In Full: ASML CEO on AI, China, Chips, US Restrictions
You said on the call today without AI, the market would be very sad if you asked me. That definitely spread some sadness in the market in the stock price. Could you give us a sense when you when you talked about the new expectations, you focused on a couple of things, some of your customers, customers like Samsung and Intel, but also China. And so I'm wondering if you could give us a sense of, on balance, how much of your resetting expectations was related to changes in China versus weakness in some of the large customers that I mentioned? Yeah. So I think last week we said a few
things and I think the first thing we said was that we still see AI as a huge opportunity for the industry. And we still see even upside on the short term. So I think AI it's been with us now for a couple of years, has created a lot of excitement. And like all our peers in the industry, I do believe that AI will bring a tremendous opportunity for this industry.
So that's the first thing we have said. So I said, on the long term, AI is going to build up a really strong opportunity for the industry. Now, in the short term, I think you're aware of that. Not everyone is surfing yet on the AI wave. I think some companies have been doing extremely well and of course NVIDIA comes to mind TSMC because they are providing the first one, the products, the second one, the process that is going to manufacture the product.
So they're very strong, I think you see and I'm sure you will see that in the weeks to come, a very, very strong demand. Now what we also said is that not every semiconductor manufacturer is involved in AI. Some of them are doing chips for more traditional application, mobile, PC automotive. And that part of the market is not, you know, recovering as quickly as we hoped. Right. And it does recover. So we see progress and people are reporting that.
But it's slower maybe than a lot of us expected. And I would say that concerned quite a few companies, all the companies, in fact, that today are not directly benefiting from from the AI boost. That's a bit the split of the market we see today at least. Does that mean that we shouldn't expect a full recovery until 2026? Well, that's always very difficult to say because, you know, it is something we we watch every quarter. You know, last week people were maybe a bit disappointed that we didn't give more heads up. But that's something we just keep
watching. And, you know, in the last three months, for example, we have seen that people we were starting to push the brake. When do they start to push the accelerator again? We don't know, because, you know, when you become cautious, it usually take a bit of time before you look at the the longer term again.
But, you know, long term, we see growth. We see tremendous opportunity, which means that at some point of time this has to come. 2025. By the way, we still see as a growth year for the industry. For ASML. the growth is not as high as I think a lot of people expected by then, but still growth here. And we also expect 26 to be a growth
year. But it's too early to, you know, to quantify that in in very high accuracy or very high precision. Got it. I want to talk a little bit about China. At one stage, they accounted for almost half of total revenue. What you shared with the market was that we have to think about China in its more historical perspective, where it accounted for something more like 20% of total revenue for you as a company. I'd like to get a sense from you of how
much of that changing viewpoint on China is coming from weakness internally versus limits on what Chinese manufacturers can buy given export controls? Yeah. So maybe first a step back. If you look at China, you know, we just talked about II. Asia is a very advanced semiconductor manufacturing, very advanced chips. China for a few years now is focusing on
what we call mainstream semiconductor. So semiconductor we used to talk about Iot, you know, all those chips that you have in all the product. Making those products smarter. That's the type of chips China is basically manufacturing today. One part of the reason for that is, of course, the restriction that have been applied on EUV, as you know, which have prevented China to move to a very advanced node.
So the focus of China is on mainstream semiconductor, Right. The demand for that as boomed in 2021, 22 as a result. The demand we had in China in 2021/22 has boomed as well, and 21/22 was the time. If you recall, where the demand was extremely high and ASML was struggling to deliver tools to everyone. Therefore, we couldn't deliver a large part of the demand in China.
Backlog in China grew. 2023 the rest of markets. Often we got tools we could ship to China. Same in 2024. So the level of business we had with China in 23/24 was more of a resource of the non-delivery capacity in 20 to 21 than anything else.
And we have always explained that, you know, our quarter at 50% of the business in China, almost 50%. This was not normal. This was a peak resulting from from basically a lack of delivery before. Right. The normal business in China is around 20, 25%. And that's the normal demand related to mainstream semiconductor. And I think we're going to go back to those number all the time. But there is something artificial about
the way we think about China as a as a purchaser of equipment like Asml's, in the sense that that many countries around the world, U.S. principally have it have have been exerting. Export controls on the country. Are those controls going to become even
more stringent, for example? China's not able to buy your most advanced machines. Is it possible that those restrictions could extend to the duty level, which is which is used for more mainstream products? Yes. If you go back to the the root cause or the I would say the prime motivation of export control, it was to prevent China to have access to advanced technology.
