Storage stocks have fallen for four days, while contract prices are still rising. So who’s wrong? | Pre-market report on the memory sector
【Sector view】Neutral to slightly bullish: prices haven’t deteriorated, but valuations are compressing. It’s still too early to take a heavy position.
【What happened】Binance futures, 24h: $MU -1.0%, SK Hynix $SKHYNIX -2.4%, Samsung -0.3%, SanDisk +0.3%, Western Digital +2.9%, memory ETF $DRAM -1.0%. U.S. pre-market: Micron 932 (-2.2%), SanDisk 1425 (-3.9%), Western Digital 369 (-3.1%). Korean stocks closed today: SK Hynix -1.7%, Samsung unchanged. Micron closed at 953.19 yesterday, marking its second consecutive day of losses.
【Breaking down the drivers】
1. Source of the panic: China’s ChangXin Memory Technologies (CXMT) says it will launch high-density 4F² DDR5 server memory by year-end, prompting concerns that it could disrupt the supply landscape. However, Wccftech reports that its DDR5 prices are broadly in line with Samsung, SK Hynix, and Micron, with no price war underway. The threat is capacity, not low prices.
2. Fundamentals remain intact: TrendForce’s 9/30 forecast calls for conventional DRAM contract prices to rise 10–15% QoQ in 4Q26, and NAND prices to rise 15–20%; enterprise SSD price increases are accelerating further.
3. But keep two things in mind: First, price increases are slowing. TrendForce says some suppliers, constrained by long-term contract mechanisms, will see increases lag the market average. Second, consumer demand is weak. NAND wafer prices are already at historic highs, making downstream module makers more cautious about purchasing.
4. Micron’s 9/30 earnings report disclosed 26 strategic customer agreements running through at least 2030, estimated to account for more than 35% of revenue before 2030 (as cited by The Motley Fool).
【Why you can’t look at P/E alone】Memory is a cyclical industry. P/E is lowest when profits peak, so a low P/E is often a signal that the cycle is near its peak. Micron’s forward P/E is only 5.4, which sounds cheap, but the market is actually pricing in “peak profits” in advance. So I look at P/B relative to ROE, EV/EBITDA, and free cash flow yield.
【Cycle valuation, stock by stock】(stockanalysis, based on 10/13 closing prices; Korean stocks based on 10/14 closing prices)
- SK Hynix: P/B 4.0, ROE 93%, EV/EBITDA 7.1, FCF yield 8.7%. HBM share is about 50% (Counterpoint); it has the highest ROE and one of the lowest valuations.
- Samsung: P/B 2.5, ROE 31%, EV/EBITDA 5.8, FCF yield 9.8%. The cheapest, but earnings quality is inferior to SK Hynix; DRAM revenue share is 39.4% (TrendForce).
- Micron: P/B 7.8, ROE 88%, EV/EBITDA 9.3, FCF yield 5.5%. Long-term contracts lock in pricing, but its HBM share is only 18%, and a strike authorization vote by a Taiwanese union is a risk.
- SanDisk: P/B 14.0, ROE 92%, EV/EBITDA 16.7, FCF yield 5.3%. Pure-play NAND; it’s up about 11-fold in a year, with the greatest upside sensitivity and the weakest downside protection.
- Western Digital: P/B 15.5, EV/EBITDA 28.4, FCF yield 2.5%. Its ROE of 131% was inflated by a one-off gain (pre-tax profit of $9.9 billion, far above operating profit of $4.6 billion), making it the most expensive of the five.
【Bull-bear debate】Bulls: rising contract prices, long-term contracts, and accelerating enterprise SSD demand. Bears: slowing price increases, consumer demand near a peak, CXMT expanding capacity, and stocks’ huge gains over the past year (Micron +425%). I’m on the bullish side, but I’d call it “valuation compression,” not “the cycle is over.” I’d know I was wrong if the next TrendForce price report shows a QoQ decline, or if enterprise SSD price increases slow markedly.
【Catalysts】Western Digital earnings on 10/22; SanDisk, Samsung, and SK Hynix earnings on 10/29 (SK Hynix date is an estimate).
【Ranking】SK Hynix > Samsung > Micron > SanDisk > Western Digital.
【Trading approach】ETF 4H chart: 50 is key support; 52.8 and 54.7 are resistance levels above. Don’t chase until it recovers 52.8; reduce your position if it breaks below 50. Scaling in is safer than going all in.
Do you think this is a chance to get in, or is the hype peaking? Share your thoughts in the comments.
⚠️Just my personal rambling; this is not investment advice.
