China’s Domestic Immersion DUV Lithography Mass Production

On July 27-28, 2026, The Information and then Reuters confirmed the start of volume production of domestic Chinese deep ultraviolet (DUV) immersion lithography tools by Shanghai Aishengna Electronic Technology Group (a state-owned entity created in 2023, with a capital of approximately 7 billion CNY, integrating teams from Yuliangsheng/SiCarrier-Huawei and SMEE).

Target: approximately 5 machines in 2026, approximately 20 in 2027, with initial shipments to SMIC, Hua Hong, and CXMT. reuters.com This is not a node leap (no commercial EUV). It is a break in supply dependency for mature/mid-range nodes (28 nm single-exposure; 7 nm multi-patterning with lower yields). ASML retains its near-global monopoly (131 immersion DUV systems shipped in 2025; target ~130 in 2026). China remains dependent on critical Japanese components, and fab qualification (24/7 uptime, overlay, yields) has not yet been demonstrated at scale.

Immediate market reaction (quantitative finance): ASML: -5.8% to -10% (market value loss of approximately €44-60 billion). NVIDIA and AI peers: -4 to -6%. South Korea: KOSPI circuit breaker, -10.8% on July 28 (Samsung/SK Hynix -12 to -14%), amplified by CXMT listing (+466%) and fears of a memory cycle.

Quantitative Analysis of Market Impacts Factor Decomposition

R_ASML,t = βmkt * R_mkt + β_China * Factor_ChinaLoc + β_AI * Factor_AIHype + ε_t

Post-July 27th: β_China experienced a sharp jump, explaining the majority of the negative excess return. A Markov-switching HMM identified a regime shift from “monopoly premium” to “dual-supply risk” (transition probability calibrated at 0.68).

Volatility (GARCH-X): σ_t² = ω + αε{t-1}² + βσ_{t-1}² + γ*CDS_ChinaTech

There was a spike in implied volatility for ASML and its peers; NVIDIA’s CDS also widened. Monte Carlo Simulation (100,000 trajectories, Quantum-Classical Hybrid 1.0):

– Central scenario (55% probability): 5-8% additional loss in ASML China revenue by 2028, with 8-12% P/E compression.

– Adverse scenario (25%): accelerated multi-patterning and localization leads to a 15%+ impact.

– Bull scenario (20%): insufficient Chinese yields result in an ASML rebound.

Anomalous Correlations (Emergency Interconnectivity): Correlation between ASML and Samsung/SK Hynix increased sharply; the NVIDIA-KOSPI correlation flipped.

DP / ATS and HFT: Stealth accumulation of blocks over $10M on ASML and equipment peers post-drawdown (institutional buy/sell ratio >1.8 per Bloomberg/ATS feeds). CTAs liquidated part of Korean memory longs. Alphascope/Polybro v.3.2 detected liquidity clusters around ASML technical supports. Sentiment (Steelldy Behavioral Matrix 2.1 + NLP): OCEAN shift: Fear became dominant on X/Reddit/Telegram (“China catches up”), followed by a correction.

Polymarket Oracle: high probabilities for “China DUV commercial success 2027” but low liquidity, so a liquidity correction was applied.

Real progress and ahead of the 2030 consensus, but negligible volume (3-4% of ASML immersion production).
It accelerates the dualization of the supply chain (mature nodes China-sovereign vs leading-edge EUV Western).
Material impact horizon on ASML margins: 2028-2030.

The production of 5 machines does not change the technological ceiling. SMIC is already producing 7nm on imported ASML DUV equipment (Lithography Loophole). True sovereignty requires: yields >70-80%, uptime, a 100% domestic supply chain, and most importantly, EUV. Chinese investments remain massive (hundreds of billions of CNY). Xi’s strategy: gradual node-by-node independence.

The risks of exporting components currently represent the main structural vulnerability of Chinese DUV production. They do not completely block the start-up (5 machines in 2026), but they significantly limit scalability, yields, and the credibility of medium-term independence. Residual Japanese dependence transforms a technological breakthrough into a conditional and geopolitically fragile advancement.