Continue Reading With Our 7-Day Free Trial
Sandisk reports fiscal fourth-quarter and full-year results after the U.S. market closes on August 5, with Wall Street expecting revenue of roughly $8.39 billion and normalized earnings of $34.52 per share. This Sandisk earnings preview begins with the obvious question: can another quarter of extraordinary NAND pricing, enterprise SSD demand, and AI infrastructure spending restart a stock that has fallen sharply from its June peak?
That framing is understandable, but it may be too narrow. Sandisk has already guided to revenue of $7.75 billion to $8.25 billion and a non-GAAP gross margin of 79% to 81%, so investors are not merely waiting for strong numbers. They are waiting for evidence that these numbers can persist.
The immediate quarter will determine whether management has retained pricing momentum. The next two quarters will determine whether Sandisk has actually changed the economics of the NAND industry—or merely reached the most profitable point of another memory cycle.
Sandisk Earnings Preview: The Priced-In Scenario
Wall Street is already modeling a result above the top end of Sandisk’s previous guidance. Consensus revenue of $8.39 billion implies that stronger pricing, improved product mix, and enterprise demand continued after fiscal Q3 revenue nearly doubled sequentially.
Margin expectations are even more demanding. Management previously projected a non-GAAP gross margin of 79% to 81%, following 78.4% in the third quarter. Investors therefore appear prepared for another quarter in which pricing power overwhelms the traditionally cyclical economics of NAND manufacturing.
The narrative supporting those expectations is familiar. AI infrastructure customers need more high-performance flash, enterprise SSD supply is constrained, and Sandisk can direct production toward customers willing to sign longer commitments and pay for assured access.
Yet the Sandisk earnings preview cannot stop at the reported gross margin. The market is pricing continuity, not simply another exceptional quarter. A backward-looking beat may matter less than management’s description of pricing, contract economics, and customer demand entering fiscal 2027.
The Metric That Actually Matters
The core swing factor is Sandisk’s forward gross-margin floor—the level profitability could hold after the most aggressive phase of NAND price increases begins to moderate. Investors already know the company can generate remarkable margins when supply is scarce and customers are competing for premium products.
What remains uncertain is how much of that profitability has become contractual. Sandisk has signed five multi-year supply partnerships covering more than one-third of expected fiscal 2027 bit shipments, with financial guarantees exceeding $11 billion. The first three agreements represented approximately $42 billion of minimum contractual revenue, although pricing contains both fixed and variable elements.
The key mechanism is straightforward. If multi-year agreements preserve attractive pricing while guaranteeing volume, Sandisk’s earnings become less exposed to quarter-by-quarter NAND negotiations. That would lower perceived cyclicality, improve cash-flow visibility, and potentially support a higher valuation multiple even after near-term earnings estimates peak.
This is why the Sandisk earnings preview hinges on management’s language around the margin floor rather than the reported peak. A sustainable margin structure would change what investors believe the business is; a temporary pricing spike would only change what it earns today.
The surprise could come from how management defines “sustainable.” Near-term contract pricing is reportedly more fixed, while later years contain greater variability. That creates protection today but leaves open the question of how margins behave if market pricing weakens in 2027 or if customers renegotiate future commitments around lower industry benchmarks.
What An Upside Surprise Would Look Like
An upside surprise would require more than revenue exceeding consensus. Sandisk would need to show that enterprise SSD demand is absorbing additional supply without weakening price, while providing guidance that keeps gross margins near current levels for at least the next quarter.
The strongest signal would be evidence that additional customers have entered multi-year agreements or that contracted coverage is moving meaningfully above one-third of fiscal 2027 bits. Management previously suggested that coverage could exceed 50%, which would make future output easier to plan and reduce the need to clear inventory through aggressive spot-market pricing.
A stronger-than-expected Stargate QLC ramp would reinforce that outcome. Sandisk’s recent data-center acceleration was driven primarily by TLC enterprise SSDs, so meaningful QLC revenue would broaden the AI storage story beyond performance-intensive compute workloads into high-capacity storage and data-retention use cases.
