Executive Summary

China is making progress toward reducing its dependence on imported semiconductor lithography equipment. Building an EUV scanner that can compete in commercial production, however, remains a formidable engineering challenge. Recent reports describe advances in EUV source development and an early system prototype, but there is still no public, independently verified evidence that China has a fully domestic EUV scanner producing advanced chips at commercial scale.

The nearer-term opportunity is deep ultraviolet (DUV) lithography. Chinese companies already supply equipment for some mature-node processes, and reports of new immersion DUV machines suggest that domestic alternatives may be moving closer to deployment. Whether they become useful production tools will come down to yield, throughput, reliability and customer qualification—not simply whether a machine has been assembled.

EUV is a much longer-term challenge. China would need to bring together a stable, high-power light source, ultra-precise optics, fast and accurate wafer stages, masks, photoresists, metrology and control software in a system that can run reliably in a high-volume fab. Progress in individual components matters, but it is not proof that the complete scanner is ready for production.

Limited experimental use of domestic EUV may be conceivable near the end of this decade. Dependable serial production of competitive scanners could take well into the 2030s, or longer. These are analytical estimates, not announced or confirmed industry schedules.

The stakes extend beyond China’s semiconductor industry. Domestic DUV tools could help Chinese fabs sustain selected processes with less reliance on imported equipment. A viable EUV scanner would have much larger strategic consequences, changing the competitive landscape for lithography and potentially weakening the effect of export controls. For now, the most useful signals are independently verifiable wafer-patterning results, sustained source power, evidence of system integration and proof that domestic DUV tools can meet production requirements.

China’s lithography effort is beginning to produce more tangible milestones after years of research. Still, generating EUV light, building a working prototype and manufacturing advanced chips reliably at industrial scale are three very different achievements.

For decades, the most advanced semiconductor lithography market has depended on an exceptionally concentrated supply chain. ASML, based in Veldhoven in the Netherlands, is the only company that has commercialized extreme ultraviolet (EUV) lithography systems for high-volume semiconductor manufacturing. Its position rests not only on the scanner itself, but also on a network of specialized suppliers, most notably Carl Zeiss for precision optics.

China has spent years trying to reduce that dependence. Export restrictions have made the effort more urgent, while government-backed investment, domestic equipment companies and research institutes have expanded the country's semiconductor capabilities.

Two recent developments make this a useful moment to take stock of China’s progress.

First, a Reuters investigation published in December 2025 reported that a Chinese team had built an operational EUV prototype in Shenzhen. The machine was reportedly generating EUV light, although it had not yet produced working chips. Second, in October 2026, Shanghai-based startup Langdao Silex publicly demonstrated a prototype EUV light-source system, claiming more than 15 W of EUV output and approximately 90% domestic content. The company reportedly estimates that mass production of the light source remains three to five years away.

Meanwhile, China's domestic deep ultraviolet (DUV) effort has moved closer to industrial deployment. Reuters reported in July 2026 that Shanghai Aishengna Electronic Technology Group had begun producing domestically developed immersion DUV machines, with initial production expected to be small and further testing still required.

These developments are worth watching, but they do not show that China has matched ASML’s production capability. The questions are practical: what has actually been demonstrated, which organizations are doing the work, how far the systems are from dependable serial production, and what a competitive Chinese lithography industry would mean for equipment suppliers and export controls.

1. Why lithography is the critical bottleneck

Lithography transfers circuit patterns onto a silicon wafer using light, a patterned mask and a highly controlled optical system. The process is repeated many times to create the layers that make up a semiconductor device.

The two principal technologies relevant to China's current efforts are DUV and EUV.

Deep ultraviolet (DUV): 193 nm for advanced ArF systems

DUV remains essential for logic, memory, analog, power and mature-node chips, including applications with feature sizes in the tens to hundreds of nanometres. Immersion lithography places a thin layer of highly purified water between the projection lens and wafer, improving resolution. More demanding patterns can also be printed through multiple patterning, but each additional step adds cost and process complexity.