EUV is not in China by not having you in China. We pretty much stop the ability of China to go beyond the five, maybe three nanometer nodes. It may be able to do a few chips and you may see some of that in the press because, of course, everyone is watching very carefully. But fundamentally, by not having EUV, we
have placed China. The regulation have placed China, I would say 10 to 15 years behind when it comes to advanced technology. We always talk about China. We usually forget to mention the fact
that while those restrictions have applied on China and EUV, a lot of work has been done with other player in industry, Intel, TSMC, Samsung on the most advanced EUV technology as we know it today etc., etc.. So there's been a lot of work done every single day to continue to widen if you want this technology gap. An EUV is a key element to be able to do advanced technology. You know, I think Intel has proven to the ward a few years ago that without EUV, life was getting impossible. That's also true for China. So this limit, part of the activity
there, based on your interaction with the prime minister of of the Netherlands with the foreign ministry. Is it your expectation that the restrictions will become even more limiting on China's ability to purchase machinery like Asml's? Well, if you look at the geopolitical landscape, I think it's clear that the United States will continue to apply pressure on their allies for more restriction. Now, the question, the discussion we had with the Dutch prime minister was, I would say the European political leader, because at the end of the day, we are a European technology technology champion company. Right? So these are the people we are having this discussion with. And the question is what is right for the Netherlands? What is right for Europe? And as a result, what is right for ASML? So what restriction makes sense when it comes to national security, which has been the prime argument? And where does it stop? Right. Because like I say, a lot of the focus in China today is on mainstream semiconductor. And this is very different from, you
know, AI. So that's the discussion we have received. The pressure will continue to be there because, you know, if you look at the landscape, that's what it is. But I think that the discussion will get more sophisticated over time. The stake will become higher. And I think that a country like the Netherlands, all of Europe, where again, we belong to. Will also more and more start to
discuss what makes sense for us. Does it make sense for the European Union to be to work in such close concert with the U.S. in in in making it difficult for China to get access to these these machines? At what stage and do you find yourself wishing that the Dutch government and maybe the EU government would push back harder on some of the restrictions that the U.S. would like you to impose? Well, I think, you know, as a European citizen, right, Dutch? I would say resident, not citizen, as you know.
I like to know that they do what is right for Europe and for the Netherlands. I think that's the most important things. And we have always said in ASML we don't make the law. You know, I'm not a politician. I don't know half of what they know, of course. And there may be good reason to restrict. And if there is a good reason to do so and our government believe that, then we would go with it, of course. I think what's very important is to have
the discussion with them so that they understand what we do and maybe understand the industry a bit better and they are better equipped to really decide what is good again overall for European countries and for for the Netherlands. Based on your interaction with your customers in China and globally. Is it your perception that the motivations for export controls are truly about national security interests and regional security interests versus economic competition? In other words, how much do you think this is about keeping China from becoming economically more competitive with the rest of the world versus these are legitimate security threats? Yeah. So, Tom, I could almost turn your back the question way to almost have a referee question. I think in in here and what you see
happening is more and more people are asking themselves exact question. And if you look at the press, I think there is a real debate happening. That's one of the debate. Is it really about national security?
We know that there is, of course, a strong economic competition between, you know, the US, Europe, China. We see it everywhere. People question how much of that has a play in the in the decision. The other debate you start to to to see which is I think is also related is does it help us or does it hurt us? I think there are a lot of companies in the US that start to also question is that a good thing for us? And I think that's the key debate to have because like I said, I think at the end of the day we need our governments to do the right thing for their people.
And I think those debate are far more I see that active, far more lively than they were two years ago. And I think it's very good. I think it's very good because most probably this bring us to what I will call a more rational discussion and a more balanced set of decision moving forward. So I think we can only welcome debate in such complicated issues. Thank you. I want to talk a little bit more. Let's bring things a little bit more local to the Netherlands, where ASML has discussed with the Dutch cabinet it wants to operate successfully in the country. Your predecessor, when asked, said if
the Netherlands shuts down and we cannot get immigrants or foreign students, then fine, you should accept the consequences. Now, the Dutch government, to its credit, has set up a task force that it will spend billions of euros in the Eindhoven region to keep ASML at home. My question is, is that enough? Particularly in a world where there is a coalition led by a far right lawmaker that's looking to implement the strictest ever anti-immigration policies, the cabinet's working on areas to restrict labour migration. You are dependent on labour from not just domestic, but around the world. How concerned are you about the forces at work bringing this anti-immigrant? Rhetoric and policy to the Dutch.