Under that scenario, the Sandisk earnings preview would shift from a shortage trade to a business-model story. Investors who reduced exposure during the summer decline might focus less on whether earnings are peaking and more on whether Sandisk is building a recurring, contract-backed earnings base.
The psychological shift would be from “sell the peak” to “recalculate the floor.” That distinction matters because the stock’s recent volatility suggests positioning is no longer uniformly bullish, even though operating estimates have continued moving higher.
Where The Thesis Could Break
The downside scenario does not require a conventional earnings miss. Sandisk could exceed consensus revenue and EPS yet disappoint investors if the result depends on unusually favorable spot pricing that management cannot confidently extend into the next quarter.
Gross-margin guidance would be especially sensitive. A material step down from the current 79% to 81% range could imply that pricing gains are slowing faster than product mix is improving. It could also raise questions about whether multi-year contracts protect demand more effectively than they protect profitability.
Investors may also be underestimating the importance of bit shipments. Sandisk’s fiscal Q3 shipments declined sequentially as the company built inventory for QLC launches and new customer commitments. That was strategically reasonable, but the fiscal Q4 report needs to show that inventory is converting into high-value revenue rather than accumulating ahead of slower demand.
The analysis must also account for the uncertainty within AI architecture. The industry has not fully settled how inference workloads will divide data between DRAM and SSDs. Longer context windows, retrieval-augmented generation, KV cache, and autonomous agents support higher storage intensity, but the precise economics remain dependent on latency requirements, cache-hit rates, and storage duration.
The most damaging outcome would be evidence that earnings are rising faster than their durability. In that case, investors could compress the multiple despite strong reported growth because the market would begin discounting a sharper normalization in fiscal 2027.
What Matters After The Print
The first post-earnings signal will be management’s outlook for NAND pricing through the end of calendar 2026. Supply discipline has supported the current cycle, but the sustainability of that discipline will depend on whether producers continue prioritizing node transitions over aggressive greenfield expansion.
Sandisk plans to grow bit supply in the mid-to-high teens, largely through technology transitions. That approach should require less capital than constructing entirely new fabrication capacity, but stronger pricing inevitably creates incentives across the industry to increase output.
Competitive supply from Chinese NAND manufacturers also deserves attention. New capacity may not immediately displace Sandisk in premium enterprise SSDs, where qualification cycles, controller technology, reliability, and customer relationships matter. Over a six-to-12-month horizon, however, additional supply could pressure less differentiated parts of the market and influence contract negotiations.
The Sandisk earnings preview also extends to the company’s August 13 Investor Day. Investors will be looking for a longer-term financial model, a clearer explanation of normalized gross margins, capital-allocation priorities, the pace of the $6 billion repurchase authorization, and greater detail on high-bandwidth flash.
Most importantly, management must explain how AI agents translate into repeatable storage demand. Autonomous systems may generate and retain vastly more data than conventional chatbot interactions, but the investment case requires a bridge from theoretical token growth to actual SSD capacity deployed by customers.
The next phase will be determined by architecture and contracts, not simply by another quarter of higher NAND prices.
Conclusion
Sandisk enters earnings with one of the market’s widest gaps between trailing performance and forward expectations. As of August 4, the company traded near 48 times trailing earnings but below eight times next-12-month normalized earnings, reflecting an extraordinary forecast increase rather than a conventionally inexpensive valuation.
That makes the headline beat-or-miss calculation less useful than usual. Strong revenue and margins are already central to the consensus view; the more important question is whether management can establish a credible earnings floor once NAND pricing stops accelerating.
An upbeat outlook supported by contract coverage, enterprise SSD growth, and disciplined capacity could make the forward multiples appear more defensible. A weaker margin bridge or uncertain pricing commentary could push investors back toward the much richer trailing multiples and the historical cyclicality of memory manufacturing.
The earnings release will show how profitable Sandisk has become. The guidance, contract disclosures, and Investor Day will show whether those profits deserve to be treated as durable.
Disclaimer: We do not hold any positions in the above stock(s). Read our full disclaimer here.
