Apple tests CXMT memory chips for China devices as DRAM maker gains clout: FT

Apple is testing DRAM memory chips from China’s state-backed ChangXin Memory Technologies (CXMT) for devices sold in China, even as U.S. technology restrictions remain in place. CXMT has become the world’s ~4th-largest DRAM producer, with ~11% share of global DRAM wafer capacity last year expected to rise to 15% by 2028, potentially weighing on memory prices. CXMT is also preparing for a domestic IPO estimated at up to 3 trillion yuan and plans to use proceeds to expand capacity and develop next-gen DRAM/HBM for AI accelerators.

Analysis

This is more about supply-chain optionality than immediate revenue. For AAPL, even a limited China-only validation process reduces single-vendor/geopolitical risk and gives procurement leverage against Korean suppliers, but it also raises the chance of SKU fragmentation, higher qualification costs, and potential U.S. policy scrutiny if it looks like a de facto technology transfer.

The more important second-order effect is competitive discipline in memory. MU and SSNLF are exposed less because this is an instant volume loss, and more because a credible domestic alternative compresses pricing power in the lowest-end DRAM buckets first; that is where Chinese substitution has historically moved fastest. If CXMT continues scaling, the first visible pain should show up in China handset and consumer-electronics channels before it reaches server or AI memory, which means the market may be underestimating the lag between capacity announcements and actual share loss.

The catalyst path is policy-driven over 1-3 months and pricing-driven over 6-18 months. A waiver or broader approval would be a negative read-through for non-China memory names because it signals Washington may tolerate more Chinese localization than expected; conversely, a rejection would make this a one-off sourcing test and likely cap the trade. The real falsifier for the bearish memory thesis is not the test itself but evidence that CXMT cannot move beyond low-end DRAM while spot DRAM prices and Samsung/Micron commentary remain stable; HBM remains a separate bottleneck and should not be lumped into this story.

Consensus may be overreacting on MU while underreacting on Apple’s bargaining power. The market should treat this as a gradual China-content substitution story, not a near-term earnings shock, and the immediate winner is likely AAPL’s supply-chain flexibility rather than CXMT’s equity value.

Key Decisions for Investors

• Relative-value: long AAPL / short MU into the next 1-3 months as a hedge against China-sourcing optionality; thesis breaks if Apple publicly walks back the testing, or if MU raises guidance and DRAM spot pricing re-accelerates.
• Avoid chasing MU on the headline; wait for evidence of China-channel inventory digestion or softer contract pricing before adding shorts. Best risk/reward is a tactical short only if DRAM ASP commentary weakens into the next earnings cycle.
• For a broader semiconductor hedge, pair long AAPL with a small short in a memory-heavy basket proxy (e.g., SOXX/SMH if available internally) rather than a pure single-name short, because the first-order impact is likely sentiment and multiple compression, not an earnings miss.
• Watch for U.S. regulatory language on CXMT access over the next 30-90 days; a permissive outcome would be a negative catalyst for MU and SSNLF, while a restrictive outcome likely makes this noise and favors covering any memory short exposure.
• No high-conviction standalone long on CXMT-related exposure yet; only consider once there is verifiable commercial adoption or clear pricing evidence that Chinese DRAM is taking share in consumer devices.