Extreme ultraviolet (EUV) machine: 13.5 nm wavelength

EUV uses a much shorter wavelength and reflective optics to print particularly demanding circuit layers. It is central to today's leading-edge logic manufacturing, including advanced 7 nm, 5 nm, 3 nm and newer process generations. Its advantage is not simply smaller features: it can reduce the number of exposures and process steps needed for selected critical layers.

Generating light at the right wavelength is only one part of the job. A production scanner must bring together the source, precision optics, mask handling, wafer stages, alignment, vibration isolation, contamination control, metrology and software—and keep printing patterns accurately, run after run. That is the point at which a prototype and a commercial tool begin to diverge.

A laboratory can show that a component works. A fab needs a machine that delivers consistent results across thousands of wafers, stays available, controls defects and fits into the wider manufacturing process. It must also be maintainable and economical to run. ASML’s own timeline shows how long this can take: its first EUV demonstration tools shipped in 2006, the pre-production NXE:3100 followed in 2010, and the first production NXE:3300 shipped in 2013. Turning the underlying physics into a commercially useful platform took years.

China does not have to repeat that development path exactly. It can draw on decades of existing EUV research and on experience gained from other lithography technologies. But it still has to solve the practical problems that determine whether a machine can be used in production.

China faces a particularly large gap. Beijing has invested heavily in domestic semiconductor equipment, and local suppliers have made progress in etching, deposition and wafer cleaning. Lithography remains a weaker link. The Center for Security and Emerging Technology (CSET) estimates that Chinese manufacturers accounted for about 0.2% of the global lithography-equipment market in 2025. That figure measures Chinese suppliers’ share of the global market—not the share of Chinese fabs’ purchases supplied domestically. The distinction matters: advanced chipmaking depends on accurate, repeatable and economical patterning as well as capable etch and deposition tools.

DUV is the nearer-term opportunity. Chinese suppliers already make lithography systems for less demanding process layers, and reports of domestic immersion DUV machines point to an effort to move into more advanced applications. But building a scanner is only the first step. Before fabs can depend on it for volume production, it must demonstrate overlay accuracy, throughput, uptime, defect control and repeatable wafer yield.

An EUV scanner is far more than a light source or a precision optical instrument. It is a tightly integrated production system. The source must deliver stable EUV power; multilayer mirrors must collect and focus the light; mask and wafer stages must stay aligned at high speed; and the whole machine must control contamination, vibration, thermal drift and defects. Those subsystems have to work together over long production runs, not just perform well in separate demonstrations.

Why China cannot easily close the gap

1.1. The EUV source is a demanding industrial system, not simply a laser at a shorter wavelength.

EUV lithography generally generates 13.5 nm radiation by directing high-power laser pulses onto microscopic tin droplets, producing plasma that emits EUV light. The source must deliver enough usable power to the wafer-exposure system while maintaining stability, controlling debris and avoiding damage to surrounding components. A laboratory demonstration that produces EUV radiation is important progress, but it does not prove that the source can sustain the power, availability and lifetime needed for economical manufacturing.

1.2. The optical system is a major independent bottleneck.

At 13.5 nm, EUV light is absorbed by ordinary materials, including air and conventional transmissive lenses. The scanner therefore operates in a vacuum and uses highly specialized multilayer mirrors rather than the transmissive projection optics found in conventional DUV systems. These mirrors must achieve exceptionally precise surface quality and multilayer coatings, while maintaining performance under demanding operating conditions.

China can develop optical research programmes and manufacture increasingly sophisticated components, but producing a complete set of EUV optics that meets scanner-level specifications is a different challenge. The problem is not merely making a good mirror; it is manufacturing, measuring, aligning and maintaining an entire optical train whose errors accumulate across the system. ASML identifies Carl Zeiss SMT as its sole supplier of critical lithography optics, illustrating how concentrated this part of the global supply chain remains.

1.3. Nanometre-scale positioning must work at production speed.

A scanner must move the wafer and mask through precisely coordinated trajectories while maintaining alignment across successive exposures. Vibration, thermal expansion, stage-positioning errors and optical distortions can shift patterns relative to previously printed layers. Even if a prototype produces a sharp pattern on a test wafer, a commercial scanner must repeat that performance across large numbers of wafers, at high throughput and with very low defect rates.

