As China begins producing domestically developed immersion deep ultraviolet (DUV) lithography equipment, some warn it could quickly narrow the technology gap with ASML of the Netherlands and even threaten the high-bandwidth memory (HBM) market led by Samsung Electronics and SK hynix. However, China has already been producing semiconductors with imported ASML DUV tools. This outcome is meaningful not because China newly secured DUV, but because it laid the groundwork to keep producing with its own equipment despite Western export controls.
Chinese equipment is still in the early stage, where productivity and reliability have not been sufficiently verified. Building equipment and mass-producing with high Production yield and utilization on a production line are different matters. Lithography tools are not operated by chipmakers alone; companies like ASML, Tokyo Electron, and Applied Materials spend years tuning process conditions and stabilizing them. It is not at a level to upend the advanced chip landscape right away, but analysts say the trend of China deploying domestic tools in real production to accumulate technology and an ecosystem should be watched carefully.
1) Does China's DUV development mean it has caught up with ASML?
According to industry sources on the 3rd, state-owned ICRD in Shanghai recently began producing domestic immersion DUV equipment. It is expected to build about five units this year and around 20 next year, supplying them to SMIC, Huahong, and CXMT. China has used ASML DUV tools it already secured and multi-patterning to produce chips down to the 7-nanometer class.
Therefore, in-house DUV production does not immediately mean it has secured a new advanced process capability. The key is that, even as U.S. and Dutch controls on equipment, parts, and maintenance tighten, China has begun to establish a foundation to expand production capacity with its own tools.
The specific throughput and overlay precision of Chinese equipment have not been disclosed. Some Chinese media claim SMEE's 28-nanometer-class tool has been supplied to a local foundry, but the delivery volume and operating level have not been officially confirmed. Kim Hyeong-jun, Seoul National University emeritus professor and Director General of the Next-Generation Intelligent Semiconductor Program, said, "Securing in-house technology is a remarkable achievement, but it is still close to a prototype testing phase, and it will take years or more to secure ASML-level throughput."
2) If you build the equipment, can you mass-produce right away?
Lithography equipment is not finished once it succeeds in drawing a circuit a single time. On real production lines, thousands to tens of thousands of wafers are processed to verify hourly throughput, overlay, Production yield, utilization, and parts lifetime. This is why stabilizing a process can take years after tool installation.
You also cannot mass-produce by tuning lithography alone. Conditions must be aligned across tools from coaters/developers that apply photoresist and develop it, to etch, deposition, cleaning, and metrology equipment. The very experience of engineers at ASML, Tokyo Electron, Applied Materials, Lam Research, and KLA, who have accumulated data with customers over decades to iron out inter-tool variation and defect causes, is a barrier to entry.
Park Jae-geun, distinguished professor at Hanyang University's Department of Electronic Engineering, said it would take about three to four years for Chinese tools to reach ASML's level. Park said, "At first, Production yield and productivity will lag, but the Chinese government is pushing the use of domestic tools, and as production experience accumulates, the tool technology will also improve."
3) Can advanced chips and HBM be made with DUV?
It is not accurate to say DUV cannot implement processes at 14 nanometers and below. Before introducing EUV, chipmakers implemented 14-nanometer, 10-nanometer, and 7-nanometer-class processes with DUV multi-patterning. The issue is production cost and Production yield. If circuits that EUV would process in one pass are divided into two to four or more passes with DUV, process time grows, and overlay error and defect risk also increase.
HBM is also not impossible to produce. However, HBM is a product that combines not only DRAM performance but also through-silicon via (TSV), stacking, bonding, thermal management, and customer qualification. If the base DRAM's Production yield and performance are low, the productivity and performance of the finished goods made by stacking multiple dies also decline.
China is likely to roll out products with fewer stacks and lower speeds first, rather than top-performance HBM4 or 12- and 16-high products. Park said, "Even if high-performance HBM at the level of Samsung Electronics or SK hynix is difficult right away, China could target the inference AI market with, for example, 8-high products."
4) Is this an immediate threat to Korean semiconductors?
It is unlikely that Chinese DUV will shake Korea's high-performance memory market in the short term. However, in the commodity and legacy markets—such as DDR4, lower-spec DDR5, and automotive and appliance chips—supply expansion and price competition could intensify. If China, blocked from adding imported DUV tools, increases supply of domestic tools, it can keep expanding commodity chip capacity even with lower productivity.
Samsung Electronics and SK hynix are expected to further accelerate a shift away from commodity memory toward high-value-added products such as HBM and high-performance DRAM. With AI data center investment, high-performance memory is also in short supply, so an increase in Chinese commodity output does not directly translate into a deterioration of overall results for domestic firms.
Director General Kim said, "More than the immediate tool performance, we should be more wary that China is continuing to use domestic tools and building an ecosystem where funding, production data, and technology improvements reinforce one another." The reason to be cautious about the localization of Chinese DUV is not today's performance, but that China has begun to establish a structure to produce and improve tools on its own and narrow the technology gap.