The AI CEO No One Talks About

Kind: captions Language: en Bloomberg just interviewed the most important CEO in the entire AI supply chain. And you are pushing hard on this. What what is your message? Really critical point about Korea towards the end of the interview. Along the way we're going to see about supply and demand for AI. We're going to hear about Terafab and Orbital AI data centers. Was a surprising point about Europe's role? Hint, he's a European CEO. Critical point about India's role. And then this big bomb about Korea. Are you ready? Let's go. >> What is it that Europe needs to do to ensure that it is not vulnerable to those kind of supply chain risks? >> If you don't have big enough part of a certain ecosystem, in this case AI, you're going to be exposed to the goodwill of other parties. >> Christophe Fouquet is the CEO of ASML, which is the critical link [snorts] in the supply chain for everything that's happening in AI. >> We're going to be looking at a supply limiting market for AI for quite a few years. >> ASML makes the lithography machines that go in the chip fab. They have a monopoly on the EUV lithography machines, EUV, EUVNA, and they have a very, very large share of the DUV machines. So, basically, none of this stuff happens if his company doesn't put out machines that go in fabs. The fabs stop growing, all the growth gets cut off if this guy doesn't make it happen, if his team doesn't make it happen. >> What's very important is make sure that innovation can happen in Europe. >> We will continue to see products produced if they stopped producing or if they stopped growing. But the growth that everyone is pricing into all these stocks depends on this guy's company growing fast. At the beginning of the interview, he's asked about Europe. My sense is he thinks that Europe is not very important in the AI story, and he encourages Europe to do more to give itself a shot at participating outside of what his company does. There's a couple other things like Mistral in France and Google DeepMind in the UK partly, but his company in the Netherlands is critical, Zeiss Optics is critical, and outside of that there's not a lot happening in the world of AI in Europe. >> The European market is pretty attractive, right? You look at 22% of GDP. It's a very mature market, very good market. So, this means we are user, we are buying a lot of stuff. >> One of the things that stood out to me is basically the only real value he sees in Europe is as consumers of AI. He doesn't see a large role for Europe in the production of other than what his company does, the production of the hardware for AI, or the development of software and models for AI. It's just not a conducive market to it. So, the first thing that comes to his mind is, well, we buy stuff. That's the Europe's place in the world is we buy stuff. >> Sometime we discuss a lot semiconductor manufacturing, but most probably not the right place to start because semiconductor manufacturing is only needed if you have people to buy semiconductor. Today, if we had a 10 nanometer fab in Europe, most probably most of the wafers will go to the US. >> He basically just told you that Europe wouldn't even buy chips if they made If they made chips in Europe, Europeans wouldn't buy them. The chips would just go to the US. So, he doesn't see a point in chip manufacturing in Europe. This is This is like the leading European in AI hardware, and he's saying Europe doesn't have a place in manufacturing chips. >> Today, the US is a clear winner. Looking at champion across the entire AI semi ecosystem, they were missing out was manufacturing, and I think they have been extremely aggressive in bringing some key company to manufacture in the US. They can do that because they buy chips. 80% of the advanced chips manufactured worldwide is bought by the United States. China has been investing also across the entire ecosystem. I think most probably they they do very well on the application side, bit less on what's before that, but I think they are driving the use of AI most probably more aggressively than any other country. And again, when you look at Europe, a lot of holes, quite behind what's happening elsewhere today. >> The United States is winning. The United States buys 80% of the advanced AI chips made in the world. >> [clears throat] >> The United States covers all five layers of the stack. China is doing really well at the applications layer and they're they're trying to succeed at every layer in the stack. And Europe has a lot of holes. Europe is just not not competitive compared to the US and China. And this is from the leading European in AI infrastructure and he's saying a lot of holes. I I hope he was saying holes. >> Are we underestimating the capex spend? >> The demand for AI infrastructure is still enormous. We're going to be looking at a supply limiting market for AI for quite a few years because the build-up of the infrastructure is is huge. >> This is one of the biggest takeaways from the video. The supply is going to be limited for quite a few years. We are in a supply limited market. We are not in a demand limited market. The the challenge of building up the infrastructure to develop all this AI, to build out these data centers, is just a big big challenge. He's going to talk about this a little bit more in a minute. I can say on my end, as a user of AI, I was working on generating images for this video and for my options special video, and Gemini froze on me and ChatGPT froze on me in generating images. I hadn't exhausted my limits on the accounts. They just ran out of compute. Grok did not run out of compute. At least on me, Grok did not run