Since artificial intelligence is now dominant, it's a fight of capital. You make money from capital, then reinvest, make profits, reinvest again, and it should continue like this. The virtuous cycle has a lot of capital and profits, making it possible. However, new companies can't keep up. Money is crucial in the semiconductor industry. Even China is pouring in a lot of money. Without that, they probably couldn't have grown like this. Recently, Korea announced semiconductor support policies, not subsidies but indirect support, totaling 26 trillion. Will this keep up with competitiveness? This time, we're looking into the global semiconductor industry. Recently, China's SMIC ranking in the world's foundry.

There was news that it rose to third place. How do you see the outlook for SMC in the future? If you see it, I will show you the chart once. It's one minute. It has now risen to third place in the first quarter of this year, surpassing UMC and becoming a global pariah. It has been a long time coming. The difference is not much. Anyway, now that it has risen to third place, if you look at it, the production capacity has also significantly increased. If you look at the production utilization rate, it is now over 80%, and similarly, the wafer shipments are now at 1.79 million units. It's quite a lot. So, in my opinion, it seems that this trend will continue in the future because SMI is the number one in China's foundry. The Chinese government is providing tremendous support, giving huge subsidies. It's a state-owned enterprise, not a private one, so it has no choice but to provide support. Especially if you look at it in China, now

There are a huge number of foundry companies, around 3,500, and now many companies have no choice but to rely on SMIC. Due to sanctions imposed by the United States, China now has to nurture its own semiconductor companies in any way possible. In order to do that, SMIC must play a significant role. Foundry companies must rely on SMIC to produce certain items, so it is expected that their importance will continue to grow significantly in the future. From my perspective, the demand for semiconductor memory chips is expected to increase in the future, followed by system semiconductors improving gradually starting this year. They are receiving demand. Furthermore, as demand increases, the SMC strategy is not about becoming the number one company, but rather expanding their market share no matter what.

So, now this is offered at a very low price. The profit margin is not much. We will somehow expand the market share. So, we will probably try to catch up with Samsung Electronics. If you look at China, they account for about 50% of the global semiconductor demand. Therefore, even just based on domestic demand, there is a huge demand. So, as you can see, about 80% of SMR's revenue is currently generated domestically. There is a huge demand. Similarly, the Chinese government will continue to provide massive support in the future, so there is a high possibility that revenue will increase significantly. Currently, SMRC is developing 5-nanometer semiconductors, and they will also develop 3-nanometer semiconductors. However, internally, it seems impossible to develop 5-nanometer semiconductors using DU equipment. This is what I see as a high possibility.

The problem is that when you start investing, the costs become enormous. Usually, as the unit price decreases, the costs increase significantly, or there is a possibility that they may even go negative. If you look at the SMRC total profit margin in January 2022, it exceeded 40%. However, if you keep watching, it continues to decline. So, the profit margin is not really a concern. Due to the support from the Chinese government, even though they are working on 7nm and 5nm semiconductors, they don't really care about the profit margin. In my opinion, it seems like it will continue to decrease, and there is a possibility that it may even go negative. If the Chinese government stops supporting them, then there will definitely be problems. I see it as lacking competitiveness. In my view, companies need to make a profit, and if they can't make a profit, what will happen in the future in terms of competition? Even if the Chinese government provides a lot of support, I don't think that will continue.

If you don't watch, there will definitely be a certain limit to capital at this point, so it will definitely be a problem then. I don't think it will be a big threat. NVIDIA's market capitalization is now ranked second or third. Right now, Microsoft, Apple, and NVIDIA are strong. As much as that, even after the stock split of 1 to 10, it is now $125. Anyway, it continues to rise. Will NVIDIA continue to grow in the future? I think it will continue to grow. I can't predict exactly how long it will last, but maybe for a few years. Because the ecosystem is so strong. For example, Kakao is the same now. Almost everyone is using it, about 90% to 95%. The ecosystem has already supported it, so because it is a CUDA ecosystem, now we use CUDA.

