, warning that margins may be near a peak as price growth moderates. Both camps are reading the same report; where they part is how long today's margin lasts.

What Is Left

The memory cycle is not over. Prices are still rising, contracts are still being signed, and data center demand still runs ahead of capacity. SanDisk repurchased $4.5 billion of its own stock during the quarter and the board authorized another $14 billion — not the behavior of a company that finds its shares expensive.

What changed is not the level but the slope. For a year the market priced a world in which price increases were accelerating. It is now moving to a world in which they continue but decelerate. The distance between those two worlds cost 16% in a single session.

On the same day, JPMorgan's Jamie Dimon noted that margin debt sits at the highest level on record. On that kind of footing, a small disappointment in a crowded, heavily profitable trade produces a large move. Not everyone who sells has changed their mind; some simply cannot afford to wait.

This article draws on the companies' published quarterly results, the earnings call transcript, TrendForce's industry pricing forecasts, and market coverage published on the day. Analyst price targets are taken from news reports. It is not investment advice.