SanDisk

Sandisk hosted a bullish investor day on Aug. 13, focusing on long-term secular demand trends for NAND memory, AI, and long-term customer contracts. Management provided attractive long-term targets, including double-digit revenue growth through 2030 and 80% non-GAAP gross margin.

Why it matters: Sandisk continues to preach that this time is different for NAND, and the current upcycle will be structural and durable. We disagree and expect commodity supply and demand market dynamics to create volatile cycles into the long term. Demand isn't a question to us—we expect memory needs to rise durably in both personal devices and data centers. For memory, it comes down to the ratio of supply to demand, which governs prices. If supply exceeds demand, it creates pricing compression that can hurt sales and margins. In our view, long-term customer agreements hedge downside cycle risk, rather than preventing it altogether. We observe less than 20% of Sandisk's contract values being financially guaranteed and believe customers hold the power to amend terms if spot market prices fall.

The bottom line: We maintain our $1,000 per share fair value estimate for no-moat Sandisk. Our long-term thesis for cyclicality in the NAND market is unchanged, and we continue to worry about a downcycle beginning in the 2029 timeframe. Shares look overvalued to us. Shares are up nearly 7 times year to date as rising NAND prices have led Sandisk's revenue and profitability to balloon. The current upcycle has led our valuation higher, too, but we can't justify the current market price given the risk of a pricing downturn toward the end of the decade. We expect superb growth through fiscal 2028 for Sandisk but expect prices to peak in early 2028 and then fall precipitously in 2029 and 2030. We believe end-of-decade profitability will be lower than current levels, but we expect a better midcycle outlook than Sandisk has offered historically.

Specifically, management rooted to long-term bit supply growth in the mid to high teens, which generally agrees with our model. Management focused on this growth being productivity-led (efficiency increases in denser memory) rather than capital spending-led. We agree here, too. Management also called for modest sequential pricing growth over the next four quarters, which fits our model. Prices remain high, but we see a tapering in the increases quarter over quarter.

Where we disagree with management's targets is its outlook past 2028. Sandisk expects mid- to high-teens revenue growth between 2028 and 2030, implying flat pricing with revenue moving in line with volumes. We believe there will be downward pressure on prices starting in 2028 as a glut of new capacity comes online from Sandisk and all its other NAND peers, leading to an eventual downturn for total revenue.

This also leads our model well below Sandisk's optimistic profit targets of 80% gross margin and 75% operating margin. We model the firm outperforming these in the short term but falling well below them longer term when prices come down to earth.

Still, we believe Sandisk will be better off at midcycle than in the past. We model revenue higher in 2030 than 2026 and midcycle gross and operating margins in the low 50% range and low 40% range, respectively, compared with historical levels closer to 30% and 15%.

Finally, Sandisk outlined updated capital allocation targets. They are, in order: to invest organically in growth, to maintain a net cash balance sheet, and to return 100% of free cash flow to shareholders. We agree with these priorities but believe the cash return goal will be tough to achieve given the mass free cash flow generation we expect in the short term.

Our fair value estimate for Sandisk is $1,000 per share. Our valuation implies a fiscal 2027 adjusted price/ earnings multiple of 4 times and an enterprise value/sales multiple of 3 times. The greatest driver of our valuation is cyclical market pricing for NAND chips.

We forecast highly volatile cyclical results for Sandisk over the next five years, led by pricing cyclicality. After 120% pricing growth in fiscal 2026 and 175% revenue growth, we expect astronomical pricing growth again in fiscal 2027, more than doubling. This leads to our 150% revenue growth forecast in fiscal 2027. AI infrastructure demand creates massive demand, while supply buildouts occur much more slowly. As a result, prices climb against a relatively fixed amount of supply. In fiscal 2028, we expect prices to peak as a glut of new supply comes online across the market, and we expect a severe downturn across fiscal 2029 and 2030.

Across this pricing cycle, we expect supply growth to be more consistent. Sandisk targets bit shipments rising in the mid- to high-teens through a cycle, which it usually achieves via improvements to manufacturing efficiency. With a significant build cycle set for 2028 and 2029, we expect shipment growth to be closer to 20%. This means that results are nearly entirely driven by pricing in the medium-term.

We expect Sandisk’s gross margin profile to fluctuate significantly with market NAND pricing, as Sandisk is a fixed-cost business. We expect gross margin to reach the mid-80% range in fiscal 2027, driven by superb NAND pricing. When prices drive revenue growth, it comes at an extremely high incremental profit. Over this period, we forecast incremental gross margins around 90%, as pricing increases will exceed Sandisk’s ability to build out new production capacity. We forecast compression during a 2029 and 2030 downcycle, but we still forecast Sandisk to hold gross margins above 50% in fiscal 2031, well above our historical view of midcycle, closer to 30%.

