The US and China's trade row is continuing. Five years have passed since the inauguration of the Trump Administration that started the row but the two superpowers' economic competition isn't relenting even in the Biden era.
The US-China trade row will undoubtedly affect South Korea's semiconductor business. Will the situation be an opportunity or crisis for South Korea?
The Elec met with SkyLake Investment CEO Chin Dae-je to get his take on the situation. Chin is known in South Korea as a key executive at Samsung that laid the foundation for the company's semiconductor business. After previously working at HP and IBM, he joined Samsung in 1985 and was instrumental in the company's development of 64MB, 128MB and 1GB DRAMs. He was president at Samsung as well as the Ministry of Information and Communication of Korea. He founded SkyLake Investment in 2013.
In his interview with TheElec, Chin said while the situation for South Korea's chip industry wasn't easy, the country had the competence and potential to turn the crisis into an opportunity.
"The US controls half of the world's chip industry. South Korea controls around 20%. That means South Korea has a leading position in the global semiconductor industry. The country also has purchasing power in fab equipment and materials," Chin said.
"China is spending heavily in its chip sector. But the semiconductor sector isn't an industry where you can just throw a tonne of money into development. It won't be easy for China to catch up to South Korea because of this," the CEO said.
However, South Korea needs to invest not just in memory but non-memory chips and form a balanced ecosystem, Chin said. In the long-term, South Korea needs an environment where there are labs and talents that develop semiconductor technology as well as one that fosters startups, the ex-Samsung president said.
Following is the interview between TheElec's JY Han and SkyLake Investment CEO Chin Dae-je:
Q: What is South Korea's position in the global semiconductor market? The US and China are actively fostering their respective semiconductor industry. How is South Korea's semiconductor competence in comparison?
A: In terms of numbers, the global semiconductor market is worth around US$450 billion annually: The US controls half of this and South Korea is the second-largest in terms of share. South Korea heavily focuses on the memory sector but this still accounts for 20% of the world's semiconductor market. Europe accounts for around 10%, Japan also accounts for 10% and China and Taiwan each control around 5%. Though South Korea is behind to China and the US in terms of absolute economic power, it still has a leading position in the global semiconductor market.
In the memory market alone, South Korea controls around 70% market share. This means it has substantial purchasing power in buying fab equipment and materials. When Samsung or SK Hynix buy fab equipment or materials, they can ask related companies to assemble or produce them in South Korea. Most of the time the equipment and material companies will comply with these requests. This shows South Korea's leading position in the sector.
China is a manufacturing powerhouse __ it controls half of the world's manufacturing market. But unfortunately, it relies heavily on overseas import for components and materials that go into the products it manufactures. For chips, it relies on foreign imports for 80% of the parts and materials it uses to manufacture them. The Chinese government wants to reduce this reliance, so it has set a goal called Made in China up to 2025. It wants to internalize the parts and materials it uses in chip production. This poses a problem for the rest of the world that relies on China for manufacturing. This can be seen as the catalyst for the US and China's trade row.
China had announced its aggressive spending plan for its chip industry in 2015. It had earmarked 300 trillion won for the plan __ where did the money come from? Samsung spend around 300 trillion won in China from 1983 to 2015. So in a way, China is spending the money Samsung spent in the country to catch up to Samsung. Six years have passed since this plan was announced __ and for multiple reasons, I don't think China's plan went well. This is because the semiconductor industry today is very different from what it was in the 80s and 90s where companies and countries could catch up to the leading companies and countries.
First of all, semiconductor technology has become very difficult compared to decades ago. Chips have become so small. Microfabrication has reached its limits __ this has been said for the past 20 years. But it continues to evolve __ which means the companies and countries working in the sector are working with extremely difficult, high-level technologies. This means simply throwing in 200 trillion won or 300 trillion won won't translate to catching up to these leaders in terms of technological level. The trade row between the US and China has also blocked China from using equipment with US' intellectual properties in them. China has been merging and acquiring semiconductor companies but this road has been blocked also because of the trade row.
Q: What is your opinion on the US government requesting information from semiconductor companies?
