After CXMT’s Listing, Do the Memory Three Have Further to Fall?
Damnang
· Damnang’s Substack
· July 27, 2026 at 06:19
· ⏱ 11 min read
| Read on Substack ↗
Summary
CXMT's IPO triggered a sharp sell-off in incumbent memory stocks, but detailed analysis shows the threat is overstated: CXMT has higher costs, lower yields, and a technology gap that prevents it from undercutting prices significantly. The correction in Samsung, SK hynix, and Micron is likely to end, with SK hynix most insulated (HBM focus) and Micron holding least China exposure.
•CXMT opened at 49.50 yuan (472% above IPO price of 8.66 yuan), giving it a market cap of 3.31 trillion yuan, the largest on A-shares; by midday it hit 54.65 yuan.
•Claimed 30% price advantage over incumbents is misleading: like-for-like DRAM pricing gap is 5-10%, and CXMT is actually 2.2% higher on a 64GB DDR5 server module in Chinese online retail.
•CXMT's production cost is >30% higher than incumbents due to larger chip area (~40%) and lower yield (below industry 85-90%); its wafer starts represent 13% of global DRAM but capacity only 6%.
•CXMT's Q1 2026 revenue surged 719% YoY, but driven largely by a 57% price increase; capacity shipped grew only 11% sequentially, showing growth is price-led not volume-led.
•CXMT allocates <2% of wafers to HBM with ~25% yield vs incumbents' 70-90%; moving more capacity to HBM would reduce ordinary DRAM supply, putting upward pressure on prices.
•Author forecasts CXMT's first-day high will mark a peak (similar to SMIC's 2020 IPO), with a large correction possible around the January 2027 lockup expiry, but no fall below IPO price.
Author argues the memory correction is likely to end after CXMT's IPO, and Micron has the least remaining China exposure (revenue fell from 12% to 7% range, exited Chinese server memory in Oct 2025).
Author argues the memory correction is likely to end after CXMT's IPO, and Micron has the least remaining China exposure (revenue fell from 12% to 7% range, exited Chinese server memory in Oct 2025). Its forward multiple of 6.42x is the highest of the three, but the threat is limited. This implies a possible rebound from recent weakness.
Risk: If US procurement excludes Chinese memory, Micron may still face competition outside North America; HBM demand dependency remains.
Author explicitly identifies Chinese equipment makers Naura and AMEC as beneficiaries of CXMT's capacity expansion: 'It makes more sense to read this news as good news for Chinese equipment makers suc
Author explicitly identifies Chinese equipment makers Naura and AMEC as beneficiaries of CXMT's capacity expansion: 'It makes more sense to read this news as good news for Chinese equipment makers such as Naura and AMEC than for the memory companies.' CXMT's doubling of wafer starts (even if capacity lags) drives equipment orders.
Risk: Dependency on CXMT's capex cycle and US export controls on semiconductor equipment.
Same explicit mention as Naura: author says news of CXMT's expansion is good for Chinese equipment makers like AMEC. AMEC produces etch and thin-film deposition tools used in memory fabrication.
Same explicit mention as Naura: author says news of CXMT's expansion is good for Chinese equipment makers like AMEC. AMEC produces etch and thin-film deposition tools used in memory fabrication.
Risk: Same as Naura; also potential technology roadblocks from restricted EUV access and slower process node migration.
This newsletter, published July 27, 2026,
features Damnang
discussing MU, 002371.SZ, AMEC.
3 trade ideas extracted by AI with direction and confidence scoring.