Wending Juxin has kicked off its financing round, focusing on domestic lithography machines in the field of self-developed power chips.
Wondersight independently develops self-owned lithography machines to fill the power chip gap and seeks financing
Apart from EUV, a lithography machine track closer to production lines is heating up. Shenzhen Wondersight Technology Co., Ltd. (brand WONDER SIGHT) was founded in May 2023, focusing on the R&D and manufacturing of full sets of front-end stepper projection lithography machines and supporting metrology equipment. Its target customers are not advanced logic chip production lines, but special process production lines for silicon carbide, gallium nitride power devices, Mini/Micro LED, optical chips, and front-end processes of 2.5D and 3D advanced packaging. In September 2025, the company's first WS180i series high-precision stepper lithography machine completed assembly, debugging and was delivered off the production line. Centering on capacity expansion and R&D of medium and long-term models, Wondersight is launching a new round of financing.
Expansion of mature process capacity accelerates, and the delivery lead time and adaptation of i-line steppers become new pain points
Lithography machines are often simplified as the synonym of EUV, but in actual production lines, the most widely used models are still mature process lithography machines adopting 365nm mercury lamp i-line light source and 248nm KrF light source. The critical layer line width of power devices, compound semiconductors and optoelectronic devices mostly falls in the range from submicron to several microns, whose requirement for resolution is far lower than that of advanced logic chips, but puts forward another set of requirements for substrate compatibility, yield and long-term stability. According to the data calculated by Wondersight based on the annual reports of ASML, Nikon and Canon over the years, the global shipment of KrF lithography machines in 2024 was about 206 units, and the shipment of i-line lithography machines was about 265 units, both types of models still remain the main shipment force; China's market demand in the i-line and KrF fields in 2024 was about 15 billion RMB, while the figure in 2020 was only over 3 billion RMB.
The problem lies in the response speed of the supply side. Canon has long occupied a high share in the i-line segment market, and Nikon launched its new i-line product NSR-2205iL1 after 25 years in 2024, with a resolution of 350nm, a numerical aperture of 0.45, and a maximum exposure field of 22mm square, which is clearly oriented to silicon carbide and gallium nitride processing scenarios. For the rapidly expanding domestic power device and compound semiconductor production lines, imported new machines usually face a delivery cycle of about two years and a long after-sales service response cycle, while the existing second-hand equipment faces the problems of accuracy attenuation and maintenance. The more troublesome problem is adaptation: the machines optimized for silicon-based production lines often cannot directly reuse the focusing, leveling and overlay accuracy when facing large warpage, transparent or translucent substrates, and large step structures of silicon carbide substrates, and such process pain points are exactly concentrated in the fastest-growing new energy and industrial control chip chain in China.
Starting from 0.35 micron, taking long-term stability as a core hard indicator
Wondersight's solution is to reverse define the product specifications based on the actual demands of compound semiconductor production lines. Both WS180i and WS235i series adopt i-line stepper projection scheme, support wafers from 2 inch to 8 inch, with resolution ranging from 1.5 micron to 0.35 micron, compatible with various substrates such as silicon, silicon carbide, gallium nitride, gallium arsenide, indium phosphide and sapphire, and support lithography processes for large warpage wafers and different photoresist thicknesses. According to the company, the stable overlay accuracy of the complete machine can reach about 90nm, and the critical dimension uniformity is controlled within ±8nm.
At submicron resolution, what really determines the yield is often not the single-point limit index, but the stability during long-time operation. Wondersight focuses its technical efforts on three directions. The first is high-precision on-line measurement and automatic compensation, which implements compensation for imaging drift caused by ambient temperature changes through real-time monitoring of projection objective lens magnification and displacement of key components. The second is real-time focusing and leveling, using self-developed algorithm to suppress defocus at the edge of the field of view caused by wafer warpage, taking into account both measurement speed and yield. The third is the dynamic stability of the whole machine, which consists of an active vibration damping system, a submicron-level aberration sensor and a microsecond-level piezoelectric feedback mirror group to form a multi-parameter real-time closed loop, and separates temperature and vibration disturbances through a frequency domain filtering algorithm. The company states that the exposure energy fluctuation can be controlled within 1.8%, and the stability under the same thermal disturbance condition is increased to 2 to 3 times of the original level. In addition, the company has built a digital twin model covering optics, heat and materials, and uses adaptive algorithms to reduce the process prediction deviation of third-generation semiconductors, with the model update cycle no more than 24 hours.
