SanDisk closed at $1,692.42, down ten percent, on a set of numbers in which profit and cash flow both expanded alongside sales. If that strikes you as the market making a mistake, you have not met the memory cycle.
Start with what is not in dispute. Revenue grew. Profit grew with it rather than lagging behind, which means the growth was not purchased with discounts. Cash generation expanded too, which means the profit was real rather than an accounting posture about inventory values. In most industries that combination is the definition of a good quarter and gets treated accordingly. In memory it is treated as a data point about where you are standing, and the question the market actually cares about is which direction you are walking.
Here is the thing about buying coffee from a single harvest. A spectacular lot arrives, the cupping scores are the best in years, the farm has never performed better, and the price being asked is accordingly enormous. An experienced buyer tastes it, agrees that it is magnificent, and then asks a question that has nothing to do with the cup in front of them: what is planted for next season, and how much of it. Because if every farm in the region saw the same prices and responded the same way, the glut is already in the ground. The quality of this year’s harvest is evidence. It is just evidence about the past, and the thing being priced is the next one.
Memory is the purest commodity cycle in technology and it runs on exactly that logic. NAND and DRAM producers do not compete on product differentiation in any meaningful sense, because a bit is a bit. They compete on cost per bit and on capacity timing, and capacity takes years to build and arrives all at once. When prices are strong, every producer on earth has simultaneous access to the same conclusion and the same capital markets. The investment decisions taken during a good quarter are what destroys the pricing two years later, with mechanical reliability. Everybody involved knows this. They build anyway, because the producer who sits out the upcycle loses share permanently.
Which is why strong current results can be a bearish signal rather than a bullish one. A memory company reporting expanding margins on rising volumes is telling the market that prices are good, that returns on new capacity look attractive, and that the industry’s collective incentive to add supply is currently at its peak. The ten percent decline is not a rejection of the quarter. It is the market marking down the value of future quarters by an amount that exceeds the good news in the present one.
There is a second element at work, which is what artificial intelligence demand has done to memory valuations generally. Enormous multiples have been granted on the argument that this cycle is structurally different: that data centre buildouts create a demand floor the old cycle never had, and that high-bandwidth memory has escaped commodity economics altogether. Perhaps. But a share price that has absorbed that argument in advance has a specific vulnerability, which is that it needs the thesis confirmed continuously rather than merely not contradicted. Any hint of normal cyclicality reasserting itself removes a premium that was never attached to the current earnings in the first place.
So the ten percent and the expanding cash flow are both correct, and they are answers to different questions. The results describe a company executing well into a strong price environment. The share price describes a market counting the hectares being planted. One of those is a report. The other is a forecast, and forecasts move prices.
Source: Trefis, “Is SanDisk Stock’s Fall A Bargain Or A Warning?”