SanDisk Pitches 80% Gross Margin Target as Customers Weigh SLC-NAND for AI Agents
Published on 11/19/2026 at 14:02 | Editorial boerse-global.deSanDisk used Morgan Stanley's Asia-Pacific summit to lay out two very different halves of its story: a data-center business already delivering billions in quarterly revenue, and an emerging use case for SLC-NAND that remains firmly in the testing phase.
At the center of the long-term case is a gross margin target of 80%, which the company says its pivot toward data centers should underpin. NAND memory for data centers now accounts for nearly 50% of its business, according to company figures. Whether that margin goal holds up depends in part on how hard a potential supply glut presses on NAND prices.
Long-term contracts carry the margin ambition
SanDisk is leaning on long-term agreements with major cloud providers to make the profitability math work. In the company's assessment, those contracts should deliver at least the targeted gross margin each year over their term. That ties the build-out of the data-center business to a concrete expectation for its earning power.
Demand stretches beyond that horizon as well: customers have already placed advance payments for memory deliveries running from 2029 to 2031, according to SanDisk.
Should investors sell immediately? Or is it worth buying SANDISK?
The company put growth in data-center NAND revenue at 1,700% year over year and expects data centers to make up more than half of the total NAND market by the end of 2026. That is a forecast for the market as a whole, not a target for SanDisk's own revenue mix.
Frequent rewrites open a new front
Small write operations could turn into large business. Customers are evaluating SLC-NAND for applications in which AI agents overwrite data frequently, the company said at the summit. The evaluations are not orders.
SanDisk sees potential additional demand of dozens of exabytes over the coming years. In certain applications, SLC-NAND could even replace DRAM, in the company's view — not a forecast for the DRAM market overall, but a matter of individual tasks where customers are testing which memory suits repeated writes.
The technical lever comes with a catch. According to SanDisk, existing NAND can be run in SLC mode without developing new manufacturing processes or memory chips. The trade-off is a lower number of memory bits per wafer. In other words, the switch requires no new chip development but eats into the memory volume a wafer can yield — which makes the customers' decision more important than the mere prospect of extra demand.
At the October 29 earnings call, the idea was still in architecture review and was not included in revenue. Today's statement sharpens the picture of what customers might use the memory for, but it does not yet prove a business of the scale SanDisk sketches will materialize.
The numbers already on the table
Unlike the possible AI-agent demand, the rise of data centers is already measurable. In the first quarter of fiscal 2027, SanDisk booked $4.8 billion in data-center revenue, up 61% from the prior quarter. The figures and the SLC comments tell two separate stories: one about memory already sold, the other about how demand could shift if AI agents overwrite data more often.
Management guided for revenue of $12.2 billion to $12.6 billion for the current quarter. The forecast rests to a significant degree on rising shipments rather than on pricing alone — a distinction that matters more for sizing up the stock than assuming every bit of extra memory demand must show up mainly in higher prices.
SANDISK at a turning point? This analysis reveals what investors need to know now.
GF Securities warned on Tuesday of a possible NAND oversupply. The research house expects flat prices in the first quarter of 2027 and declines in the quarters that follow. Against that backdrop, the long-term supply agreements matter especially to investors: the key question is whether they carry the margins SanDisk expects even when NAND prices in the broader market soften.
What the stock hinges on
Infinity Curve considers SanDisk undervalued but flags risks, among them a possible capacity build-out by competitors and dependence on the Kioxia partnership. SanDisk itself says it has no aggressive capacity expansion planned.
That makes the SLC drawback relevant for investors: if bits per wafer fall, the additional business cannot be judged on expected memory volume alone. For now, the data-center business SanDisk already reports is what counts. AI agents offer a plausible explanation for why demand could keep shifting. Whether they become an extra growth driver will only be settled once customer evaluations turn into orders.
Ad
SANDISK Stock: New Analysis - 19 November
Fresh SANDISK information released. What's the impact for investors? Our latest independent report examines recent figures and market trends.
