CXMT Corp surged more than 500% on its Shanghai trading debut Monday, catapulting the chipmaker to the top of China's stock market by valuation. The stock reached an intraday high of 55.03 yuan against its IPO price of 8.66 yuan, lifting its market cap to approximately $488 to $539 billion from $85.5 billion at IPO pricing. CXMT became the first A-share stock to exceed 100 billion yuan in single-day turnover. In the same hours, US memory stocks lost $541 billion in combined market value: Nvidia fell 4.37%, SK Hynix ADRs fell roughly 8 to 9%, Micron fell roughly 4.7 to 5%, SanDisk fell 10 to 12%, Western Digital fell 5 to 7%, and Seagate fell 5.6%. The S&P 500 lost roughly $1 trillion in the same session.
The market's read is that a newly capitalized Chinese memory champion with a $488 billion valuation and $8.6 billion in fresh IPO proceeds represents an existential threat to the pricing power that has driven memory stocks to extraordinary gains in 2026. That read is worth examining carefully, because the threat is real in some dimensions and materially overstated in others.
What CXMT Actually Is and What It Is Not
CXMT is the world's fourth-largest DRAM maker at 8% global market share, trailing Samsung at 36%, SK Hynix at 29%, and Micron at 24%. It is China's leading domestic memory producer and a genuine national champion in conventional DRAM for smartphones and PCs. What it is not, at least not yet, is a credible competitor in the high-margin products driving current memory valuations.
CXMT remains constrained by US export controls and is stuck on deep ultraviolet lithography, meaning it cannot access the extreme ultraviolet technology from ASML that SK Hynix, Samsung, and Micron use to manufacture advanced nodes. HBM, the high-bandwidth memory stacked in layers that goes into Nvidia's AI accelerators and commands gross margins above 80%, requires both EUV lithography and advanced through-silicon via packaging. CXMT has neither at production scale. Its conventional DRAM competes on commodity price in smartphones and PCs. It does not yet compete for the hyperscaler AI contracts worth $100 billion that Micron has locked in at take-or-pay pricing floors.
Analysts note that CXMT remains unlikely to ease the near-term memory shortage, because the shortage is concentrated in HBM, not in the commodity DRAM segments where CXMT is active. The $8.6 billion raised in Monday's IPO will fund capacity expansion in those conventional segments, adding supply that could eventually pressure DRAM and NAND pricing in the PC and smartphone markets. That is a legitimate medium-term concern. It is not a threat to the Q4 2026 or Q1 2027 earnings trajectory for SK Hynix or Micron, both of whom are sold out in HBM through at least 2027.
The Apple Signal That Mattered More Than the IPO
The catalyst that gave Monday's selloff its additional force was not purely the IPO itself. Apple is reportedly testing CXMT's DRAM chips, adding weight to the concern that Chinese memory could reach top-tier customers sooner than bulls had assumed.
That disclosure is structurally different from the supply argument. If Apple qualifies CXMT DRAM for iPhones, it opens a channel for Chinese memory to penetrate the premium device market at scale, which would represent a meaningful revenue loss for Samsung, SK Hynix, and Micron in their conventional DRAM businesses. Apple is one of the largest purchasers of mobile DRAM globally. A qualification at Apple would legitimize CXMT's product quality at the highest consumer standard available, which in turn makes it easier to win other design-ins.
The Apple test story is early stage and may not result in a design win. Apple routinely evaluates alternative suppliers without committing to them. But its timing, landing on CXMT's IPO day, made the combination of supply growth and demand diversion concerns hit simultaneously, which explains the magnitude of the US memory selloff relative to what the CXMT IPO alone would have implied.
Why the Selloff Looks Overdone on the Fundamentals
The market is pricing as if CXMT's IPO changes the HBM supply-demand equation. It does not, at least not in any timeframe relevant to current earnings estimates.
SK Hynix's Q2 2026 results arrive Tuesday. The company is expected to report the strongest quarter in its history, with HBM revenue representing roughly 50% of total DRAM revenue and pricing confirmed at levels consistent with Micron's 84.9% gross margin disclosure in June. None of that changes because CXMT raised $8.6 billion and debuted at 49 yuan. The physical reality is that CXMT cannot manufacture a single HBM4 chip competitive with what SK Hynix is shipping to Nvidia this quarter. The capacity build that CXMT's IPO proceeds will fund takes 18 to 36 months to reach production, and the EUV bottleneck means that capacity will land in conventional segments, not HBM.
The valuation math reinforces the overreaction case. SanDisk's 505% year-to-date gain and Micron's 223% run made both stocks ripe for profit-taking, and the CXMT news provided the permission structure for sellers who had been waiting for a credible-sounding bearish narrative. That is the classic dynamic of crowded momentum trades: the selling is legitimate position management using a plausible excuse, not a fundamental reassessment of the earnings outlook.
Micron's 45 analysts maintaining a Strong Buy consensus with an average target of approximately $1,486 have not revised their models on CXMT's IPO. The contracted $100 billion take-or-pay revenue base, the sold-out HBM through 2027, and the Q4 guidance of $50 billion at 86% gross margins are still intact. The market is not wrong to think about CXMT as a long-run competitive risk. It is wrong to price that risk as if it materializes next quarter.
SK Hynix Tomorrow: The Answer That Actually Matters
Tuesday's SK Hynix Q2 report is the most important piece of information available for resolving the question Monday's selloff raised. If SK Hynix confirms HBM pricing held in Q2, contracted volumes expanded, gross margins came in at or above Micron's benchmark, and management maintains its 2027 demand visibility language, the CXMT narrative gets its clearest available factual rebuttal from the company that actually competes in the market CXMT wants to enter.
SK Hynix holds 58% of the global HBM market. It delivered a 49.71% EPS beat in Q1 on revenue of 52.58 trillion won, up 198% year over year. Q2 revenue consensus expects a further step up to approximately 82 trillion won, which would represent the strongest quarter in company history. If those numbers arrive intact, the gap between what CXMT can do today and what the market is currently pricing into the selloff becomes impossible to ignore.
The CXMT debut is a legitimate long-run signal about China's semiconductor ambitions and the timeline of its competitive advance. It is not a Q3 earnings event for Micron or SK Hynix. Tomorrow's SK Hynix results will tell the market which of those two readings is currently more relevant.
Explore more exclusive insights at nextfin.ai.