This is especially important for advanced chips, where small overlay errors can compromise electrical connections or reduce yield. China must therefore solve both the precision problem and the productivity problem at the same time.

1.4. A working scanner depends on an entire domestic supply chain.

The scanner requires more than its most visible components. Masks, photoresists, contamination-control systems, metrology, precision mechanics, control electronics and specialized software all affect the final pattern. Some components can be sourced or developed independently, but the complete system must be calibrated and optimized as one machine.

Export restrictions make this task harder by limiting Chinese access to EUV scanners, certain advanced DUV systems, related components and technical support. ASML has stated that it has never shipped an EUV lithography system to China. As a result, Chinese developers cannot simply acquire a commercial EUV scanner, study its operation in a domestic fab and gradually localize its components. They must develop much of the system through research, component testing, prototypes and iterative integration.

1.5. The final test is manufacturing economics, not a successful demonstration.

A prototype can show that the underlying physics works. A semiconductor fab needs a machine that delivers consistent critical dimensions, low defect rates, high availability and competitive throughput over extended periods. It must also integrate into a tightly controlled production line and be maintainable without excessive downtime.

These requirements create a difficult feedback loop. Better performance depends on repeated testing and refinement, but meaningful refinement becomes easier when a developer has access to mature tools, experienced suppliers and production-fab data. China is investing to build that capability domestically, but the lack of an established commercial EUV ecosystem makes the process harder.

Why the distinction between a prototype and production matters

ASML’s own development history illustrates the length of the journey. Its first EUV demonstration tools were shipped in 2006, the pre-production NXE:3100 followed in 2010, and the first production NXE:3300 was shipped in 2013. The progression took years of engineering, supplier development, integration and customer qualification.

China does not have to repeat ASML’s development path exactly. It can draw on decades of EUV research, existing semiconductor manufacturing experience and domestic investment. Nevertheless, the underlying requirements remain: a viable source, precision optics, accurate stages, process control, reliable subsystems and a complete supply chain must all come together in a commercially productive machine.

This is why lithography remains a critical bottleneck for China. The issue is not whether Chinese researchers can demonstrate EUV radiation or print a pattern under laboratory conditions. It is whether Chinese companies can turn those achievements into a reliable, repeatable and economically competitive production platform—and do so despite restrictions on access to key foreign technologies and equipment.

2. EUV: what has China actually achieved?

The most important distinction is between three milestones:

China has now reported progress at the first milestone and, according to Reuters, has an integrated prototype that is generating EUV light. Public evidence of repeatable chip production from a complete Chinese EUV scanner remains absent.

The Shenzhen prototype: In December 2025, Reuters reported that a high-security facility in Shenzhen housed a large EUV prototype completed in early 2025. According to people familiar with the project, the system was built by a team that included former ASML engineers and was operational enough to generate EUV light, but it had not produced working chips. The project was described as part of a wider state-backed effort involving Huawei and research organizations.

The report also described a machine considerably larger and less refined than ASML's commercial systems. One reported reason was the difficulty of replicating the high-precision optical assemblies available through established European suppliers.

This is meaningful progress, but it needs to be interpreted carefully. A prototype that generates EUV radiation is not equivalent to a scanner that can print advanced chip patterns at production speed. The next stages require accurate illumination, projection optics, mask handling, wafer positioning and control of the entire exposure process.

There is also a limit to what can be independently verified. The Shenzhen project has been reported primarily through anonymous sources rather than through published system specifications, customer qualification data or independently demonstrated wafer results.

Langdao Silex — FlintNano LPP-OXXO: A publicly displayed EUV light-source prototype

At the China International Industry Fair in Shanghai, Langdao Silex (朗道燧岩) displayed a prototype EUV light-source system. Reports attributed to the company describe an output above 15 W, a 1,000 W driving laser and continuous operation exceeding 100 hours. The company also claimed that approximately 90% of the components were domestically sourced. This is an important development because the light source is one of the most difficult subsystems in an EUV scanner. ASML's commercial approach uses a high-power laser to strike microscopic molten tin droplets, producing plasma that emits radiation around 13.5 nm. The process must run at high repetition rates while managing debris that can contaminate the optical collector.