If we don't, now it's difficult to develop artificial intelligence, especially if you look at NVIDIA recently, you can see that the development speed of their AI accelerators has increased significantly, but there are almost no companies that can follow. NVIDIA is expected to become more monopolistic because of this. The reason for this expectation is that there are no companies that can keep up with that speed. Also, there are no companies with a comprehensive ecosystem. Many companies talk about developing AI chips, but clearly, some of them are criticized for having low performance. However, we shouldn't just focus on the chips. We need to provide a full stack service. This means we need to cover everything from hardware to software, and then services and all other aspects need to be supplied together.

There is no company that can follow NVIDIA in terms of capabilities now. Will this continue for a while? NVIDIA is doing a lot more than just data centers or supercomputers. Recently, many countries in Europe, the Americas, the Middle East, and Asia are building data centers and supercomputers. Due to the risks in the US-China conflict, they are trying to independently build an AI ecosystem. These countries are all using NVIDIA chips because they have the best performance. Since all countries' companies and even the countries themselves are using them, I see a significant possibility of increasing revenue. In the future, NVIDIA will also continue to focus on autonomous driving.

I'm holding on to almost 70% of autonomous driving. If autonomous driving becomes more popular in the future, there will inevitably be a significant increase in sales. Therefore, I think it will have to be pursued to some extent in the near future. Recently, there have been news about the revival of the semiconductor industry in Japan. How is the current situation of the semiconductor industry in Japan? In the 1980s, Japan was a strong country in this field, especially in memory, surpassing 50%, 60%, or even 70%. However, due to severe sanctions imposed by the United States, Japan suffered a significant setback. Now, the government is taking a very proactive approach to regain its past glory. Recently, a factory for sushi was completed, right?

From now on, starting from the 10-nanometer process, we are likely to begin mass production from the latter half of this year. In that way, we provided tremendous support to attract TSMC. We provided huge infrastructure and financial support. It amounts to around 40% compared to that. Isn't the United States also investing now? However, even now, the construction is not completed yet. Moreover, they started in the United States first, but Japan started late. However, they have already completed it first. Japan has several advantages in this regard. For example, labor costs are quite low. Labor costs in Japan are lower than in Taiwan, but in the United States, they are significantly higher, more than double. There are considerable merits in that aspect. Similarly, Japan's small and medium-sized enterprise ecosystem is quite strong. TSMC's weakness is that the small and medium-sized enterprise ecosystem is.

It's somewhat weak compared to other things, but now the semiconductor ecosystem is number one in Japan. It's quite a warning. As we continue to give warnings, we can fill those gaps. Japan provides tremendous support. They are planning to build a second factory in the latter half of this year. This factory will produce 6-nanometer semiconductors. The amount of support is even higher than the first one. Japan needs to continuously improve its process technology, so they provided more support because they need to keep advancing to 6-nanometer, 5-nanometer, and 3-nanometer. Japan is trying to catch up. There are talks of possibly building a third factory. It seems like it will be a 3-nanometer semiconductor. Oh, it seems like it will be around 2013. In this area, Lapis, a company, is also building a foundry factory. They are currently working on nanotechnology.

So, now in 2027, we are currently making plans. Originally, in Japan, the technology level was around 40 nanometers, but now with the introduction of EUV, we are working on 12 nanometers. We need to keep pushing forward. We only have about three years left to plan for the 9-nanometer semiconductor. Since Japan lacks the technological capability, we have partnered with IBM to supplement it. IBM has experience in developing nanoscale semiconductors, so they have the technical know-how. Therefore, we are sending engineers from Leaders to them. Recently, there was news that we will also collaborate on packaging. There is a possibility that the 9-nanometer semiconductor could be ready by 2027, with a slight chance of delay. Anyway, the plan is to proceed in 2027. This is considered Japan's last chance.

Japan is now in the process of implementing the fourth industrial revolution, so I sometimes think it will not go according to plan to some extent. Recently, there seems to be an increased interest in Taiwan again, such as Computex 2024. Will Taiwan continue to hold an important position in the semiconductor industry in the future? Yes, there is a possibility. Currently, TSMC plays a significant role in the semiconductor industry, especially in silicon wafers, with over 90% market share. TSMC's market share is quite high at 62% in the first quarter. In addition to TSMC, there are many other foundry companies in Taiwan, such as UMC, Vanguard International, and PSMC. Furthermore, there are many backend companies in Taiwan as well. Taiwan holds over 70% of the global market share in the foundry sector.

