#GoldmanSachsBullishOnCXMT

I’m bullish on CXMT’s business trajectory, but much more cautious on the stock at current expectations.

Goldman’s thesis is compelling: CXMT’s monthly wafer capacity is projected to rise from 270k in 2026 → 447k in 2028 → 665k in 2030, more than doubling. Goldman also sees CXMT potentially supplying around 50% of China’s DRAM demand by 2028.

Why I’m bullish

China’s AI buildout: Increasing AI-server demand should support DRAM/HBM consumption.

Domestic substitution: CXMT is strategically positioned as China’s leading large-scale DRAM producer.

Capacity expansion: More than 2× wafer capacity over four years is a major earnings-growth lever.

HBM optionality: Goldman expects HBM revenue to rise dramatically, although from a very small base.

Memory supercycle: Samsung, SK hynix and Micron are allocating more resources toward HBM, potentially leaving conventional

DRAM tighter.

But here’s the big catch

The Yuan129 target is an aggressive bull case, not simply a conservative valuation.

Goldman effectively needs several things to go right simultaneously: sustained high DRAM prices, successful capacity ramp-up, better yields, product upgrades, meaningful HBM adoption and a huge improvement in profitability. Its model assumes CXMT's gross margin rises from 41% in 2025 to 82% by 2030.

And there is a striking disagreement among analysts. The current market data shows CXMT around ¥56.65, while the analyst consensus is Buy with an average target around Yuan 75.11;Yuan 129 is the high end. Meanwhile, Morningstar's fair-value estimate is only Yuan16.10, arguing that DRAM is cyclical and CXMT lacks a durable moat.

So I'd frame it like this:

CXMT the company: Bullish

CXMT's 2030 earnings potential: Very bullish

CXMT at ~Yuan 57: Interesting, but expectations are already high

Yuan 129 target: Possible in a strong memory bull market, but definitely not guaranteed

The most important variable for me isn't actually the capacity doubling. It's HBM. If CXMT can move from being primarily a conventional DRAM challenger into a credible HBM supplier, the valuation framework changes substantially. Goldman itself acknowledges that CXMT's HBM technology is still at an early stage.

My take: I would be bullish on CXMT over 3–5 years, but I wouldn't blindly equate Goldman’s Yuan 129 target with fair value. The better question is whether CXMT can convert enormous capex + domestic demand + HBM development into sustainable margins and free cash flow.

If that happens, Yuan 129 could eventually look conservative. If DRAM prices roll over or HBM fails to scale, the stock could look extremely expensive even around today's price.