out of compute at a time when both ChatGPT and Gemini were refusing to provide me images, even though I was well within the bounds of my paid accounts. So, this supply limitation on the world of AI, it's going to be biting all of us for two, three years if not longer. This guy's at the center of that supply chain limitation. He sees it both in his company and he sees it in his supply chain and he sees it in those around him. >> Early in this still, where are we in in that? >> Even a year ago, the AI company were already very vocal about the need for this architecture, the need to build up capacity. My industry, the semi industry was still a bit hesitant to move and the move really happened at the end of last year. And now we all need to build capacity, so our customer have to build fab. We have to make sure we build enough equipment to fill those fab. And a catch up will take quite a bit and that's only for the infrastructure. >> So, the reason for this delay in the supply chain is the semiconductor industry was hesitant to grow. There's a cyclical nature to the semiconductor industry. They've seen these rises and falls. We've seen this with the memory makers. We're seeing this with the chip makers. Now, we're seeing this with the infrastructure, the supply chain. They were hesitant to grow fast because they just didn't trust that it wasn't another cycle. And the AI companies came to them and they said, "No, we really need this." And it took until the end of the year before they really started to turn things around and it's going to take them a while to catch up. And I I suspect the growth is going to happen so fast that these guys will not grow as fast as they need to and we're going to be supply limited for even longer than he's talking. >> You're seeing a pull forwards of orders for your DUV and EUV machines from customers. >> We see a pool with see customer continuing to create visibility, longer term visibility telling us, "This is a bit here to stay." Trend we see is still very strong towards basically demand. >> Create visibility means some kind of actual formal written commitment. Make this stuff and we'll buy it. The big fear of the supply chain is that they commit to making improvements in increasing their scale and then the buyers dry up. They need contracts, written contracts saying you're going to buy these things that we're going to make. ASML's machines are like three, four hundred million dollars a piece. So, if they overbuild capacity to build these machines and then people don't buy them, they're in trouble. This is a problem that the semiconductor industry has faced again and again. So, increase visibility is code for they gave us a commitment that we can rely on. >> Data centers in space >> doesn't necessarily change the number. I think the discussion about data center in space is to maybe address another potential bottleneck in the ecosystem, which is energy. There's an idea that if you do that, the energy consumption or the energy availability will be there. Maybe not too much to do with the total data center capacity, maybe more about trying to solve another potential big issue, which is energy for data center. >> Short-term, I don't think he's wrong. Orbital data centers are not going to be a big impact on total demand for data center hardware. Long-term, I think he's missing the story that where Elon's going for much, much, much more scale. There's a limit to how much energy you can get on Earth, and if you really want to do the Dyson swarm, if you if you want to use as much of the energy of the sun as possible, only a little bit of that energy hits the Earth. So, he's not on board with the the Dyson swarm yet. We'll get back to the ASML CEO in a minute with all his updates and that big story coming about Korea. But first, our primary AI stocks index, ASI Stock Report X, was up 3%. You see TSMC up big, AMD up, Broadcom up. The only ones that are down are Palantir and SpaceX, which isn't part of the index, but we've added it, down to $185, still well above the $135 IPO price. The AISRY, our secondary index, up 4%. Micron, Marvell, continuing to be the highly volatile stocks that we've seen before. Surprising that TSMC uh jumped so much, cuz that's been a quiet slow-moving stock. The only big news I see about TSMC is they got a big deal for developing more infrastructure in Arizona. Positive, but it doesn't seem like something to me that would drive the stock up that much, and I don't see anything else in the news. The hyperscalers, Amazon, Meta, Oracle, and Microsoft up slightly. Nvidia and Google as usual in the middle of the pack. Just TSMC for some reason an outlier here. I think we can just call this a volatility day, and it's an upside volatility day. Let's get back into this king of the supply chain and his big news both about TeraFab next and then about especially about what's happening in Korea. >> Elon Musk is also looking at a tera- a TeraFab. What's the potential upside for ASML on that? >> You know, any added capacity of a of course a potential upside for for us. New projects are opportunities as long as you're not supply limited. The TeraFab is is an example of a major fab project. >> So, he's not jumping and not getting too excited about orbital data centers or TeraFab yet. It's too early yet in the process. He's making sure they got all their eyes dotted and tees crossed before he starts to worry about that. It's not an imminent demand situation for ASML. They have customers who are getting who are already building fabs and already need the machines now. Apparently, he did not get the clear signal he needed from Elon to start increasing capacity more for