SanDisk's Cheap Multiple Hinges on Whether NAND Margins Outlast the Building Boom

SanDisk's valuation case rests on a single, uncomfortable question: are the extraordinary profits of the past year a durable feature of the business, or a snapshot of a shortage that new factories will eventually erase? Wall Street's answer is split, and the stock now trades on how long the company can keep pricing power intact.

Cantor Fitzgerald's C.J. Muse frames the recent semiconductor pullback as a correction within the ongoing cycle and advises investors to start drawing up a shopping list. His math is striking. On 2028 earnings estimates, SanDisk's price-to-earnings ratio comes to 2.6, and he projects cumulative free cash flow of USD 82 billion by then. That is a deeply confident view of the company's ability to convert the memory boom into spendable cash.

The appeal of the calculation lies in the link between multiple and cash generation. SanDisk looks cheap from that vantage point as long as the forecast earnings actually arrive as cash flow. The fault line runs through the same spot: the bullish case assumes high margins survive beyond the immediate demand surge.

A Quarter That Explains the Optimism

Results published roughly two weeks ago show the starting position. In the quarter through early October, SanDisk booked revenue of USD 11.0 billion and an adjusted gross margin of 85.1 percent. Figures like those explain why analysts are willing to credit the company with enormous future cash generation. Operating cash flow for the first quarter of fiscal 2027 reached USD 6.36 billion.

The risk is treating an exceptionally profitable year as the new normal. NAND memory is a business where overcapacity and price wars are recurring hazards, and they can turn an apparently cheap earnings multiple expensive in a hurry.

Demand-side evidence arrived today from CoreWeave, whose results reinforced hopes for storage consumption. The AI cloud provider reported revenue of USD 3.66 billion, up 168 percent year on year, with a backlog of USD 132.5 billion. The numbers pushed memory makers back into investors' sights. For SanDisk, data centers are a key growth engine: segment revenue climbed 61 percent quarter on quarter to USD 4.8 billion, and management expects data-center bit demand to rise 40 percent in 2027.

The Bottleneck May Move From the Chip to the Construction Site

The larger trend behind the valuation question stretches well beyond the memory industry. Building AI infrastructure takes data centers that require years to complete. Brent Tomlinson, a manager at Kroll, warned Business Insider today that technological progress could erode the economic value of installed hardware before it has even been paid off.

For SanDisk, the demand for memory that goes into this infrastructure is a tailwind — provided the projects are built on time and put to productive use. Delays that shift expected demand into a later period argue against extrapolating profits in a straight line. The quality of the demand therefore depends partly on construction schedules and permits. UBS estimates that around 30 percent of the US data-center capacity planned for 2028 could be at risk from construction pauses — a figure that matters to SanDisk because data centers accounted for 43.5 percent of its revenue in the latest quarter.

That makes the stock a wager on two clocks running at different speeds. Memory prices can move in weeks; infrastructure projects have long lead times. SanDisk's high profitability is an advantage as long as the two stay in sync. If construction slips, the durability of long-range earnings estimates becomes the deciding factor.

Contracts Must Outlast the Cycle

Against that uncertainty, SanDisk can point to multi-year agreements with floor prices totaling USD 88.9 billion. Management expects these to support a gross margin of roughly 80 percent. For investors, that is a weighty argument — if the arrangements deliver the price and volume stability they promise. They give the earnings outlook a contractual foundation and make the valuation thesis sturdier than a simple extrapolation of rising memory prices.

The supply side sets the timing. BofA expects a memory supercycle running through 2028, a call that favors SanDisk so long as demand keeps absorbing new production. Fresh capacity, though, needs lead time: Micron's Singapore fab is scheduled to begin wafer production in the second half of 2028, while SK hynix's M17 plant is due to start in early 2029.

Warnings circulating today target the assumption that today's exceptional profitability persists. A model cited in media reports illustrates the risk: if the operating margin fell from 73 percent to 10 percent, the price-to-earnings ratio could climb to 100. That is a scenario, not a forecast for SanDisk's actual business.

The yardstick, then, is the combination of contract quality, margin stability and free cash flow. Cantor's optimistic arithmetic has substance if SanDisk profitably delivers its contracted volumes and data-center buildout carries the demand. Under those conditions the low multiple is a compelling argument; with weaker prices or delayed projects, the earnings base would have to prove itself all over again first.