ASML Holding NV ADR (ASML.US) Q1 conference call: Expected growth years in 2025 and 2026, future AI demand gradually becoming stable.

Recently, ASML Holding NV ADR (ASML.US) held a conference call for the first quarter of 2025 financial report. The company's total revenue for the first quarter of 2025 was 7.7 billion euros, in line with expectations. The revenue from lithography systems reached 5.7 billion euros, with EUV sales at 3.2 billion euros and non-EUV sales at 2.5 billion euros. The logic chip sector accounted for 58% and the storage chip sector accounted for 42%. Equipment service business revenue was 2 billion euros. Executives stated that the semiconductor market has recently experienced strong growth driven by artificial intelligence, and future AI demand is gradually solidifying. However, global economic uncertainty has increased due to tariff discussions. Based on discussions with customers, it is expected that 2025 and 2026 will be growth years, with product portfolio trends leaning towards advanced logic chips and DRAM. Revenue is expected to be between 44 billion and 60 billion euros in 2030, with a gross margin of 56%-60%. The company's gross margin for this quarter was 54%, higher than expected, benefiting from installed EUV systems achieving customer production efficiency goals, favorable EUV product mix, and higher ASP from high configurations. Research and development expenses were 1.161 billion euros, sales, general and administrative expenses were 0.281 billion euros, both in line with expectations. Net profit was 2.4 billion euros, accounting for 30.4% of total revenue, with earnings per share of 6 euros. In terms of revenue guidance, the company expects revenue to be between 30 billion and 35 billion euros in 2025, with 2026 being a growth year. For the Q2 of 2025, the company expects total revenue to be between 7.2 billion euros and 7.7 billion euros; installed equipment management business revenue to be approximately 2 billion euros; gross margin is expected to be between 50%-53%, with a wider range than usual due to uncertainty surrounding tariffs, scale, and value chain impact; research and development expenses around 1.1 billion euros, and sales, general and administrative expenses around 0.3 billion euros. The growth of artificial intelligence remains a key factor in industry growth. If AI demand is strong and customers can increase capacity to meet demand, it is expected to reach the upper limit of performance expectations; otherwise, it may fall to the lower limit. Q&A: Q: Are you currently considering flexible pricing to promote faster adoption of high NA technology by customers? A: Generally speaking, single exposure lithography is better for customers than multi-exposure lithography, as it can simplify processes and reduce costs. We want to promote the adoption of new lithography systems as quickly as possible. Low NA technology is still under development, and high NA technology is also being promoted. The main reason customers do not adopt new systems for single exposure is not the price of the tools, but the level of maturity. If the maturity is not there, the cost of the technology cannot be optimized. High NA technology is much more mature at the moment compared to low NA, so that is where our efforts are focused. Currently, we are working on improving the maturity of high NA technology. Lowering prices when maturity is low will only create confusion for the customers because the tool might not be reliable. Q: You mentioned that in the low NA area, EUV single exposure is being adopted more compared to multi-exposure lithography, especially in the DRAM sector. Can you talk more about the timeline? Is it around two years, or can we see opportunities for this application sooner? A: Every time we bring tools like the 3800E that can optimize the cost of technology for new customer nodes, it is a good opportunity for promotion. This is a long-term effort that we have been working on with customers for some time. Currently, the low NA EUV technology is mature and efficient in production, with the 3800E being 30% faster than the 3600D. Q: You mentioned growth next year, what should the order booking rate be for the next two quarters? If the upper limit of performance is reached in 2025, can we continue to grow in 2026? A: We will not discuss the growth rate specifically. Based on technology, customer feedback, and market demand, although there are macro factors affecting this, we still believe we can achieve growth in 2026. Regarding the amount of orders needed, it is hard to give specifics, and although we can see the current backlog of orders, many of these are related to post-2025, the order situation for next year looks good, but new orders are still needed to ensure growth. Next quarter, our focus will be on this, as the nature of the orders cannot accurately reflect the business momentum, so we will not quantify it. Q: Do you still expect the sales in the Chinese market to account for around 20%? Can you talk about the composition of the backlog orders? A: This year, the share of the Chinese market will slightly exceed 25%, it was mentioned before