I think we will continue to move forward in the future. Of course, to some extent, now, Japan, well, because they are also doing business overseas, there are some aspects that are a bit scattered. However, I believe that in the future, we will continue to grow to a certain extent due to TSMC's revenue. Recently, ASM's stock price has also risen a lot. And some time ago, news came out that TSMC and ASM had purchased new equipment. As for ASM's future, I think it looks promising. I believe it will continue to improve. AS has just released new equipment, right? Yes, that's correct. So, they have increased the size of the lens. In the past, the number was 0.33, but now they are increasing it to 0.5. What's good about increasing it? It improves the focusing power. It means that it can gather light better. The resolution increases. You can draw circuits more finely. When you draw circuits more finely, you can place transistors

Now, it is possible to put in a lot more, much better semiconductors are coming out, more superior performance semiconductors are coming out. There are no companies that can do that now. Originally, in the past, they said they wouldn't introduce it in the market because the price is too high. The reason is that the equipment costs about 200 billion won, but the high NA equipment costs over 500 billion won, almost double the price. They said they wouldn't introduce it because it's too expensive. Until recently, but now there is competition in the field of artificial intelligence, right? Intel introduced it first, and now they are using it. Samsung is the same, now they are introducing it, so they can lose their leadership. It's not going to work, so they will introduce it. The three major companies will all introduce it. Probably, as early as the second half of this year, or soon, the quantity is also not large.

The amount that can be produced in a year is only about six due to the price competition, so there is no choice but to face competition. Memory companies are also now preparing for introduction, especially Micron or then again SK Hynix, these companies are also preparing for introduction. In the future, equipment needs to be brought in first, so this kind of competition arises again. Eventually, it will increase to around 20 or 30. Then, the revenue will have to continue to increase. Recently, many companies are making a lot of efforts to replace the EV process and ASML equipment. Many companies, like SMEs, are also doing this, but the technology is significantly inferior. There is almost a difference of more than 10 years. Recently, Canon, Nikon, and other companies are also doing this. If you look back, Canon, Nikon

These companies, like duv equipment in the 1990s, were much stronger back then. As ASML introduced EU equipment, the situation has now reversed. Canon is now developing new equipment with a slightly different approach. While ASML uses EUV equipment, Canon is now focusing on nanoimprint equipment. Normally, EU equipment uses photoresist to apply light and create patterns. However, this new method involves stamping patterns like making a stamp impression. It can achieve up to 5-nanometer processes and is relatively inexpensive. In my opinion, to some extent, they will continue to do well. Moreover, it's not just Samsung doing this, right? They are also in this competition. Competition among companies like this is inevitable. NVIDIA almost has a monopoly, it's the same thing. From what I see, they can take a lot of the market share.

Because there are so many, but now it's not easy to replace, maybe for the next 10 years or so, we will continue like this. In this session, many global big tech companies like TSMC, ASM, NVIDIA, and others have emerged. Among these big tech companies, where are some good companies to invest in now? Companies like Apple, Microsoft, or Meta are all in the era where artificial intelligence is dominant, so it's a fight of financial strength. Companies that invest in financial strength will inevitably take the lead. In this financial strength competition, there is no company that can surpass big tech companies because a virtuous cycle needs to happen. Capital is earned, then invested, profits are made, then reinvested, and it needs to continue like this. These companies have a lot of capital and profits, so such a virtuous cycle is possible. However, for new companies,

It's impossible, we can't follow. What can we do but go with the flow? I'm trying to expand my semiconductor market. Recently, Chairman Wei of TSMC and Chairman Chee of SK met. What was the purpose of their meeting? Let me show you a chart. Isn't NVIDIA working on accelerators? GPUs are produced in Sushi. When NVIDIA designs, they manufacture it, and now they are producing it. HM also needs to supply, and currently SK Hynix is supplying HBM 32. So, we are cooperating. Then, TSMC will take the GPU HBM 3 and make packaging to create accelerators. We have been doing well so far, but the problem ahead is, what is it? From HBM 4 onwards, the form changes slightly again because we need to go to the next level, which significantly increases the cost. SK Nix is in the foundry business.