But the headline number needs context. The reported output of more than 15 W is far below the hundreds of watts of usable EUV power associated with commercial production systems. It is also not clear from the public demonstration whether the Chinese figure is measured at the same optical reference point used for commercial scanner specifications. The reported 100-hour operating period is an early endurance milestone, not evidence of production-level availability over extended fab operation.

The distinction matters because light generated at the source must pass through the optical collection and illumination system before it reaches the wafer. Losses along this path reduce the power available for exposure. Higher power alone is also insufficient: the source must be stable, repeatable and compatible with contamination-control requirements.

Langdao Silex reportedly estimates that mass production of its light source is still three to five years away. Even if that schedule is achieved, it would refer to the light-source product, not necessarily to a complete EUV scanner.

Langdao Silex FlintNano LPP-OXXO: reported technical specifications

Chinese research is addressing more than the source: Official Chinese research programs show that the effort extends beyond a single machine or company.

In June 2025, the Shanghai Science and Technology Commission published a research call for a high-power EUV source based on an electron-beam-driven plasma approach. The program targeted more than 100 W of EUV radiation at the intermediate focus and was scheduled to run from October 2025 to September 2027. This is a research target, not evidence that the specified output has already been achieved.

Separately, a Chinese Academy of Sciences institute has reported work on EUV optical technology, including an EUV principle exposure system and a laser-produced-plasma source prototype. Its published results include static exposure patterns and source-conversion measurements, demonstrating domestic research activity at the component and experimental-tool level. They do not establish the existence of a production-ready scanner.

China is also working on materials needed for EUV processing. In August 2025, the National Natural Science Foundation of China highlighted a Tsinghua University team's research on a tellurium-based EUV photoresist. The researchers reported resolving 18 nm line widths at a dose of 13.1 mJ/cm², with a line-edge roughness of 1.97 nm. These are material and patterning research results, not a demonstration of high-volume manufacturing with a Chinese EUV scanner.

Taken together, these efforts suggest a broad strategy: develop the light source, optics, resist chemistry and related process technologies in parallel. That is the right direction for building a domestic ecosystem. The challenge is to make those separate advances work together in a single, reliable manufacturing platform.

3. Who is leading China's EUV development?

China’s wider lithography strategy: Building an ecosystem, not just a machine.

ASML has a defined commercial product line and an established supplier network. China’s EUV effort, by contrast, appears to involve a mix of state-backed programmes, research institutes and private technology companies. That distributed structure is not necessarily a weakness: different teams can work on sources, optics, materials and other difficult components in parallel. The harder question is who is responsible for integrating them into one machine and proving that the complete system works in a fab.

This approach reflects the structure of the technology itself. A commercial EUV scanner is not a standalone invention that can be completed by solving one difficult problem. It is a tightly coupled system whose performance depends on the interaction of dozens of specialized subsystems. Improving the light source, for example, has limited value if the projection optics cannot collect and focus the radiation efficiently. High-quality optics are insufficient if the wafer stage cannot maintain precise positioning at production speed. And even a scanner that prints a convincing test pattern is not commercially viable unless it can do so repeatedly, with acceptable defect rates, throughput and uptime.

China's response has been to distribute parts of the problem across research institutes, equipment developers, specialist suppliers and semiconductor companies, while using state-backed coordination to align their work.

3.1. A division of labour across the technology stack: Several organizations and programmes illustrate how this strategy is taking shape. Their roles should not be confused: some are developing research components, others are building lithography equipment, and still others are working on materials or manufacturing processes. Public evidence does not establish that every organization listed below supplies the Shenzhen EUV prototype.

3.2. Why coordinating these efforts could accelerate development: The advantage of a distributed programme is that several technical bottlenecks can be addressed in parallel. A light-source team can work on plasma generation, conversion efficiency and output stability while an optics group develops multilayer mirrors and optical metrology. Another team can investigate wafer-stage positioning, control software or contamination management. Materials researchers can develop photoresists and exposure processes suited to the radiation produced by the source. These efforts need not wait for a complete scanner to exist before beginning.