that it was over 20%, but now there have been changes in DUV demand. In the backlog of orders, the portion from China is still roughly in the range of 20%-25%. Q: Regarding the tariff issue, how have your discussions with customers been over the past few weeks? Do customers have any interest in changing delivery times? How do you consider delivery times with regard to avoiding tariffs by shipping early? A: After the announcement of tariffs, our business discussions with customers have not changed. On one hand, the uncertainties mentioned earlier still exist, and in the short term, customers and suppliers are still uncertain. Everyone is trying to understand the impact, but so far these discussions have not fundamentally altered the business plans. On the other hand, many customers find it unrealistic to act as advised, mostly due to space constraints. Having a semiconductor fab take in equipment early for tax reasons is not feasible. Q: You mentioned that the 3800 model with low NA has become the main model for shipments, how much inventory is left from the 3600 models that were built last year? How are the sales of these devices? Can they be upgraded to achieve productivity close to the 3800 model and be resold? A: No need to worry about that. Q: The comprehensive average price for low numerical aperture EUV is around 230 million euros, with a gross margin of approximately 55%, is that correct? Previously, you mentioned equipment upgrades, are there any one-time or special circumstances affecting the average price, gross margin, or involving IBM (services for installed equipment)? A: The calculation is correct, the average price is about 227 million euros. Considering the products include the 3800 and 3600 models, and various configurations, the price is at a slightly higher level. However, looking ahead, the comprehensive average price is expected to decrease, particularly for low numerical aperture lithography machines. When modeling, the aim is to enhance the comprehensive average price, so there could be a decline in the future.The average price is set at around 2.2 billion euros, which is more suitable.Regarding gross profit margin, the company will not disclose on a product-by-product basis. It has been previously mentioned that the gross profit margin of low numerical aperture lithography machines is currently significantly higher than the overall gross profit margin of the company. This is indeed the case, but specific data is not separately disclosed at this time. Q: Previously, it was expected that the revenue share of the Chinese market in the group would exceed 20%, now it has risen to slightly over 25%. What is the reason for this? Also, due to limitations in wafer factory maintenance services, can you provide a rough indication of the revenue related to services for installed equipment for IBM in China? Please inform about the plans for this in 2025. A: In terms of these few percentage point changes, it actually involves changes in the hundreds of millions of Euros. However, this falls within the normal range of business dynamics and not significant fluctuations. It is evident that the demand in the Chinese market remains strong, with mainstream business demands being particularly prominent. Currently, the demand for mainstream chips in China is strong, meeting both domestic consumption needs and supporting exports, making this demand highly resilient and even more substantial than expected three to six months ago. Q: There were reports in the middle of this quarter that China is close to developing its own EUV lithography tool alternatives. Can you share your latest views on this and discuss its scalability and reliability aspects? A: As it stands now, there have been no substantial new breakthroughs. It is expected that there will be further reports on China's progress in this field in the future, mainly driven by China's strong desire to showcase achievements. Essentially, even though some images have been shown, they mostly represent research progress rather than product results. Certainly, the development of extreme ultraviolet light and even the preparation of extreme ultraviolet reflection mirrors are achievable, but this does not indicate that mature products are on the verge of being released. Therefore, we still believe that it will take many years for China to successfully develop EUV lithography machines. Furthermore, such reports are expected to continue to emerge, as showcasing interim results is common. Q: It was mentioned that AI demand is stabilizing, and NVIDIA Corporation's GTC event indicates that the increase in inference models is driving more semiconductor demand. Can this have a positive impact on ASML's long-term potential market expectations? A: According to information from logic chip and storage chip customers, demand in this field is still strong. In the past few quarters, the industry has been highly focused on the artificial intelligence training phase, and now, with the further development, the focus is shifting more towards the inference phase, which is logical. In the long run, the share of the inference part in artificial intelligence