In addition to the main unit, Wondersight has simultaneously developed overlay and critical dimension optical metrology equipment (with metrology accuracy of 10nm), ultra-precision optical imaging quality evaluation system, thin film measurement system and high-performance active vibration damping platform. For a newly built compound semiconductor production line, this complete delivery capability of main unit plus metrology equipment and environmental control often has a greater impact on the introduction decision than the parameters of a single machine.
This track does not only have one player. Shanghai Micro Electronics Equipment's 90nm ArF lithography machine has achieved mass production, and its 28nm immersion model is in the verification stage; Shanghai Xinshang Weizhuang shipped its first 350nm stepper AST6200 in November 2025, with a positive overlay accuracy of 80nm. Wondersight's differentiation is more reflected in the delivery and service side: the company states that the price range of the complete machine is between 12 million RMB and 50 million RMB, the delivery cycle is about 8 months, the after-sales service responds within 24 hours, and its target customers are clearly locked in non-IC chip production lines in Chinese mainland to avoid head-on competition with leading manufacturers in advanced processes.
Prioritize non-IC production lines, and roll out to KrF and advanced packaging step by step
In terms of business path, Wondersight adopts a rolling development strategy: in the near term, it focuses on 0.8μm and 0.35μm i-line models, in the medium term, it extends to 0.13μm KrF and 28nm ArFi immersion DUV, covering mature process nodes from 0.8μm to 28nm. Its business will first serve power semiconductors related to automobile and new energy chip manufacturing, as well as front-end processes of 2.5D and 3D advanced packaging. The revenue is mainly from complete machine sales, and supporting metrology equipment, key components and technical services form the second growth curve. Universities and scientific research institutes are another market with clear demand, and driven by the equipment update policy, this part of procurement is basically oriented to domestic equipment.
The team composition is the most valued part of this project. The founder, Dr. Qi, graduated from Huazhong University of Science and Technology with a doctor's degree in optical engineering. He used to work at the Academy of Opto-Electronics of Chinese Academy of Sciences, Shenzhen University and Southern University of Science and Technology. He is a core member of the national "02" special project, an expert of the expert database of the Ministry of Science and Technology, has published 16 papers in authoritative domestic and foreign journals, and holds nearly 350 authorized or published patents in China, the United States and Europe. Previously, he led the development of excimer laser light source for 193nm ArF lithography, cost-effective Mask Aligner lithography machine and 308nm excimer laser light source. There are five other doctors in the company's core team, who are responsible for overall design, simulation, optical design, metrology design and optical system respectively; the executive team covers lithography machine operating system and software integration, semiconductor device process optimization, high-end manufacturing, as well as whole machine vibration control and structural design.
In terms of progress, the company has applied for 32 patents and software copyrights, including 28 invention patents and 4 software copyrights; it currently has more than 2,000 square meters of R&D sites, and plans to expand a 1000-class and 100-class clean manufacturing workshop, equipped with 30 VC-E grade vibration-proof lithography machine integration platforms, with a planned maximum annual production capacity of 80 units, and currently has an annual production capacity of about 20 units. According to the company's plan, the delivery volume will gradually increase from 2 units in 2025 to 50 units in 2029, it is expected to achieve profitability in 2026, enter stable operation around 2028, with a total R&D and production investment of about 298 million RMB. In January 2026, Wondersight signed a school-enterprise cooperation agreement with Shenzhen Institute of Information Technology and unveiled the industry-education integration training base, paving the way for subsequent process verification and talent cultivation.
For investors, the real variable lies in: under the background that Canon and Nikon continue to increase investment in mature process models and domestic manufacturers are also promoting related R&D, whether a company established for less than three years can stably replicate the debugging experience accumulated from the first delivery to mass production delivery with a scale of dozens of units.