The Dutch government. Yeah. So here also maybe you should take a bit of a step back. I think that this is almost an issue for Europe. What you describe as it's happening in Netherlands, it's happening in many countries as we speak. It's happening in the U.S., it's happening in the U.S. And what we try to do again, in our
discussion with our government, every government wants to have companies like ASML in Europe, in the Netherlands, right? Everyone wants that. The more the better. But to be able to have company like this, you need a few things. You need to have access to capital, access to people, access to energy.
You need to have a place to build your factory. All those conditions have to be there. And if you want to compete with other country, China, the US, whatever, not only you need to have those conditions, but those conditions should be as good as possible so that you can be competitive because this industry is extremely competitive and no one should ever take any success for granted. Right. Of course, the story of ASML is great, but it's only great after 40 years of very hard work. And this is, you know, something you have to to to preserve. So we need all those conditions to be
there. I think, you know, the Draghi report was also a good way to summarize what the entire industry in Europe think. And our government has to walk that. And, you know, of course, there's a bit of you know, immigration is a bit of a topic of the day because that's that's a way to get people attention. And I think what we have said is we have built our company with one word, more than 100 nationalities. So bringing talents from everywhere has
been an absolute condition for success. And this has to continue. Now, the good news is, as you say, and I think people are listening and maybe when, you know, you get down to business, people are more reasonable that maybe when they run a campaign or try to get elected. Does it make sense for ASML to consider locating or relocating some operations outside of the Netherlands as at least a contingency measure to ensure that you have access to those ingredients that you mentioned that over 40 years have made ASML a great company. So the threshold to do that is pretty high for us because we are developing extremely sophisticated machines.
Technology, in fact, is a combination of many, many technologies that we have to bring together. Our supply chain is 80% of it around in Italy and Germany and Italy and France. Italy very close. Our engineer, our R&D is now, of course, almost fully in the Netherlands. We have R&D center also in the US. But I would say the large part of it is
in the Netherlands and one very important thing is because of the complexity of our tools, we want to keep our research development very close to our manufacturing center, because what you may not realize is every two years we make a new step on our machines, so we have to improve them every couple of years, like our customer. And it's made that every two years we build a new machine and we have to teach our manufacturing people very quickly to be able to build a new machine. So there's been a long lasting principle in ASML that we want to keep R&D and manufacturing together. And we would like to do that in the Netherland because this is where we are. So the threshold to change that I think is very high. And I would say new things like geopolitics give us even more reason to stay in the Netherlands so that we can build this very strong relationship with the Dutch government, with the European countries around us to to also get the support of this government in our activity.
So it's not really being considered now to maybe locate some operations elsewhere. We have operation elsewhere further, but we will only grow those operation based on the activity we have in the different countries, but we don't at all consider to move a large part of our operation outside of the Netherlands. Not at all. European Union has a multi tens of billions of dollars plan to expand local manufacturing capacity with the goal of doubling output from the region to 20% of the global market by 2030. In your estimation, is this a realistic expectation and what are some of the hurdles that the region needs to overcome? This is the idea of increasing the region's capacity to be a bigger producer of global of of chips to the global market. Yeah. So I think it's very simple. If you if you want to achieve this kind of target, you have to build capacity. If we look at, for example, where we
ship our system today, most of our system, 80% of our system are still being shipped to Asia. So Asia remains at least for quite a few years. And I would say the key manufacturing place for semiconductor to get to 50% share of the output. You have to take more than 50% of the tools that are going to be shipped. We don't see that yet. So I think that most probably the progress
not as fast as some people hope. I think you have also heard, of course, some some major delay, right, in some of the factory in the US. So this is pointing to more superbly a slower execution. But, you know, you only can start to believe that targets would be achieved when the capacity build is exceeding 30% of the total worldwide capacity, which is not the case yet. In our introduction, Caroline reminded
us that we have an election in the U.S. in a couple of weeks. You might have heard about it. Do you? Do you? It's very close so far in the polls. Would you expect conditions for the global chipmaking world to become better under one candidate versus the other in the U.S.? Well, you know, there's a lot of fear. Of course, like everyone else. You know, I follow the election in the US and you can see a lot of polarization, I think, in the debates. There's one topic that still is, I would
say, shared by both sides. In the US, it's China. That's one. I think one element of agreement. So we don't expect major changes there. We expect a pressure on export control on China to remain. To tell you exactly how this would play
out. I really don't know. To be honest, I think. Do you think it'll be ratcheted up? Do you think that pressure will be you talked about it's going to remain.