The work can also generate useful feedback between disciplines. Source engineers need to know the optical system's collection requirements. Optical engineers need information about source stability and spectral characteristics. Resist developers need the actual exposure conditions and image characteristics. Stage and control engineers need the scanner's imaging and overlay targets. Coordinating these requirements early can reduce the risk of developing components that work individually but cannot function together.

Huawei's reported coordinating role is relevant in this context. A central organization can define interfaces, allocate resources, establish technical milestones and ensure that component teams work toward a common system architecture. This is different from asking every supplier to develop its own independent solution and attempting to assemble the results at the end.

It also helps explain why China's strategy extends beyond a single company. The goal is not necessarily to create a vertically integrated manufacturer that produces every component itself. It is to build a domestic network capable of delivering a complete machine, with an organization responsible for the overall architecture and performance.

3.3. The harder problem is system integration: However, parallel development does not automatically translate into a working production scanner. In fact, system integration can become the most difficult phase precisely because the subsystems interact.

Consider the EUV source. Increasing its output power may improve the potential throughput of the scanner, but it can also increase the thermal load on optical components and complicate debris management. A more powerful source is not necessarily useful if the collector degrades rapidly or the illumination system cannot deliver a stable field across the wafer.

The same principle applies to precision optics and wafer stages. The optical system must produce a sufficiently accurate image, while the stage and control system must position the wafer correctly relative to that image. The required performance cannot be established by evaluating either subsystem in isolation. Both must be calibrated and validated as part of the complete machine.

These interactions create a difficult engineering loop: subsystem performance determines the scanner architecture, while the scanner architecture determines the requirements imposed on each subsystem. Changes to one component can require changes elsewhere. A successful integrator must therefore manage interfaces, tolerances, calibration, thermal behaviour, vibration, software and failure modes across the entire platform.

ASML's position illustrates the importance of this role. Its advantage is not limited to access to advanced optics or an effective EUV source. It has developed a commercial platform around a network of specialized suppliers and has accumulated extensive experience integrating those technologies into machines that operate in customer fabs. Replicating individual components is not equivalent to replicating that system-level capability.

3.4. Domestic suppliers must also become production suppliers: A domestic supply chain creates another advantage: Chinese developers can work toward replacing foreign components that are difficult or impossible to procure. But localization introduces a second challenge. Each substitute must meet the performance, reliability and consistency requirements of the complete scanner.

A prototype component may be adequate for laboratory experiments but unsuitable for a machine expected to run continuously. Optical coatings must maintain their characteristics across production batches. Precision stages must repeat their motion accurately over long operating periods. Sensors and metrology systems must remain calibrated. Software must coordinate the subsystems and detect faults before they compromise wafers.

Supplier maturity matters as much as component capability. A commercial platform needs repeatable manufacturing, documented quality control, spare parts, maintenance procedures and a route to improve performance as problems are discovered in the field. These capabilities usually develop through repeated integration and customer use rather than through a single successful demonstration.

This is why domestic content percentages should be interpreted cautiously. A claim that a source or machine contains a high proportion of domestic components can indicate progress in localization, but it does not establish that those components match the performance, reliability or lifetime of established foreign alternatives.

3.5. DUV provides a nearer-term route to integration experience: China's DUV programme is strategically relevant because it offers a more immediate opportunity to develop lithography-equipment engineering, supplier relationships and customer-qualification capabilities.

DUV systems do not face all the same optical and source challenges as EUV. Nevertheless, advanced immersion DUV tools still require sophisticated projection optics, precision stages, alignment systems, control software and manufacturing-process integration. Developing these capabilities can strengthen the broader domestic equipment industry, even if the resulting systems cannot simply be converted into EUV scanners.

Reuters reported in July 2026 that China had begun producing domestically developed immersion DUV machines under the leadership of Shanghai Aishengna Electronic Technology Group. The report also noted that the machines still faced performance and reliability challenges and would require further testing. This is an important distinction: reported production is a milestone, but customer qualification and repeatable manufacturing performance determine whether the equipment becomes a viable alternative.

