CXMT is reportedly preparing to expand beyond DRAM into NAND flash memory, a strategic move that could broaden its market but also raises the execution bar after an extraordinary stock-market debut. Reuters reported that the Chinese memory producer plans a 3D NAND research-and-development line at a second plant near Beijing. The project would put CXMT closer to the core market of domestic rival YMTC and increase its exposure to storage demand from data centres and solid-state drives.
From a 466% debut to a new growth narrative
CXMT shares closed their July 27 Shanghai debut at 49 yuan, up 466% from the 8.66-yuan offer price, after touching 55.03 yuan intraday. The gain lifted market capitalisation to about 3.3 trillion yuan ($488 billion at the time). CXMT raised 57.92 billion yuan ($8.6 billion), the largest mainland Chinese semiconductor IPO on record.
The size of the move needs context. Only 6.73% of the enlarged share capital was freely tradable at listing, while turnover reached about 141.1 billion yuan on the first day. A limited float can magnify both demand and volatility, so the debut return should not be treated as a clean measure of long-term value.
| IPO / operating datapoint | Reported figure | Investor interpretation |
|---|---|---|
| Offer price / first close | 8.66 yuan / 49 yuan | 466% debut gain |
| IPO proceeds | 57.92bn yuan | Major funding base for capacity and R&D |
| Free float at listing | 6.73% | Amplifies price moves and valuation risk |
| First-half revenue guidance | 110–120bn yuan | More than sevenfold expected growth |
Why NAND matters
CXMT has been focused on DRAM, while YMTC leads China’s NAND production. Entering NAND could increase CXMT’s addressable market and give customers another domestic source during a global shortage. It could also improve strategic relevance as China seeks semiconductor self-sufficiency under U.S.-led restrictions on advanced manufacturing equipment and high-bandwidth memory.
But an R&D line is not the same as qualified mass production. NAND requires different process expertise, controller ecosystems, customer testing and sustained capital spending. Investors should separate the announcement effect from evidence of competitive yields, product reliability and commercial shipments.
Earnings leverage cuts both ways
CXMT’s prospectus linked the memory upswing to AI infrastructure and expected first-half revenue of 110–120 billion yuan, with net profit of 66–75 billion yuan after a year-earlier loss. Tight supply and supplier diversification can support pricing and margins. However, memory remains deeply cyclical: weaker AI spending or aggressive capacity additions could reverse price gains quickly.
The expansion can be read positively if CXMT converts IPO proceeds into qualified NAND output without undermining DRAM execution. The bearish case is that the valuation already discounts years of growth while export controls, low free float and competition from Samsung, SK Hynix, Micron and YMTC leave little room for setbacks.
What investors should monitor next
- Formal company confirmation, project budget and timing for the Beijing NAND line.
- Wafer yields, customer qualification and the transition from R&D to volume production.
- DRAM and NAND contract-price trends through 2027.
- Changes to U.S. export controls and access to advanced fabrication equipment.
- Post-lock-up share supply, free float and valuation versus global memory peers.
Sources: Reuters, September 18, 2026; Reuters IPO report, July 27, 2026.