Do you expect it to be ratcheted up? I think, you know, I told you the pressure will be there after I don't know. It's speculation. If, you know, one candidate may may want to do more than the other, I don't know. But like I said, that most probably would not come without a debate. And I think that, you know, this is what we've seen happening now for for a few months. But we'll see. At Bloomberg Tech, we like to talk about new products.
You've got a new one coming out high NA EUV. We showed an image of it earlier, $380 million. One of our sizes of the scopes. It's the it's the size of a bus or a size of a small Amsterdam flat. Some of ASML customers have raised concerns about the price tag. One won SVP at a major customer said, I
like the high NA EUV's capability. I don't like the sticker price. What's your what? Are you concerned that the price tag could affect demand for it? And I'm asking this of someone who has about 100% market share in this. So I understand. No. So, you know, I had the question many, many times in the last few months, and I always gave the simple answer. We never build a product in a small with the idea that a new product would be more expensive than the previous one. Of course, the price is higher, but the price of all our product has been higher over time.
And the reason for that is that the value they bring is even higher. So I think the reason why we make a NA is basically to enable customer to look at new design. So to create more flexibility, to look at the ability to go down another two, three, four nodes, which for this industry at the end of the day is the highest value we can ever generate. You know, someone a customer told me, you know,AI can happen today because we have this three nanometer node. We couldn't do a was a previous node because this would have been too expensive, this would have been too much energy consumption. And in fact, for AI to really span out in
all application like we all dream of, there would be a need of another two three nodes to make sure that the cost goes down, to make sure that the energy consumption goes down a NA is going to contribute to that. You know what we dream of in ASML for the next 5/10 years is going to contribute to that. And the value of NA is to continue to enable Moore's Law to go with, of course. Now there's a bunch of other technology, but this is huge value. You know, I was also discussing with Caroline to say, you know, AI is coming, it came, you know, like it's a big bang, right? And we know that as long as we continue to drive the industry, there will be another big bang at some point which no one can even start to dream of. But this has been the story of this industry, and this is what going to contribute to, right? So customer will adopt a NA when the value of it is right and they will continue to use all the other lithography product. Of course, at the same time.
My last question for you before I let you go. We saw during the pandemic the the what can happen when we lose control over the supply chain. Do you feel like it's it would be necessary for ASML to to somehow use its its cash pile to do more to gain control over the supply chain. Have a little bit more power in decision
making whether that would be making acquisitions or increasing the stake in one of. Your suppliers, such as Zeiss. Yeah. So it's a very good question, by the way, because, you know, we said the market is going to be a bit less than we expected in 25. At the same time, we are continuing to build the capability we were missing in 2022. So that part estimate has not stopped because it takes time to do that. And we have, of course, had those discussion with our suppliers, most of our suppliers, our partners. Zeiss is a true partner.
They understand our industry as much as we do. We call a relationship with Zeiss two company one business. I mean, the business is the same and we still have a bit our different ways to do things. What I'm trying to say is that for most of our supplier, we have a very strong alignment to make sure that we can meet the supply chain end of the industry. Only in very rare exception, we are
acquiring some of the suppliers. The reason for that is usually if they cannot scale fast enough, if they cannot invest to the level we need them to invest. I think Cymer was a good example of that. When Cymer had to develop EUV, this was just too much too quickly. Therefore, we brought them in. We have another few example, but that's the exception. We try to keep the supply chain
separated because this allows us even to deliver as we need to be a lot more agile on our product and with our customers. So the model is very powerful, we believe. Christophe, thanks for being with us today.
Please join me in thanking ASML CEO Christophe Fouquet.
2024-10-25 08:21
