Apple's Chinese Memory Push Is a Precedent Problem for $MU and $SNDK, Not a Volume Problem
Apple’s reported move toward Chinese memory is turning into a much bigger story for the sector than the first headlines suggested, and I think the market is worrying about the wrong part of it.
The reports say Apple is testing CXMT for DRAM and YMTC for NAND, aimed at iPhones and other devices sold specifically inside China, with Washington possibly deciding whether to allow it after a Trump-Xi meeting in September. None of this is confirmed by Apple or by the U.S. government. The trail runs back to a Weibo account, and the sourcing deserves to be held at arm’s length. Monday’s reaction did not hold it at arm’s length. SanDisk fell as much as 9%, Micron and Western Digital around 7%.
The reaction is understandable. The size of it is not. Nobody is losing Apple as a global customer. The arrangement as described is ring-fenced to Chinese-market devices, which is a slice of Apple’s memory buy, not the whole of it. CXMT is also capacity constrained, running flat out and serving domestic customers first, with meaningful expansion not landing until 2028. Qualification adds another gate on top of that: one analyst’s checks found CXMT approved on a single low-volume Mac and not on iPhones at all, and YMTC not qualified on anything Apple ships. Replacing an established supplier in phone memory is not a decision, it is a program, and programs take quarters.
The politics cut the same way. Earlier this month Commerce Secretary Howard Lutnick made clear the administration was not in favour of Apple buying from CXMT and YMTC, and a bipartisan group of senators separately pressed Apple to commit that it would not use either supplier anywhere in the world. Both companies sit on the Pentagon’s Section 1260H list. So what the weekend report actually describes is a reversal, or a negotiated exception carved out as a summit gesture. It is not a policy that exists. Markets repriced a rumour of a decision as though it were the decision.
Where I think the real risk sits is one layer up, and it is not about Apple’s volumes at all. It is about precedent.
If Washington signs off on Apple using Chinese memory in China, the principle it establishes is country-specific sourcing. Any multinational with a large China business gets a template and a reason to ask for the same treatment, and some of them have already started without asking, since HP, Asus and Acer have been putting small amounts of CXMT DRAM into laptops for non-U.S. markets during the shortage. Once the carve-out has a marquee name attached to it, it stops being an exception and starts being a category. That matters because pricing power in memory is not built on any single account. It is built on the fact that there are only a handful of places to buy, and that buyers have nowhere else to go. A legitimised Chinese alternative, even a slow one, even a regional one, gives the largest buyers a second door to point at during negotiations. It would arrive precisely when AI demand has made the market as tight as it has been in years, which is the worst possible moment to hand the buy side an argument.
And then there is the detail that makes the whole thing land differently. YMTC’s parent, CCSH, has just had its application accepted for a Shanghai STAR Market listing to raise around 33 billion yuan, roughly $4.9 billion, at an implied valuation near $41 to $49 billion, targeting 2027. That would be the third-largest listing the STAR Market has done, behind SMIC and CXMT, which floated last month.
Read the prospectus numbers and the shortage stops being an American talking point. CCSH reported 47 billion yuan of revenue in the first quarter of 2026, close to five times the year-earlier level, and 33.38 billion yuan of net profit in that single quarter, more than double what it earned in all of 2025. Average NAND selling prices in the quarter ran 173% above the 2025 average. Gross margin went to 76.8% from 35.3%. TrendForce data in the filing puts YMTC third globally in NAND revenue and shipments.
Two things follow from that. First, it explains why CXMT has reportedly refused Apple any price concession, and why I would not expect YMTC to behave differently. A supplier earning a 76.8% gross margin in a sold-out market does not discount to win a politically radioactive account. The idea that a China option gives Apple leverage over Micron requires the Chinese option to be cheap or available. Right now it is neither. Second, and more importantly, look at where the IPO money goes. Around 20.8 billion yuan is earmarked for mass-production line upgrades and 12.2 billion for R&D. That is the supply response, and capital markets are funding it. The bear case for memory pricing has never been Apple’s China-market bill of materials. It has always been new capacity arriving before demand catches up, and this is what the first stage of that looks like.
Which is why an Apple approval would matter more to YMTC than it would cost Micron. Validation from the most demanding hardware qualifier on earth, landing while the parent is marketing a $4.9 billion raise, is worth more than the order. It reprices the company, it reprices the raise, and it accelerates exactly the expansion that eventually pressures everyone’s margins.
So I read Monday as a geopolitical risk premium plus profit-taking in a very crowded trade, not as evidence that Apple is about to dent anyone’s earnings. The near-term threat is thin enough that at least one analyst called the reaction an overreaction, and I agree with him on the near term.
The precedent is the part I would price. Not the order.