SanDisk After a 3,000% Run: 4 Tests to See If the Margin Boom Is Real

SanDisk now has to prove the margin boom is durable

SanDisk is no longer a hidden storage story. After a more than 3,000% rally since standalone trading, a 52-week intraday high of $2,354.39 on June 22, and a retreat of more than 30%, the stock is back in focus for a clear reason: investors now need evidence that the margin expansion is more than a cyclical burst. SanDisk reports on Aug. 5, then holds Investor Day on Aug. 13. Those are the near-term events that will do most of the proving.

Bulls see a business that is finally capturing higher-value demand. Bears see the old memory-cycle risk, with NAND pricing and supply tightness still doing a lot of the work. That is why these next two events matter more than another stock-point chase.

Test 1 and 2: Was Q3's breakout driven by better mix, or just stronger pricing?

The first question is not whether margins jumped. They did. The real question is whether SanDisk improved the business or simply rode a very strong part of the cycle.

What Q3 actually showed

Q3 was hard to miss: revenue of $5.95 billion rose 97% sequentially, and management said growth came from both a mix shift toward higher-value customers-led by Datacenter up 233%-and higher pricing. That is the best kind of headline, but it is also the trap. A cyclical peak can look a lot like a better business if you stop at the top line.

The test is straightforward. If better customers were doing the heavy lifting, datacenter and other high-value mix gains should show up as durable demand. If pricing and tight NAND supply were doing most of the work, the margin jump will look less like a transformation and more like a very strong spot market.

Why expectations now matter more

Before earnings, consensus sat at just $14.17 EPS, while management had already pointed to $30.00 to $33.00 non-GAAP EPS for Q4 on $7.75 billion to $8.25 billion of revenue. That raises the bar sharply. Investors are no longer asking whether the quarter was good; they are asking whether the next quarter can support that level of earnings power without leaning too heavily on favorable pricing.

Goldman Sachs' view is a useful lens, not settled fact. His CY26 estimate sits roughly 30% above Street consensus and relies on continuing NAND supply tightness, improving NAND pricing trends, and customer agreements disclosed after Micron. If that bull-case view holds, mix and contracts should do more of the work over time. If it is too early, the market will find out quickly.

Test 3: Are the new business-model contracts stacking up?

Five contracts are not a moat, but they are a useful scoreboard.

SanDisk ended Q3 with three signed New Business Model agreements and added two additional NBM agreements in the fiscal fourth quarter, for a total of five. That matters because management has described these as multi-year customer engagements backed by firm financial commitments, not one-off spot orders.

Why the count matters on an expensive stock

On a stock already up 574% year to date with a $236.8 billion market cap, investors do not need more hype. They need signs that SanDisk is locking in demand before the memory cycle does what memory cycles usually do.

Contract count is a signal, not proof. Bears are right to press on that. A signed agreement can still be small, back-loaded, or vulnerable if pricing rolls over. The real question is whether these deals are spreading across enough customers and segments to make earnings less dependent on the spot market.

That is why the Aug. 5 call and Aug. 13 Investor Day matter. Investors need to hear whether five is the start of a steady pipeline or just a one-quarter bump. SanDisk reports on Aug. 5, then holds Investor Day on Aug. 13.

Test 4: Can the product roadmap keep the story grounded?

This is where common sense helps. If the product is valuable, customers will keep buying it. If it is only convenient, they will shop around when supply eases.

SanDisk says it is sampling BiCS10 1Tb TLC 3D NAND flash memory, with the company saying the product is pushing density, power efficiency, and performance to support data-intensive workloads. Sampling is early-stage evidence, not shipped revenue. But it is still the kind of operational proof investors want to see before paying up for a sturdier business model.

What to watch

Bull case: - Management shows a path from five agreements to a larger backlog. - Customers expand commitments quarter after quarter. - BiCS10 sampling turns into real product adoption and mix support.

Bear case: - Agreements are small or uneven enough that five does not materially change the risk profile. - Revenue still leans too heavily on pricing and tight supply. - Product news stays theoretical because sampling does not lead to adoption quickly enough.

If contracts keep stacking and the product line stays credible, the margin story gets harder to dismiss as pure cycle luck. If not, investors should assume the business is still telling them more about the market than about itself.

What would make the stock work from here

With results due on Aug. 5 and Investor Day on Aug. 13, the stock only works from here if management passes a simple checklist. Investors are no longer paying for a vague AI-storage story. They are testing whether the margin boom is becoming a sturdier business.

The checklist for the bull case

• Durable mix improvement. Management needs to show the mix shift toward higher-value customers is holding, not just flashing once.
• More detail on long-term agreements. Investors are expected to focus heavily on long-term agreements. SanDisk should explain how many are now locked in, how firm the commitments are, and whether customers are deepening them.
• Investor Day should connect contracts to product adoption. The roadmap should tie demand to execution, ideally through BiCS10 sampling and clearer signs of customer traction.

The main danger signs

• Upside still comes mostly from improving NAND pricing trends rather than contracted demand.
• Management leans on strong timing while giving little substance on how many long-term agreements are really locking in future business.
• Investor Day is hype-heavy, and the company still cannot show that better mix and product progress are translating into durable customer commitments.

That is the setup now: this is a proof story, not just an upside story. If those checks are met, the stock can keep earning its reputation. If not, the market is likely to treat the recent stretch as cyclical luck and reset expectations quickly.