Sandisk (SNDK) has been one of the hottest tech stocks of the year, surging more than 650% since January. Yet even after that rally — and a 24% pullback from its 52-week high — the shares look cheap on paper, trading at roughly 8 times forward earnings versus about 20 for the average S&P 500 stock. The bigger question for long-term investors is whether the AI storage boom driving those numbers can keep lifting the stock well beyond the broader market.
The bull case rests on a data center business that has gone from negligible to enormous in little more than a year. Datacenter revenue jumped from $325 million in fiscal 2024 to $960 million in fiscal 2025, and then to $5.2 billion in fiscal 2026. Total revenue hit $20.2 billion last fiscal year, up 175% year over year.
Margins tell an even more dramatic story. Gross margin climbed from 16.1% in fiscal 2024 to 71.5% in fiscal 2026, and reached 84.6% in the fiscal fourth quarter alone. Management is guiding for fiscal 2027 first-quarter revenue of $10.3 billion to $10.8 billion and adjusted EPS of $44 to $46.
At an August investor day, Sandisk laid out a model through fiscal 2030: revenue growing at a mid-to-high teens annual pace, adjusted gross margin near 80%, and adjusted free cash flow around half of revenue. Holding the fiscal 2027 revenue midpoint flat for four quarters implies about $42 billion for the year; compounding 15% annually for three years puts fiscal 2030 revenue near $64 billion and free cash flow around $33 billion. At today’s roughly $260 billion market capitalization, the stock trades at only about 8 times that projected cash flow.
The weak spot isn’t demand — it’s the assumption that today’s economics persist. Sandisk’s gross margin was just 16% two years ago, and the company posted net losses in both fiscal 2024 and fiscal 2025. Even in fiscal 2026, profits arrived late in the year: roughly two-thirds of the fourth quarter’s sequential revenue growth came from higher prices. In effect, management’s model asks the pricing power enjoyed over just a couple of quarters to hold for four more years.
Long-term supply agreements with eight customers — including committed volumes and minimum financial guarantees — offer some cushion, covering about half of expected supply in fiscal 2027 and roughly two-thirds in fiscal 2028. But those contracts have yet to be tested through a memory downturn, and a guaranteed minimum is not the same as an 80% gross margin.
If the model holds, valuing $33 billion of fiscal 2030 free cash flow at 12 to 15 times implies a market capitalization of $400 billion to $495 billion, or roughly $2,700 to $3,400 per share. That’s a 50% to 90% total return over about four years — about 11% to 17% annually. If a normal memory correction cuts free cash flow in half to $16 billion, a 10 to 12 multiple would value the company at $160 billion to $190 billion, or roughly $1,100 to $1,300 per share — 25% to 40% below where the stock trades today.
A middle outcome — shares in the low $2,000s — would still amount to a below-market return from current levels. The stock already appears to price in a largely successful execution of management’s plan.
The Federal Reserve raised rates last week and signaled further hikes could be ahead. If higher borrowing costs prompt companies to dial back AI-related capital spending, demand for Sandisk’s memory and storage products would feel the impact quickly, given how much of the stock’s valuation depends on continued hypergrowth. Sandisk is arguably executing as well as any memory company in recent history. But at current levels, investors are being asked to underwrite four years of peak economics for a business historically prone to sharp cyclical contractions.
