BlockBeats reported on July 16 that Counterpoint Research Director Hwang Minsung said a 15% global DRAM market share is the baseline for ChangXin Memory Technologies, or CXMT, to remain viable over the long term. Below that level, he said, the company could repeat the trajectory of Taiwan DRAM makers in 2008, when they failed to cover investment for next-generation fabs and eventually fell to 3% market share.
CXMT’s global DRAM bit shipment share is currently about 9%, and the target is to lift that figure to 11% by 2028.
IPO proceeds are earmarked for G5, HBM3, and capacity growth
Counterpoint said the funds raised in the IPO will be directed to next-generation G5 process technology, research and development for HBM3 high-bandwidth memory, and capacity expansion. The company plans to increase monthly wafer capacity from the current 320,000 wafers to 420,000 by 2027, then double that level by 2030 and triple it by 2035.
In its product mix, LPDDR5 and DDR5 are expected to account for about 75% of total output. DRAM products for PCs and servers have already begun shipping to global customers.
Four issues Counterpoint is watching after the IPO
Counterpoint listed four main variables to watch after the IPO:
- HBM mass-production yields and the pace of revenue ramp. The firm said this segment could benefit from expanded production of Huawei Ascend AI chips, with HBM revenue potential reaching about $2 billion in 2028.
- Whether CXMT can secure large-scale procurement contracts from major global customers outside China. Supply to Apple still depends on approval.
- Whether an escalation in trade conflict could trigger additional export controls on equipment.
- Whether the company can, over the medium to long term, reach 20% market share by bit shipments and 15% by revenue.
Neil Shah says restrictions may also force technological divergence
Neil Shah said U.S. export restrictions on equipment for CXMT are a real constraint, but added: “Ironically, this may instead push CXMT to move ahead of traditional giants on differentiated technology paths such as vertical channel transistors and wafer-to-wafer bonding, while the incumbents often delay innovation in order to protect returns on existing equipment investments.”