Written by Cassian Vance
DRAM pricing delivered one of the most extreme quarters on record, yet memory shares are trading near the top of their range and the sector sold off while the Dow rose. The gap between the commodity and the equity is where the analysis now sits.
On Monday 12 October the US market finished mixed. The Dow Jones Industrial Average rose while the Nasdaq slipped, and part of the drag came from losses in memory chip stocks even as rising oil prices weighed on the broader tape. The same day, one of the industry’s mid-sized DRAM makers reported a quarter that by any historical standard was extraordinary.
Nanya Technology told analysts on its third-quarter call that average selling prices rose by more than 60% quarter over quarter and by more than 500% year over year. Its president, Pei-Ing Lee, attributed the sales and profit gains mainly to pricing. The shares trade near the top of their 52-week range, close to a high that is several times the year’s low.
That is the paradox worth sitting with. If pricing power of this magnitude does not lift the equity, what is the market pricing?
Three answers emerge from the company’s own disclosure. First, capex. Nanya said one new fab will require roughly $16 billion including construction, clean rooms, equipment and future extreme-ultraviolet capability, with a second facility for 3D wafer packaging and testing running to about $10 billion over the long run. A business earning windfall pricing is simultaneously committing that windfall to capacity that will not produce wafers until the second half of 2027 at the earliest. Second, volume. Shipments fell mid-single digits in the quarter because inventory was depleted, and management expects bit shipments to stay flat for several quarters. Revenue is rising on price alone. Third, the cycle. Management told analysts the shortage will continue into the December quarter and that 2027 is likely to remain tight, which is precisely the setup that invites the capacity response now being funded.
The market is therefore doing something rational: valuing a peak-price year against a 2028 that will contain all of this new supply.
Where the money is going instead is instructive. The scarcity has migrated to the steps that happen after the wafer is printed. Equipment lead times for some South Korean component suppliers have stretched to 40 months, meaning a fab planned today is not equipped until 2030. DISCO reported record July-to-September shipments, up 60%. Lam Research raised its quarterly dividend 27%, payable 14 October, which is not the behaviour of a company bracing for a capital-spending collapse. Tokyo Electron’s finance chief has signalled a record year. TSMC posted record September revenue, up 54.6% year over year, and a record quarter of NT$1.49 trillion, with full results due 15 October.
The pattern is familiar from every commodity cycle. The producer that owns the scarce input captures the spot price, but it also carries the obligation to fund the next generation of capacity, and its equity is discounted for the day the price normalises. The suppliers who cannot be substituted — deposition and etch tools, test and metrology, substrates, high-voltage power equipment on 128-week lead times — collect a toll on every path to capacity without underwriting the glut.
For investors, the practical question is not whether the memory shortage is real. It plainly is. The question is which part of the value chain is being paid for scarcity rather than for expansion. Memory makers are being paid for the former and charged for the latter.
Three data points will settle it. December-quarter DRAM contract prices, which management expects to improve again. HBM4 negotiations for 2027, where the outcome determines how much conventional capacity gets diverted. And the capex lines in the next round of earnings, because every dollar committed today is supply arriving in 2028.
This article is for informational and educational purposes only. It is not financial advice, and nothing here is a recommendation to buy or sell any security. Readers should conduct their own due diligence and consult a qualified professional before making any investment decision. Past performance does not equal future results.