However, this is not advanced, it is a traditional process. Because of this, we cannot perform advanced processes. We need to create a test chip for advanced processes. The base die is now logic, right? We stack HBM like this, don't we? Then we go in underneath. We control and communicate with HBM. SK Hynix cannot do this alone, so we need to collaborate. It is crucial with MC and that is the most important. It's essential because the performance will vary significantly depending on how we package it in the future. We have gone through this partnership to warn you about that. So, if you look at it, now we are also working with AMD, right? NVIDIA is also making efforts to deliver. So, Samsung is also making continuous efforts like this, so we don't know how this will turn out, but in this way, moving forward, to some extent.

Now, both companies may not see benefits anymore. From the perspective of SK Hynix, they definitely need NAND, and then they also need DRAM. Not long ago, SK also visited DiA, right? It has no choice but to do so. Initially, 8-inch foundry companies were in a serious recession, but do you think they are recovering now? It's getting a little better. Domestic companies are struggling, especially DB HiTek, which is currently operating in the 8-inch sector. Looking at the first quarter, although they are not in the red, they still suffered a significant blow. Companies like SMIC or Nexperia, and also Hwahong Semiconductor, are abundant in China. These companies offer their products at extremely low prices, causing a lot of pressure on 8-inch companies. In this recession, competition is fierce.

It seems that there is not much profit, but the semiconductor market will continue to improve in the future. Even the 8-inch process benefits from the 4th industrial revolution. So, not only advanced processes but also older analog semiconductors will see market improvements as the 4th industrial revolution progresses. Companies with 8-inch facilities, as well as those with 10-inch facilities, are all used in various applications. Especially, in the 7nm and 7nm under semiconductor sectors, there are only a few players, so despite the recession, the growth potential is high and will continue to improve in the future. This year, the growth rate is expected to be around 13%. In the advanced process area, like silicon under, the market growth may exceed 20%. Companies in advanced processes benefit significantly as many new applications emerge due to the development of the 4th industrial revolution.

For example, smart factories, IoT, and supercomputers like autonomous vehicles, various applications like these will inevitably advance. Many founder companies are likely to benefit greatly from these situations. AI is currently a major trend, and in the future, it can be expected that the AI semiconductor market will continue to grow. What types of AI semiconductors are there? AI semiconductors are now being developed, and in about 10 years, it is projected that one-third of semiconductors will be AI semiconductors. It is highly likely that they will be used a lot in the future. If I show you a chart now, it is expected to grow significantly. By 2032, it is estimated to grow to around 27.4 billion dollars. In my opinion, not only domestically but also internationally, there are many opportunities.

If it's happening, the flow will likely change a lot. There are many types of AI semiconductors. Usually, they are divided into generations like 1st, 2nd, 3rd generations. The 1st generation typically has better CPU performance than GPUs. The downside is that it processes serially, unable to process simultaneously. However, GPUs have lower CPU performance but can process in parallel, making them quite suitable for AI semiconductors. This is the 1st generation, still widely used for GPUs. Moving on to the 2nd generation, there are FPGAs and then ASICs used for semiconductors. These are mainly what we refer to as MPUs, specialized for artificial neural network semiconductors, tailored for AI. Unlike CPUs, which are not specialized for AI, MPUs are quite expensive due to their power efficiency and price. They are designed specifically for AI, making them significantly cheaper in terms of power consumption and price. In terms of inference, MPUs excel.

I think it will be used a lot, especially recently. If you look at it now, there is on-device AI, right? It's being used a lot in areas like smartphones, and recently, there are also AI laptops coming out, right? Because things like PCs are coming out, it will be heavily integrated in that area. Next, we have to move on to the 3rd generation. Up until now, memory and semiconductor subcontractors have been doing everything, so it's quite efficient in terms of processing and storage. It's also very good in terms of power efficiency and avoiding bottlenecks. This is the only way we can go forward. Many companies like Intel and Samsung are working on this, so I think we will continue like this in the future. Especially with autonomous driving at level 5 and humanoid robots, they have to come out, so to make this happen, the semiconductor industry must account for at least %.