Such potential feedback loop is significant. Domestic fabs can help equipment developers identify failures, refine process recipes and improve reliability. Those experiences can strengthen local suppliers and the ability to manage complex equipment projects. They do not eliminate EUV's unique challenges, but they can improve the industrial foundations on which a future EUV programme depends.

3.6. The remaining uncertainty: who owns the complete system? China's distributed approach may help it develop missing technologies faster than a fragmented collection of independent projects would. But the public record leaves an important question unanswered: how the country will turn these capabilities into a fully integrated, production-qualified EUV platform.

Reuters has described Huawei as a central coordinator and identified contributions from Chinese research institutes. Public reporting has also identified separate efforts in EUV-source development and domestic DUV equipment. What remains unclear is the precise allocation of responsibility for the complete Shenzhen machine: who controls its system architecture, who qualifies each subsystem, who manages the integration programme and which organization will ultimately support the scanner in a production fab. That responsibility is critical. The organization that owns the complete system must resolve conflicts between subsystem requirements, validate the integrated machine and take responsibility when performance falls short. It must also manage the transition from a research prototype to a manufacturable product with predictable performance and service requirements.

In Summary: China's lithography programme should be judged neither as a single machine project nor as proof that a domestic EUV ecosystem is already complete. The more meaningful development is the emergence of a coordinated effort spanning research institutes, equipment developers, specialist suppliers and semiconductor manufacturers.

This structure could accelerate progress by allowing multiple bottlenecks to be addressed in parallel and by reducing dependence on foreign suppliers. Domestic DUV development may provide earlier benefits, including experience in equipment integration and customer qualification. EUV remains a more demanding objective because its optical, source and system-level requirements are substantially harder to satisfy.

The decisive evidence will be the transition from individual component demonstrations to an integrated scanner that produces repeatable wafer patterns, meets overlay and defect requirements, achieves useful throughput and operates reliably in a customer fab. Until those milestones are demonstrated, the ecosystem is best understood as a potentially powerful foundation for future progress—not proof that China has already matched the capabilities of established commercial EUV platforms.

4. How long until China can manufacture EUV systems in volume?

This is the question that matters most to the market, and it requires a more precise definition of “production.” A machine can be assembled in a factory and still be years away from commercial qualification. For this assessment, serial production means that a manufacturer can repeatedly build complete scanners, install them at customer fabs, qualify the tools for defined process layers and support them with acceptable uptime, throughput and defect performance.

The central question is what China is trying to achieve: an EUV machine that works in a laboratory, one that can manufacture advanced chips for domestic use, or a system that can eventually compete with ASML on performance and cost.

China’s goal appears to extend beyond meeting immediate demand for EUV lithography. The broader effort is to build a semiconductor-equipment supply chain that China can control, reduce reliance on foreign technology and, over time, support advanced chip manufacturing with domestic tools.

However, three milestones must be kept separate:

China can make progress on the first milestone without having achieved the second, and can achieve the second without coming close to the third. That distinction is essential to evaluating the timelines.

1. What China has actually committed to: The most useful way to assess China's intentions is to separate official national policy, state-backed industrial investment, and reported EUV project targets. They provide different levels of evidence.

China's 15th Five-Year Plan, 2026–2030:  China's official plan calls for decisive breakthroughs in key technologies across the integrated-circuit supply chain, with extraordinary measures to strengthen technological self-reliance. The scope explicitly includes semiconductor equipment, not just chip design or fabrication. The plan also links research, technology transfer, commercialization and industrial deployment.

What it means for EUV: EUV fits directly into the national strategic objective, even though the public plan does not provide a verified, machine-specific EUV production deadline.

National Integrated Circuit Industry Investment Fund — Phase III: Established in May 2024 with registered capital of 344 billion yuan (about US$47.5 billion). The fund is part of China's effort to reduce dependence on foreign semiconductor technology. Its capital is not an EUV-specific budget, so it would be incorrect to claim that the entire amount is available for lithography. But it establishes the scale of the state-backed financing environment in which domestic equipment suppliers operate.