We believe operating margin will rise and fall primarily with volatile gross margin and pricing. We forecast nonGAAP operating margin peaking above 80% in fiscal 2028, before falling back down to the 40% range in fiscal 2031, which remains exceptional compared with Sandisk’s historical levels before the AI boom in the midteens.

Our forecast is rooted in our long-term expectations for cyclicality, but we believe that Sandisk will exit the current AI-driven cycle much better off than before. In fiscal 2031, we expect more than $25 billion in revenue (compared with $7 billion in fiscal 2025), gross margin above 50% (30% in fiscal 2025), and operating margin above 40% (9% in fiscal 2025). The root of these higher midcycle estimates is that prices fall from current lofty levels but remain above NAND’s historical trendline of annual declines, creating structurally improved results.

Our model comes in mostly below Sandisk’s lofty long-term targets as given at its 2026 investor day. We agree with the firm’s long-term goal of mid- to high-teens bit supply growth, but our expectations for pricing cyclicality put us well below management’s 2028 to 2030 goals for revenue growth and profitability. Sandisk expects 80% non-GAAP gross margin and 75% non-GAAP operating margin long term. We model that in the short term during the AI upcycle, but we come in well below it by 2030. Similarly, we model well below the target for mid- to high-teens revenue growth between 2028 and 2030, given our expectations for pricing to compress after 2028.

We don’t believe Sandisk holds an economic moat. We view flash memory chips as commodities that don’t command any pricing power, instead being governed by market supply/demand dynamics that determine pricing. Furthermore, flash chip manufacturing is incredibly capital-intensive, which weighs down returns on invested capital. We believe Sandisk can earn positive economic returns for short periods in a strong market cycle, but don’t believe these are durable over the long run.

NAND flash chips are commodities in our view and we see no pricing power for Sandisk. Every flash chipmaker will tout different specifications as advantages over the competition (layer counts, memory density, and so on) but we see virtually no differentiation between competitors. That isn’t to say flash memory isn’t complex, but that the seven major suppliers all match each other’s development stride for stride. Additionally, we don’t see it as complex enough to ward off new entrants. We point to YMTC’s entrance to the market and quickly taking share as a prime example of using a large capital infusion to rapidly catch up technologically. This technological fungibility results in market-driven pricing dynamics. When demand slows versus producer expectations and there’s an oversupply of chips in the market, prices fall and drive down supplier revenue and profits. We saw this most recently with the precipitous downturn in 2023. This lack of pricing power and subsequent cyclicality erodes our confidence in durable economic returns for Sandisk.

Flash chipmaking is also a capital-intensive process, which weighs on Sandisk’s ability to extract economic returns. We estimate Sandisk committing roughly 10% of revenue toward capital expenditure (between cash capital expenses and loans to its joint venture with Kioxia) over the course of a cycle. We also expect it to continue investing more than 10% of sales in research and development over our forecast. Combined, this high level of organic investment weighs on Sandisk’s economic returns. We don’t expect Sandisk to earn cumulative economic profits over the course of a cycle. Significant capital investment alone isn’t enough to erode a moat—we point to vertically integrated logic chipmakers that offer attractive and consistent returns on invested capital as a key example. However, when significant investment requirements are paired with commodity pricing, low profitability, and volatile cyclicality, we don’t believe in durable economic returns.

Finally, we see competitive dynamics differing very little across Sandisk’s end-market exposures. While we see better profitability in enterprise SSDs that sell into data centers, we view these as similarly fungible across flash competitors. In addition, Sandisk has a relatively low share of these higher-margin enterprise drives and is much more concentrated in PC, mobile, and consumer markets, which we believe are the most commoditized.

Robust AI infrastructure demand is driving incredibly tight NAND supply, massive pricing growth, and terrific sales and profitability. If AI demand growth remains rapid, NAND supply could stay tight through 2030, carrying Sandisk to new highs

Sandisk is investing to raise its market share in enterprise SSDs for cloud infrastructure, which we see as a higher-profit business.

New AI model architectures that utilize higher NAND memory capacity could further augment the demand for NAND in an AI era and boost results even further.

Sandisk sells commoditylike NAND flash memory chips and offers little differentiation or pricing power, exposing it to volatile cyclicality.

Sandisk faces steep and volatile risk from longer-term NAND market cyclicality. If additional supply coming online creates a pricing crash, Sandisk risks steep declines to sales and profits.

Sandisk holds the lowest share among peers in the higher-profit, highest-growth enterprise SSD segment of the market.