The world is betting on nurturing the semiconductor industry. So, what are the key factors that South Korea needs to address to stay ahead in the semiconductor market? Well, first of all, the most important thing in the semiconductor industry is money. Capital is crucial; it's basically a war of money. Companies with a lot of money have an advantage. Even now, China is pouring in a huge amount of money. If they hadn't done that, they probably wouldn't have grown this much. Money needs to be invested significantly because capital is needed to build fabs, conduct R&D, and most importantly, to attract talented engineers. If you offer a lot of money, you can attract them. Not only China, but Japan as well. Recently, they have been providing huge subsidies. The United States is also providing massive subsidies. This is how they directly

They are giving money, but recently Korea announced semiconductor support policies. It's not just a subsidy policy anymore. Even indirect support of 26 trillion won, can this keep up in terms of competitiveness? Because if a lot of capital is invested, they have to make it cheaper. Prices go down. When it comes to the semiconductor industry, it's not just competing within a country. From a capital perspective, companies might be somewhat disadvantaged. There is a concern that once you fall behind here, it's hard to catch up in the gap. Yes, so if you follow, you have to invest more money. There is no guarantee that following will work. I've heard Professor Shi talk about semiconductors, and this time, the DX leader has also published a must-read book on the semiconductor world. Lastly, in terms of semiconductors, promising domestic and global companies.

If I were to give a word about stocks and investment, as I mentioned before, if someone has a tendency towards high risk and high return, then I think it's not a bad choice to consider investing in ASM. I see it as a good choice. If someone still prefers stable investments, then investing in companies like Samsung Electronics, which continue to grow steadily, could be a pretty good option. Clearly, when a company performs well, the stock price goes up. However, this is not reflected in the short term, but rather over a period of 2 to 3 years or even longer. Even if a company performs well, external factors or negative events can cause the stock to drop. So, despite good performance, it is always important to consider a long-term perspective of at least 3 years.

Invest in stocks because in about 3 years, most companies go through certain variables, usually within a year. Therefore, companies with good performance in these 3 years will definitely grow and their stock prices will rise. So, it's recommended to invest with at least a 3-year perspective like this. Today, I saw that NICE rose another 3-4% during trading hours, hitting a new high. Of course, there are many investors who have made a lot of money with NICE. It is hoped that when NICE takes the lead in market capitalization, like Samsung Electronics, many people in the country who hold Samsung Electronics will benefit. I hope that HBM, which is at the forefront of the pillar in the semiconductor industry, will have its day. Next time, I will invite a professor again to create good semiconductor content. Today, our special guest is Professor Kwon Young.

So, thank you all for spending your valuable time with me. It was Kim Ja, thank you for your hard work. Thank you.

AI, 반도체 아직 못 산 분들은 '이 3개 주식을' 사 모으세요 (권영화 교수)

출연 신청/제휴 문의: smilekdy7@naver.com

🌟 김작가 TV를 구독해
'재테크, 자기계발, 동기부여 콘텐츠'를 간편하게 받아보세요.
- 평일 오후 5시 / 주말 오전 10시 영상 업로드

📚 김도윤 작가 도서 출간 (11권)
[1번째 책: 동기부여] 날개가 없다, 그래서 뛰는 거다
[2번째 책: 취업] 인사담당자 100명의 비밀녹취록
[3번째 책: 몰입] 최후의 몰입
[4번째 책: 기획] 기획에서 기획을 덜어내라
[5번째 책: 공부법] 1등은 당신처럼 공부하지 않았다
[6번째 책: 유튜브] 유튜브 젊은 부자들
[7번째 책: 에세이] 엄마는 괜찮아
[8번째 책: 자기계발] 럭키
[9번째 책: 경제경영] 머니 트렌드 2023
[10번째 책: 경제경영] 머니 트렌드 2024
[11번째 책: 에세이] 내가 천 개의 인생에서 배운 것들


AI, 반도체 아직 못 산 분들은 '이 3개 주식을' 사 모으세요 (권영화 교수)

Scripts