The Shenzhen EUV prototype: Reuters reported in December 2025 that a prototype in Shenzhen had been completed in early 2025 and was generating EUV light. The project reportedly involves former ASML engineers, Huawei, research institutes and government-linked coordination. Reuters also reported a target of producing working chips by 2028, while people familiar with the project considered 2030 more realistic. Chinese authorities did not publicly confirm those project details.

This is the most specific reported timetable for the integrated Shenzhen machine, but it is not a publicly confirmed government delivery commitment.

Langdao Silex's EUV light source, October 2026: Shanghai-based Langdao Silex publicly displayed a domestic EUV light-source prototype. Reports describe a claimed output above 15 W and roughly 90% localization of components. These are reported prototype specifications, not independently validated production performance; crucially, a light source is only one subsystem of a complete lithography scanner.

The company-reported localization figure is encouraging for component development, but it does not demonstrate that the source is operating at ASML-like power at the wafer, or that a complete scanner can use it at production throughput.

The distinction matters: China has an explicit national policy to achieve semiconductor-equipment self-reliance, but its exact EUV project schedules remain largely opaque. It is reasonable to infer that EUV is a strategic priority; it is not reasonable to present every reported milestone as an official target.

2. Is China aiming only to satisfy its own demand?: My assessment is that the strategy has three layers, in this order.

Goal 1 — Strategic independence: Reduce exposure to export controls and ensure that Chinese fabs can access advanced lithography without depending on ASML or its international suppliers. Highest-confidence objective. This follows directly from the national policy of semiconductor self-reliance and the reported Shenzhen effort.

Goal 2: Domestic industrial deployment: Build enough capability to install, maintain, repair and improve domestic systems in Chinese fabs, initially accepting lower throughput, narrower applications or higher operating costs.

This is the logical next step after a working prototype. It would give China practical experience and reduce dependence even if the first systems are commercially uncompetitive.

Goal 3 — International competitiveness: Eventually offer lithography systems that compete on performance, cost, reliability and service—and potentially sell them abroad.

This is a plausible long-term industrial ambition, but I have not found public evidence establishing an official date for China to export a commercially competitive EUV scanner. It should be treated as a strategic possibility, not a confirmed near-term target. Why I would not assume China intends to stop at domestic supply: its published policy emphasizes technological leadership and the development of globally competitive industries, not merely emergency substitution. The 15th Five-Year Plan's stated ambition is broader than building one machine for one domestic customer.

Nevertheless, the most rational sequence is domestic deployment first. China's first-generation EUV system would have a valuable strategic purpose even if it never became a serious export competitor: it could enable advanced domestic chip production that might otherwise be unavailable.

3. China's short-, medium- and long-term goals: In the Context of EUV below is our assessment of the likely development sequence. The dates are analytical estimates, except where I explicitly identify a reported target.

Short term: 2025–2028: Prototype;

Likely objective: demonstrate the complete EUV process chain at experimental scale.

Milestone to watch: repeatable wafer exposure, followed by successful patterning and functioning test chips.

Reuters' reported target is 2028 for working chips, with 2030 considered more realistic by sources close to the project. Neither date establishes production readiness.

Medium term: 2028–2035: Qualification

Likely objective: move from experimental exposure to a machine that a domestic fab can use for selected production tasks.

Milestone to watch: repeated production of usable chips across multiple wafers and lots, followed by sustained operation at a Chinese fab.

My estimate is that limited domestic production use sometime around 2032–2035 is plausible, but it is a scenario rather than a confirmed schedule.

Long term: 2035–2040+:Industrialization;

Likely objective: build a repeatable domestic EUV manufacturing and service ecosystem.

Milestone to watch: multiple machines installed at multiple fabs, delivering predictable production performance with a largely domestic supply chain.

My base case is that a more credible path to this stage extends into the late 2030s or beyond.

4. How large is the technical gap to ASML?

The main problem is not generating 13.5 nm light. It is turning that light into a reliable, high-throughput lithography process.

