CXMT Surges 523% in IPO Week as Analyst Price Targets Show Widest Divergence in Asia
Alina Collins
CXMT soared 523% in its debut week, yet analysts' highest and lowest targets diverge by nearly 7× — a split that lays bare the market's fundamental disagreement over whether China can redraw the global memory-chip map.
A 7× target gap — what are they actually arguing about?
The most bullish target implies a market cap of roughly $1.1 trillion; the most bearish, about $160 billion. The gap is among the widest ever seen for an Asian stock.
This means → bulls and bears are not debating "how much upside." They are debating whether CXMT becomes the next Samsung-tier player or a latecomer crushed by the cycle.
Three brokerages rate the stock buy, one hold, one sell. The bear camp is small, but its logic strikes at the core issue.
What is the bull case built on?
China consumed one-quarter of global DRAM — the most common memory chip in phones and PCs — last year, yet domestic producers supplied less than one-third of that. In plain terms = most of the memory China uses still comes from someone else.
Nomura analyst Donnie Teng set the street-high target at ¥116, more than double last Friday's close. His core bet: that import-substitution gap.
Some bulls go further, wagering CXMT can produce China's first HBM — high-bandwidth memory, the premium DRAM that powers AI servers, currently dominated by Samsung and SK Hynix.
No EUV machines — how do you make high-end chips?
Export controls block access to EUV lithography — machines that etch the finest circuits with extreme-ultraviolet light. But multi-patterning DUV — repeating exposures with older machines to approach EUV's single-pass precision — offers an alternative route.
Aberdeen portfolio manager Bush Chu argues that stacking DUV exposures could let CXMT catch up with Korean rivals in HBM.
Last week a Chinese state-backed firm reportedly began mass-producing immersion-DUV equipment; the news sent global chip stocks into turbulence. This reflects how sensitive markets are to every step China takes in lithography.
What is the bear case?
Morningstar analyst Jing Jie Yu is the sole sell rating, with a target of just ¥16.1. She does not deny CXMT will gain share — but projects that continued capacity expansion across the industry will drag DRAM prices down in 2027–2028, hurting every producer.
In plain terms = one more big player at the table, same-size pie — everyone's slice shrinks.
Bold Wealth Partners CIO Jason Lemire frames the concern more broadly: "The supply discipline built over decades is eroding." This means → the memory industry's profits have long rested on a tacit output-control pact among Samsung, SK Hynix, and Micron. CXMT's rise could break that equilibrium.
CXMT — trillion-dollar import-substitution play, or a capacity-glut time bomb?
BULL
Massive substitution gap
China takes a quarter of global DRAM yet makes under a third domestically.
HBM pathway exists
Multi-patterning DUV is proven; the tech gap is narrowing.
IPO cash fuels expansion
Fresh capital significantly boosts capacity-build firepower.
BEAR
Expansion crushes prices
DRAM pricing downturn projected for 2027–2028, industry-wide margin hit.
Supply discipline breaks
The big-three's tacit output pact faces a structural shock.
Tech gap persists
Chips still lag two to three generations; catch-up pace uncertain.
In plain terms = the bulls bet 'China will use its own chips no matter what'; the bears bet 'one more mega-seller and prices won't hold' — they are not even arguing about the same thing, which is why targets diverge 7×.
What happens to Samsung, SK Hynix, and Micron?
All three suffered a historic sell-off last month as the AI rally faded; CXMT's rise will further test their pricing power.
Bahrain Investments' Asia-Pacific strategy head Robert Li is cautious: "Chips still lag two to three generations, even though the gap is narrowing. Chinese memory makers' global share remains quite limited."
This means → the big three's moat holds for now, but every CXMT breakthrough will trigger a market re-pricing. Over the coming years, memory-sector valuations will revolve around one central question: how far can CXMT actually go?
Content is for reference only, not financial advice.