Invest1 publisher2 min readPublished
Monthly DRAM contract gains shrink from 14% to 4% at record prices
DDR4 contract prices rose 4% in September to a record $26 after a 14% gain in July, while MLC NAND rose 0.44%. Analysts tie Micron's shrinking earnings beats to multi-year supply deals, so the peak may show up in results before it shows up in contract prices.
The Investor · Invest desk
Drafted by a language model from the sources cited here and checked against its claim ledger before publication. How we use AISend a correction
What happened
- The DDR4 contract price is up about 126% since January, when it stood at $11.5.
- MLC NAND's monthly gain has fallen every month since April, from 36.29% to 9.73%, 8.72%, 4.26% and then 1.42% in August.
- DRAMeXchange said DRAM spot purchase inquiries nearly dried up before the end of September, and NAND buyers held off as some product prices weakened.
- Micron's fiscal fourth-quarter revenue of $54.23 billion beat forecasts by about 6%, against a beat of roughly 31% the quarter before.
- FnGuide's consensus for Samsung's third-quarter operating profit has slipped 4.5% in three months to about 105.6 trillion won ahead of October 8 results.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- constraint If Samsung reaches its 60% to 70% contract target, only about 30% to 40% of its memory output sells at monthly or spot prices, so further price spikes reach earnings through that remaining share.
- cost Analysts say PC and smartphone makers are near the limit of the price increases they can absorb, so consumer memory is where demand gives way first while capacity goes to AI parts.
- contradiction TrendForce's NAND forecast implies about 4.8% to 6.3% a month against 0.44% for MLC in September, so the forecast depends on AI-linked flash outrunning the consumer parts the index tracks.
Converted back into dollars, the DDR4 series runs from about $21 in June to $24 in July, $25 in August and $26 in September [2]. The chip added $3 in one month and $1 in each of the next two. MLC NAND, at $30.61, gained about 13 cents in September [8].
Both series track consumer parts. DDR4 is mainstream PC memory and the MLC chip goes into memory cards and USB drives [1][4], while high-bandwidth memory and server DRAM appear in neither. TrendForce says producers are allocating capacity to those AI parts first, keeping PC and mobile supply tight [12].
Morgan Stanley said Micron's results had beaten forecasts by 20% to 40% in each of the three quarters before this one [9]. Some analysts attribute the smaller surprise to supply agreements of three to five years, under which Micron guarantees volume to big tech customers in exchange for lower prices [10]. Samsung is reportedly trying to put 60% to 70% of its memory capacity under such contracts [13]. Under that plan, most of Samsung's output would no longer be available to sell at spot prices.
The data fit more than one ending. If the softness in consumer parts spreads to server DRAM, TrendForce's fourth-quarter forecast of 10% to 15% for conventional DRAM will miss [12]. If the forecast holds, it needs no reacceleration: 10% to 15% a quarter is about 3.2% to 4.8% a month [5], and September's 4% held for three months compounds to about 12.5% [7]. A third case is that prices keep rising and profits flatten anyway because most output is contracted. Analysts expect Samsung's profit growth to be capped on exactly that logic [16].
I think the third case is the one the numbers already show. Micron's smaller beat [8] and the cut to Samsung's consensus [11] both arrived while contract prices were setting records [1]. On this evidence the peak has come in the monthly pace of consumer memory prices and in the size of Micron's surprise. For AI server parts, the record holds only TrendForce's forecast of further increases [12].
The counter-case rests on that AI demand and on enterprise SSDs, expected to see larger price increases as AI inference grows [14]. Micron also played down the chance that the shortage ends soon [15]. If Micron's next beat returns to the 20% to 40% range of earlier quarters [9], the long-term-contract explanation for the shrinking surprises is wrong.
What to watch
- Samsung's preliminary third-quarter results on October 8: operating profit above the reduced FnGuide consensus would weaken the case that long-term contracts are capping earnings.
- Whether October's DDR4 contract gain holds near September's 4%, the pace TrendForce's fourth-quarter DRAM forecast requires.
- Whether fourth-quarter enterprise SSD contract prices move away from consumer MLC flash, as TrendForce's AI inference argument implies.