The Shenzhen prototype is reportedly already generating EUV radiation, but Reuters described it as crude relative to ASML's machines and identified precision optics as a major obstacle.

A useful new indicator is the Langdao Silex light source publicly displayed in October 2026. Reports put its claimed output above 15 W, although the precise measurement location and independently verified operating performance are unclear.

ASML reported that its NXE:3800E reached 230 wafers per hour at full specification in 2025, while its High-NA EXE platform was progressing through customer qualification. The company reported more than 400,000 wafers processed on High-NA systems by the end of 2025.

That illustrates the gap China must close: ASML is not just selling a light source or a scanner that can expose a wafer. It is delivering an integrated production platform with a mature supplier network, process knowledge, service infrastructure and years of customer-fab experience.

The critical gaps are:

A machine could succeed at several of these and still fail commercially because the system is only as productive as its weakest critical subsystem.

5. When could China match ASML—and when could it dominate? I would distinguish three different meanings of “match.” Match a basic technological milestone. China could demonstrate working chips made using a domestic EUV prototype. This would be a major achievement, but it would not mean that the system matches ASML's production performance.

2032–2038: My optimistic window: Match selected capabilities for domestic applications China might qualify a lower-throughput system for specific chip layers or process nodes. It could be strategically valuable even if its performance, reliability and economics remained far behind ASML.

Late 2030s–2040s: Base-to-optimistic case; Approach competitive production performance. A domestic platform could begin approaching older or contemporary ASML systems in selected metrics if China solves the major optical, source, stage and process-integration problems. A broad, apples-to-apples match across throughput, overlay, availability and cost would take longer and is not assured.

2040+: Potential to compete globally: China could eventually establish a serious alternative supplier, particularly for customers who prioritize supply-chain independence or accept lower performance. But global commercial leadership—and especially sustained technical dominance over ASML—is a much higher bar.

These windows are  of course our judgment based on the technical progress. The optimistic end assumes rapid progress in the precision optical system and reliable source, and successful domestic integration. The base case assumes that some of the hardest problems require several successive machine generations.

Could China actually dominate ASML? I would not make that the base case for the next 15 years. I would assign greater probability to China first building a strategically useful domestic alternative, then closing selected performance gaps.

Because: First, ASML's lead is cumulative. The company reported that its first working EUV prototype dates to 2001 and that it took nearly two decades and billions of euros in R&D before commercial EUV chip production began in 2019.

Second, China is not starting from zero. It can draw on existing EUV knowledge, reverse-engineered hardware, a large domestic semiconductor market and coordinated state investment. Those advantages could shorten the development cycle compared with the original EUV effort.

Third, ASML is continuing to improve its own technology. China's target is moving: it must catch not just today's scanner but the platform ASML will be selling when a Chinese machine is ready. ASML's 2025 reporting describes the NXE:3800E's 230-wafer-per-hour performance and ongoing High-NA qualification.  A domestic Chinese system could therefore be a strategic success without being the world's best scanner. And a Chinese system that matches one ASML generation would not necessarily match the next also.

6. Our central forecast as Photonics Intelligence:

The biggest uncertainty is how far ahead the secret Chinese programme really is. If the reported prototype has already solved more of the optical and stage problems than public evidence indicates, the schedule could move forward. If the current machine mainly demonstrates EUV generation while relying on immature optics and components, the path to reliable production could be considerably longer.

The strategic conclusion is this: China does not need to defeat ASML commercially to achieve its first major objective. It needs to build an EUV platform that Chinese fabs can operate and improve without depending on foreign suppliers. That is a more attainable first milestone—and the one I would watch most closely over the next decade.

Next: China’s lithography ambitions are about more than building a machine that can print smaller features. In Part 2, we examine the engineering bottlenecks still standing in the way, the race to commercialize domestic DUV systems, and how China’s emergence as a lithography supplier could reshape competition with ASML, fragment global semiconductor supply chains, and create new opportunities and challenges across markets from Southeast Asia to Europe. Continue reading in Part 2 to explore what this could mean for the future of the global semiconductor industry.