#存储芯片 #美股 #Semiconductor
$DRAM $MU $SKHYNIX
【Sector view】Neutral to slightly bullish: prices haven’t deteriorated, but valuations are compressing. It’s still too early to take a heavy position.
【What happened】Binance futures, 24h: $MU -1.0%, SK Hynix $SKHYNIX -2.4%, Samsung -0.3%, SanDisk +0.3%, Western Digital +2.9%, memory ETF $DRAM -1.0%. U.S. pre-market: Micron 932 (-2.2%), SanDisk 1425 (-3.9%), Western Digital 369 (-3.1%). Korean stocks closed today: SK Hynix -1.7%, Samsung unchanged. Micron closed at 953.19 yesterday, marking its second consecutive day of losses.
【Breaking down the drivers】
1. Source of the panic: China’s ChangXin Memory Technologies (CXMT) says it will launch high-density 4F² DDR5 server memory by year-end, prompting concerns that it could disrupt the supply landscape. However, Wccftech reports that its DDR5 prices are broadly in line with Samsung, SK Hynix, and Micron, with no price war underway. The threat is capacity, not low prices.
2. Fundamentals remain intact: TrendForce’s 9/30 forecast calls for conventional DRAM contract prices to rise 10–15% QoQ in 4Q26, and NAND prices to rise 15–20%; enterprise SSD price increases are accelerating further.
3. But keep two things in mind: First, price increases are slowing. TrendForce says some suppliers, constrained by long-term contract mechanisms, will see increases lag the market average. Second, consumer demand is weak. NAND wafer prices are already at historic highs, making downstream module makers more cautious about purchasing.
4. Micron’s 9/30 earnings report disclosed 26 strategic customer agreements running through at least 2030, estimated to account for more than 35% of revenue before 2030 (as cited by The Motley Fool).
【Why you can’t look at P/E alone】Memory is a cyclical industry. P/E is lowest when profits peak, so a low P/E is often a signal that the cycle is near its peak. Micron’s forward P/E is only 5.4, which sounds cheap, but the market is actually pricing in “peak profits” in advance. So I look at P/B relative to ROE, EV/EBITDA, and free cash flow yield.
【Cycle valuation, stock by stock】(stockanalysis, based on 10/13 closing prices; Korean stocks based on 10/14 closing prices)
- SK Hynix: P/B 4.0, ROE 93%, EV/EBITDA 7.1, FCF yield 8.7%. HBM share is about 50% (Counterpoint); it has the highest ROE and one of the lowest valuations.
- Samsung: P/B 2.5, ROE 31%, EV/EBITDA 5.8, FCF yield 9.8%. The cheapest, but earnings quality is inferior to SK Hynix; DRAM revenue share is 39.4% (TrendForce).
- Micron: P/B 7.8, ROE 88%, EV/EBITDA 9.3, FCF yield 5.5%. Long-term contracts lock in pricing, but its HBM share is only 18%, and a strike authorization vote by a Taiwanese union is a risk.
- SanDisk: P/B 14.0, ROE 92%, EV/EBITDA 16.7, FCF yield 5.3%. Pure-play NAND; it’s up about 11-fold in a year, with the greatest upside sensitivity and the weakest downside protection.
- Western Digital: P/B 15.5, EV/EBITDA 28.4, FCF yield 2.5%. Its ROE of 131% was inflated by a one-off gain (pre-tax profit of $9.9 billion, far above operating profit of $4.6 billion), making it the most expensive of the five.
【Bull-bear debate】Bulls: rising contract prices, long-term contracts, and accelerating enterprise SSD demand. Bears: slowing price increases, consumer demand near a peak, CXMT expanding capacity, and stocks’ huge gains over the past year (Micron +425%). I’m on the bullish side, but I’d call it “valuation compression,” not “the cycle is over.” I’d know I was wrong if the next TrendForce price report shows a QoQ decline, or if enterprise SSD price increases slow markedly.
【Catalysts】Western Digital earnings on 10/22; SanDisk, Samsung, and SK Hynix earnings on 10/29 (SK Hynix date is an estimate).
【Ranking】SK Hynix > Samsung > Micron > SanDisk > Western Digital.
【Trading approach】ETF 4H chart: 50 is key support; 52.8 and 54.7 are resistance levels above. Don’t chase until it recovers 52.8; reduce your position if it breaks below 50. Scaling in is safer than going all in.
Do you think this is a chance to get in, or is the hype peaking? Share your thoughts in the comments.
⚠️Just my personal rambling; this is not investment advice.
#存储芯片 #美股 #Semiconductor
$DRAM $MU $SKHYNIX