demand is expected to increase, and future developments in this area are worth continuous attention. Q: Due to tariff impacts, the gross profit margin range for the second quarter has been widened. What are the expectations for the full year? How is the full-year gross profit margin directly affected by tariffs? What factors have been taken into account in making this full-year forecast? A: Tariff dynamics are constantly changing, making it difficult to provide exact numbers. On one hand, the final tariff situation between regions is unclear, especially in the semiconductor field, as the US government is still reviewing and the scale of tariffs is uncertain. Secondly, even if it becomes clear, how the entire value chain will absorb and digest the tariffs is also a question. We are closely working with all parties to minimize the impact of tariffs on the value chain, and we believe that once the impact is minimized, the main burden should not fall on ASML Holding NV ADR, but on the next part of the value chain. With so much uncertainty, it is impossible to judge the full-year impact, and as the second quarter is short, we can try to incorporate the impact broadly, but we cannot provide specific numbers for the full year. Q: There is a viewpoint suggesting that in the future, if wafer fabs are more diversified geographically, such as having more production in the US and China, it will ultimately benefit wafer fab equipment expenditure (WFE), and the current tariff situation seems to support this. I would like to know your latest views on this, including the impact of geographical diversification on business and communication with customers. A: I think this viewpoint is correct to some extent. Building wafer fabs globally would likely require increasing capacity to ensure wafer production. However, during this process, there would be efficiency losses in the ecosystem, as evident in semiconductor manufacturing and data center distribution. The trend towards ecosystem diversification is clear. In the long run, this will increase semiconductor and wafer demand, as having wafer production in multiple locations will naturally increase demand, which is beneficial. However, the current uncertainties should not be underestimated and must be considered comprehensively, weighing the pros and cons. Q: From the development stage of the EXE 5000 to the mass production of the EXE 5200, what are the key milestones? Can you discuss production volume or other related factors? A: Generally, there are three stages: the first stage where customers receive the first equipment or use our ASML Holding NV ADR laboratory equipment to test new technologies. Since it is brand new and involves imaging and performance aspects, customers need to verify if the design meets the standards. Most customers are currently in this stage, which is why we are happy when they share results, indicating smooth progress. In the second stage, customers choose a small number of layers and products to test with the more mature and higher productivity EXE 5200 in early production. We have already started delivering this equipment. Once this stage is completed and customers confirm the equipment is usable and mature, it can be fully used for future large-scale manufacturing nodes by 2027-2028. The process is long, but it is necessary across multiple nodes, and currently customers are progressing as expected. Q: Given the key points you mentioned earlier about the February SPIE conference (International Society for Optical Engineering), the advanced logic sector seems to be the first adopter of this high NA technology. So, who will be the second adopter of this technology, foundry factories or DRAM? A: It is difficult to say. To be honest, I have mentioned before that the time is very close. Previous discussions involved customers who could use laboratory equipment, including logic and memory chip customers. Once the equipment is mature and cost-effective, they both have reasons to adopt high numerical aperture technology as soon as possible. It was often thought that the logic chip sector would take the lead, but this time it is hard to distinguish between the two in terms of timing. Q: I would like to know the percentage of revenue the company receives from the top four chip manufacturers in China. Has this percentage decreased over time or remained unchanged? A: Over time, the percentage of revenue from the top four chip manufacturers in China has decreased. More participants in the Chinese market (with a more pronounced long tail effect) have emerged, but a large portion of the products shipped to China still go to large enterprises, while small customers purchase equipment.Despite the small quantity, a large portion of China's sales revenue still comes from large enterprises.2021ASML5000EXE:52002025ASMLEUVDRAMASMLASMLASMLWill we be able to start receiving third-stage orders for these high NA devices by the end of the year or in the second half of this year?A: There won't be too much, maybe just a small amount, but it's still early. We are mostly still in the first stage, still installing some equipment, which will take more time.