Sandisk's $8 Billion Quarter Looks Great-4 Questions Decide if the Margin Boom Is Real

Sandisk's Q4 bar is high after a huge prior beat

Sandisk is now asking investors to underwrite $7.75 billion to $8.25 billion in Q4 revenue and $30.00 to $33.00 in non-GAAP EPS. After a quarter like the last one, that raises the bar significantly.

The bar is high because Sandisk previously delivered $23.41 non-GAAP EPS versus a $14.36 estimate, a 63% beat. The Street is also already leaning bullish, with more than 23% upside in the average analyst price target. At that level, a merely good quarter may not be enough; investors are looking for proof.

Why this report matters now

Sandisk reports on August 5, 2026, then holds Investor Day on August 13, 2026. That compressed window gives investors a chance to test whether the company is building a durable profit engine around datacenter demand, mix shift, and pricing - or whether expectations have run ahead of the evidence.

Test 1: Did Q3's profit surge come from better customers and pricing?

Sandisk's third quarter was explosive. Revenue jumped 97% sequentially, and datacenter revenue rose 233%. Management also said outperformance was driven by a mix shift toward higher-value customers and higher pricing. That is the core bull argument: the company was not just benefiting from a strong market, but selling into better ends of the business at better economics.

The key question is whether that mix shift was durable. If consumer demand was driving most of the upside, the story would be more fragile. But the stronger signal was that the datacenter end of the portfolio accelerated sharply alongside better pricing.

Test 2: Are the new datacenter deals sticky enough to support margins?

A mix shift can reverse. Contracts backed by customer commitments are harder to dismiss. Sandisk ended Q3 with three signed New Business Model agreements and added two additional NBM agreements in Q4. Management has described those arrangements as multi-year customer engagements backed by firm financial commitments.

If those deals are becoming a repeatable source of demand, the story changes from one great quarter to a more durable growth model. That is the cleanest way for margins to look less cyclical and more structural.

Why BiCS8 matters to the story

There is also a product-ramp test inside the narrative. Sandisk said earlier in the year that BiCS8 technology would account for the majority of bit production exiting fiscal 2026. At the same time, the customer signal still sits at the qualification stage, not in a disclosed hyperscaler revenue breakout. Still, two hyperscalers in qualification is the kind of incremental traction investors want to see before the numbers fully arrive.

Test 3: Did operating leverage really improve in Q3?

The real question is not whether revenue can jump again. It is whether the expense base stayed small enough for those gains to flow through to earnings.

In Q3, revenue rose 97% sequentially to $5.95 billion. Management also said non-GAAP operating expense rose just 8% to $448 million. That is an important sign: the earnings expansion did not require management to spend freely. Instead, fixed costs were spread across a much larger revenue base.

Test 4: Can management credibly defend the Q4 expense range?

For Q4, management outlined $480 million to $500 million in non-GAAP operating expense against $7.75 billion to $8.25 billion in revenue. At the top end, that works out to roughly 6% operating expense against sales. If that pattern holds, the margin story starts to look more structural and less like a one-quarter spike.

The split between bulls and bears is not really about whether the expense ratio looks low. It is about whether future spending supports durable customer commitments, product ramp, and datacenter demand - or simply reflects the cost of maintaining a stronger cycle.

What management needs to show on the call

Investors should listen for three things:

• Evidence that the expense range is tied to higher-value work, not just a favorable scaling effect.
• More concrete detail on how two additional NBM agreements translate into visible demand.
• Progress markers on the hyperscaler file, building on the earlier two hyperscalers in qualification.

What would confirm the story - and what would weaken it

After such a huge prior beat, the next move higher likely comes from confirmation rather than another blowout quarter. With results due August 5, 2026 and Investor Day on August 13, 2026, investors have a short window to separate real business progress from a setup that may have peaked.

Signs that would strengthen the case

• More concrete detail on two additional NBM agreements and how they support visibility into demand.
• Further evidence behind two hyperscalers in qualification.
• Commentary on improving NAND pricing trends that sounds sustainable rather than purely tactical.
• A spending narrative that links investment to higher-value customers and durable margin support.

Warning signs

• Vague language on mix without evidence that new business-model deals are becoming repeatable.
• Investor Day that stays light on detail when the market is focused on long-term agreements.
• Signals that higher operating expenses are needed just to sustain the cycle.

The positioning takeaway is simple: after the prior 63.02% EPS beat, this is no longer a story about surprise upside. It is a test of whether Sandisk has built something durable.