the TeraFab project. At least that's what he said they're they're evaluating it. So, they're not they're not there yet. But, he drops a big bomb next. >> If you look at some of the DRAM project in Korea, you also look at million of wafer fabs. So, that's also pretty big. >> That's the big bomb drop. The millions of wafers fab projects in Korea. This is something that I haven't seen many people talking about. When we talk about memory, a lot of people in America focus on Micron, which is the American memory chip maker, and they're big. But, the largest memory makers, the most important memory makers are in Korea. SK Hynix is the leader and the developer of high-bandwidth memory. Samsung is also very big in high-bandwidth memory and DRAM. Micron's important, but what's happening in Korea is there's massive investment in building more fabs. And that means massive demand for ASML's hardware to to fill out fabs that they're building in South Korea. So, something I haven't seen anyone talking about is the South Koreans are getting ready to build much larger fabs for much more memory because there's a massive memory shortage. And Micron's doing whatever they're doing, but everyone is underestimating the South Koreans. Samsung and SK Hynix are going all out. They're competing with each other ferociously, and at the same time there are projects where they're working together with the government to build out massive fabs so that they can really gear up and deliver more memory. And I just saw that Apple is getting ready to raise prices because the the drive in the impact of AI demand for memory is so big that Apple is now struggling to get memory for iPhones. HBM memory is so valuable that they're crowding out production of DRAM memory. You have NAND memory also that's critical for iPhones that there's a lot of demand for that in AI, and all of this is driving up the demand, and there's a shortage. And we're talking about potentially two or three hundred dollar price increases on iPhones. And on iPhones, you might be able to absorb that price increase. When that hits the Android market, people aren't paying as much for the vast majority of Android phones. This could be a huge challenge for the cell phone market. And so, Samsung and SK Hynix see this, and they're ramping up their production very, very fast. And that's going to require lithography machines. So, when you think we're going to need lithography machines for Terafab, we're going to need lithography machines for TSMC, don't forget the Koreans. The Koreans are going all in. And that's that's the the big bomb that I haven't seen anyone talking about. South Korea's where the action is in memory. China is working on it, but China's well behind. Micron is working on it, but I haven't seen the scale of investment coming from Micron that's happening in South Korea. >> Korea and the build out there is going to be a driver as well. Staying with the US though. >> Do you see it? He's presented with this huge issue with Korea and he says, "Okay, but staying with the US though." Like His audience isn't interested in Korea. He's not interested in Korea. This guy just dropped a massive bomb on you. He's like, "Well, yeah, but let's get back to the US. That that Korea stuff that we can worry about that later." No, that's the big story. You missed it. That's it's so shocking to me when I when you see it happen just like that. Millions of wafers. That sounds like a big deal. He seems to think it's a big deal. Maybe I should ask him more about that. No, no, no, let's stick with the US. >> The Indian Prime Minister, Narendra Modi, will be here in Paris tomorrow. You've partnered with Tata. How much of a growth story potentially is India? >> We are partnering at the very start, at the very beginning. India today doesn't have manufacturing for semiconductor. By 2030, they believe that 10% of any chips will be used by India. First fab will come up next year. Growing opportunity for the not the next 2 years, but for the next 5 and then 10 years in India. >> At least cautiously optimistic. I think he's painting an overly rosy picture. We'll see if India's able to succeed. What I've seen in India is number one, they have tremendous talent, they have great people, and the bureaucracy in India it just always seem to find a way to shoot themselves in the foot. I think for humanity, for Earth, for everything that we're doing, for the future of AI, it would be great if India was able to succeed. I just I can only say that what I've watched from India so far is when it comes to competing on a global stage in any industry, they tend to fall apart somewhere along the way. Primarily not not because of a lack of talent, but the government, the bureaucracy, the oligopolies, whatever it is, something in the Indian economy and the Indian society that just keeps them from really succeeding on something big like this. I hope I'm wrong. I hope Tata's able to deliver and we see something big out of India, but my prediction is this doesn't go anywhere and it's not a big deal for ASML and it's not a big deal for AI stocks and and AI in the world that we're paying attention to. I published my latest options special video today for YouTube channel members, X subscribers and Patreon supporters. Please consider supporting me on one of those platforms. This was my my longest option special video yet. Went in-depth both incorporating SpaceX now in the analysis and looking at three different strategies for dealing with these long-term call options that we haven't talked about before. YouTube channel membership, X subscriber, Patreon supporter, please consider joining. Thank you so much for watching.