BlockBeats News, July 28th - Samsung Securities' semiconductor and IT analyst Lee Jong-wook released a report stating that multiple media outlets reported on July 27th that a Chinese state-backed enterprise's immersion DUV lithography equipment may enter the mass production stage. The plan is to produce 5 units in 2026 and 20 units in 2027 for SMIC and Yangtze Memory, but mass production capability still needs to be verified. For comparison, ASML shipped a total of 131 immersion DUV machines in 2025, mainly used for producing 7nm, 14nm, and 28nm chips.

The report stated that this development is not entirely new. In 2025, there were reports that a Shanghai-based company related to Huawei was developing 193nm immersion DUV technology. The current mass production project is also believed to be led by the same development team. The significance lies in this being China's first achievement of domestication of an immersion platform; disregarding performance differences, it is equivalent to mastering the technology level of ASML around 2008, posing a negative impact on global equipment companies like ASML.

The impact of Chinese semiconductors on this AI chip cycle is limited because the core of this cycle is still the U.S. hyperscale cloud providers and NVIDIA's data center ecosystem. AI chips and server DRAM are not homogeneous goods, and quality and on-site failure rates will directly affect the total cost of ownership. Product validation also takes a long time, plus geopolitical risks make it difficult for Chinese AI chips and DRAM to enter the U.S. data center ecosystem in the short term.

The news about Chinese DUV technology spreading from equipment stocks to U.S. semiconductor stocks declining is a typical case of the "tail wagging the dog," more reflecting the current semiconductor sector's fragile sentiment. The impact of the Chinese factor is on the depth of the next cycle's trough, rather than when this cycle will peak, and it has already been largely reflected in the roughly 5x P/E ratio of the three major memory manufacturers. What truly needs to be observed now is whether there will be evidence proving that investments in this AI data center cycle have peaked.