I don't care if you've never touched a NAND wafer. The numbers SanDisk just dropped—80% non-GAAP gross margins by 2028, 75% operating margins, 100% excess cash back to shareholders—aren't just a financial target. They're a declaration of war against the commodity memory cycle that has crushed every storage company for decades. The 2017 break didn't happen in NAND; it happened in the boardroom when Western Digital decided to spin off SanDisk. Now, the standalone entity is betting its entire future on AI demand and a radical shift from a heavy-asset IDM to a cash-generating franchise. Let me unpack what this really means, layer by layer, because the market only saw the +6.3% spike. I see the structural assumptions—and the risks—hiding beneath the surface.
## Context: Why Now? SanDisk, the NAND flash and SSD arm once buried inside Western Digital, officially became independent in early 2025. Its first major investor day since the split landed on August 13, 2025, with a headline: "2028-2030 Financial Targets." The numbers: high-double-digit revenue CAGR, 80% non-GAAP gross margin, 75% non-GAAP operating margin, and 100% of excess cash returned to shareholders. The stock jumped 6.3% intraday. To understand the audacity, you need to know the history. NAND is a brutal cyclical business—gross margins have swung from -10% to +40% over the past decade. Even Samsung's memory division, the gold standard, has never sustained gross margins above 60%. 80% is territory owned by software companies and GPU monopolists. SanDisk is saying it can become the NVIDIA of storage.
## Core: The Numbers and the Hidden Levers Let's break down the core assumptions that make these targets even remotely plausible.
Product Mix Shift: The first hidden lever is product mix. Pure NAND wafer gross margins typically sit at 20-40% in good times. To hit 80%, SanDisk must dramatically tilt its revenue toward high-value enterprise SSDs (eSSDs) for AI data centers. Imagine a 30TB+ SSD for an AI training cluster—priced at a premium because it delivers the throughput and reliability that hyperscalers need. The 80% margin implies that eSSD revenue will climb from maybe 20-30% of total today to 60-70% by 2028. This is a bet that AI demand for storage will continue to outpace supply, allowing premium pricing.
Depreciation Tailwind: The second lever is depreciation. NAND fabs are capital-intensive, with equipment depreciated over 5-7 years. SanDisk's existing fabs in Japan (joint ventures with Kioxia) are aging. By 2028, a significant portion of those assets will be fully depreciated, dramatically lowering cost of goods sold. If SanDisk stops building new fabs—which the 100% cash return policy suggests—the depreciation expense could drop to near zero. That alone could add 10-15 percentage points to gross margin.
Capital Expenditure Discipline: The third lever is capital expenditure. SanDisk is signaling it will not participate in the traditional NAND capacity race. Instead of pouring billions into new wafer starts, it will rely on Kioxia's capacity and focus on controller design, firmware, and brand. This "fab-lite" model frees up cash for shareholders and reduces the risk of oversupply. But it also means SanDisk is betting that its competitors (Samsung, SK Hynix, Micron) will also show restraint—a risky assumption given the industry's history of overinvestment.
Revenue Growth: The high-double-digit revenue CAGR target is predicated on the AI-driven explosion in data creation. Every AI training run generates petabytes of data. The average AI server now uses 3-5x more NAND storage than a traditional server. If AI capex continues to grow at 30%+ annually, SanDisk's revenue could indeed compound quickly. But the company must also capture market share. With only ~13% of the NAND market (and 15% of eSSD), it faces fierce competition from Samsung (35% share) and SK Hynix.
Operating Margin: The 75% operating margin target is the most aggressive piece. It implies that after R&D and SG&A, only 5% of revenue is left for non-COGS expenses. That would require massive cost cutting—likely layoffs, consolidation of R&D with Kioxia, and extreme efficiency. SanDisk's R&D spending as a percentage of revenue is currently around 12-15%. To get to 5% total operating expenses, it would need to cut R&D by half. That's a dangerous move for a company that needs to stay competitive in stacking layers and controller design.
## Contrarian: The Unreported Blind Spots Everyone is praising the cash return promise. I see three major unspoken risks.
First, the Kioxia relationship is a ticking bomb. SanDisk's 100% cash return means it will not invest in new joint venture fabs. Kioxia, which co-develops BiCS technology and manufactures the wafers, will be expected to carry the capex burden. Why would Kioxia agree to that? They will demand a larger share of the profit, or they could divert capacity to their own branded products. The alliance is already strained after the failed merger talks. If Kioxia pulls back, SanDisk loses its manufacturing lifeline.
Second, the 80% gross margin target is not sustainable in a downturn. NAND is cyclical. When AI demand slows (and it will, at some point), the oversupply will crush prices. SanDisk's high-margin eSSD contracts will get renegotiated downward. The 100% cash return policy leaves no buffer for a rainy day. If the company needs to reinvest in capacity during a downturn, it will have to cut dividends or sell equity.
Third, SanDisk is missing the HBM gravy train. High-bandwidth memory is the premium storage product for AI GPUs, with margins even higher than NAND. But SanDisk has no HBM business—that's Samsung, SK Hynix, and Micron's turf. By focusing on eSSD, SanDisk is settling for the "storage peripheral" role, not the core AI memory. This limits its top-line growth potential and makes the 80% margin target even harder to achieve without a pure pricing miracle.
## Takeaway: What to Watch Next SanDisk's financial targets are a high-stakes bet on AI demand, fab-light discipline, and a cooperative Kioxia. The next six months will be telling: watch for any signs of friction with Kioxia, such as changes in capacity allocation or profit-sharing terms. Also watch SanDisk's R&D spending—if it cuts too deep, it risks falling behind in the BiCS9 transition. Finally, the stock will react to every NAND price tick. If the cycle turns, the 80% margin dream will evaporate. I don't think the market has fully priced in the execution risk. But if SanDisk pulls it off, it will be the most impressive turnaround in storage history.
