ACM Research looks underpriced because its Chinese subsidiary alone has a backlog equal to about half the company's market cap. ACM Shanghai's backlog rose 88.2% to $2.5B, yet the stock barely reacted. x.com  x.com 

ACM Research is the best idea under $5B in market cap right now because China is going to keep pushing for localization. Ongoing dilution of its stake in its Shanghai subsidiary is a valid concern, but a necessary one. x.com  x.com 

ACM Research carries binary China risk: China is about 97% of its business, and authorities could stop it any day, forcing management to choose between San Jose and China. The CCP's threats toward other countries make that exposure uglier, whatever the company's $4B revenue ambition. x.com  x.com  x.com 

ACM Research should benefit when its core customers CXMT and YMTC go public, because US investors will not be able to buy them and are likely to use ACM Research as a proxy. As the market starts pricing in who the real memory winners are, Micron looks set to lose out. x.com 

ACM Research should benefit as China's wafer fab equipment spending soars over the next couple of years across memory and leading-edge logic, more than offsetting lower mature-node spending. Its advanced packaging sales to non-Chinese clients are also growing well, and valuation looks reasonable after the recent pullback. x.com 

Chinese semiconductor localization names like ACM Research and NAURA look like attractive opportunities for the second half of 2026 because CXMT's IPO should channel capital their way. Chinese sell-side analysts say CXMT's market cap right after listing could reach at least RMB 5 trillion, a frenzy Western investors underestimate, and most of the money it raises will likely flow into localization stocks. x.com 

ACM Research is the cleanest second-derivative play on AI capex, and Kimi's rise makes it more bullish. A China that is frontier-competitive in models but cannot buy Nvidia's best chips must build its own compute, HBM and packaging, and that spend lands on ACM Shanghai, its STAR-listed operating company, where new orders rose 65% year over year in Q1. Its new Lingang flagship site has about $3B of annual output capacity as it ramps. x.com  x.com  x.com 

Micron's China sales risk shifting to CXMT as its domestic DRAM capacity grows. Mainland sales more than tripled to $8.3B during the shortage, mostly standard DRAM that CXMT already makes; including Hong Kong, Micron booked $14.6B, 11% of FY2026 revenue. CXMT already has roughly a third of Micron's revenue, and Reuters said its server DDR5 was priced above Samsung's. Micron is the only American maker among the three ahead of CXMT; buyers drawn by supply could switch as it expands. x.com  x.com 

China's domestic market lets YMTC and CXMT threaten Micron, Samsung Electronics and SK hynix without entering US products. China receives 15-25% of smartphones and PCs and is on a path to at least 20-30% memory share over the next several years; current leaders are much too optimistic. History suggests at least a near-term semiconductor rally even if the AI trade has topped. It likely has not: agentic AI should give the trade room to run into 2027. x.com  x.com  x.com 

Memory remains the main AI bottleneck, and the supply-demand gap will keep widening through 2027. New capacity takes years of fabs, advanced stacking, packaging and testing, so supply only starts improving slowly around 2028. That is why Elon Musk is spending hundreds of billions on TeraFab, with memory capacity a main goal. China's CXMT says in its IPO prospectus that its capacity ranks first in China and fourth globally yet falls far short of even domestic demand. x.com