A: They are requesting information but they are not likely requesting manufacturing technologies. I think they are requesting information to gauge China's semiconductor procurement trend, so they want to know how many semiconductors are sold to which customer or country. Through the information, they can control the supply and demand of chips, which can be used as an anti-China strategy. As the chip industry is one of the most important in the world, the US is feeling the need to manufacture them on its own soil. It is also feeling the need to control the needed technologies. The US is actively trying to build fabs on its own soil.
Q: Intel seems to be an active participant in this.
A: Intel is facing difficulties in its businesses. Intel's microprocessors are general-use CPUs, but companies that had used these CPUs like Amazon, Google or Microsoft are now developing their own chips because of AI. Companies had relied on Qualcomm for mobile application processors but the likes of Apple and Samsung are developing their own. The market has changed. As more companies like Google and Amazon develop their own chips, it has become a difficult environment for TSMC to be the sole manufacturer of chips. Intel has its own fab and has technologies that allow it to enter the foundry arena. So for Intel, it is natural for it to spend in its own manufacturing fabs.
Q: There are more talks that South Korea's system semiconductor fabless companies are facing difficulty. It is a high-value market __ many new companies had been formed in the past in South Korea, but their earnings are not growing.
A: There are multiple reasons why South Korean fabless companies are facing difficulties. In the past, let's say there were a hundred chips in a handset. These companies could provide some of the chips. But today, a single application processor handles all the workloads previously provided by the hundred chips. Intellectual properties have merged into one big chip. Past companies that conducted business with one or two key intellectual properties have been acquired by larger companies. There is no place for small fabless companies.
However, chips needed in small markets likely electric blankets or rice cookers are controlled by China. Chips with low value are mostly made in China this way. There are some 3,000 fabless companies that can make such chips. In South Korea by comparison there are only around 100 companies.
Q: More people are saying that it is difficult for fabless startups to form in South Korea.
A: It's true. Startups will design chips that fit certain systems and markets in their mind. But they don't understand the system. There needs to be a lot of companies that work on system chips. But in South Korea, there is only Samsung Electronics, LG Electronics, Hyundai Motor, SK and Hanwha that manufacture end-products, which is a system. This has also caused the decline of South Korea's component industry. In the 2000s, there were around 40 handset companies in the country. Now there is only one [Samsung].
However, the Fourth Industrial Revolution is giving a lot of opportunities for fabless startups to form. AI, IoT, autonomous driving and wearable sectors are active. It is important for startups to react quickly to the changes in these sectors. They need to understand the information and trends in sets and systems. It is impossible now for college graduates to start a semiconductor startup. They need working experiences at set companies before forming a startup.
For example, those working at global companies, let's say they come up with a new idea in AI. They will start their startup in Silicon Valley. This is still difficult because of the cost. However, there is a limit on how much designers companies like Google, Amazon or automobile companies can secure for their own chip development. Startups that help these companies in chip development are growing fast. These are contractors for chip design.
Q: South Korea has quite a lot of companies working on materials, components and equipment used by the chip industry. Are they competent?
A: They are not bad. Some of them were excellent in batteries or solar panels; some are great in equipment and materials. They won't catch up to ASML, but they can offer the equipment needed by Samsung or SK Hynix. In materials, they won't be able to manufacture the most advanced materials but can manufacture general-purpose materials. Also, it is important to keep in mind that ultra-high-purity materials like hydrogen fluoride are so small in terms of market size that it is just convenient for South Korea to rely on Japan. There is no need for South Korean companies to do everything. Not everything needs to be made in South Korea and market conditions must be taken into account.
In the automobile and battery sectors, which South Korea is strong in, related material, component and equipment companies are doing well. System semiconductor is different. The set needs to be strong, which is not the case. Would China buy our materials, components and equipment related to system chips? They likely won't use it because the price will be too high. For materials, components and equipment related to system semiconductors to flourish, higher-tier companies need to also be doing well.
Q: How to grade the current South Korean government's K-Semiconductor strategy?
A: The government and companies say they will spend 270 trillion won. I think they should use the opportunity to focus their energy on non-memory sectors, not just memory so that the whole of South Korea's chip industry can grow. In the US, President Biden is conducting strategy meetings for procurement of chips, South Korea should open similar meetings to actively foster its non-memory sector.
I think SK Hynix is doing well. It has said it plans to spend in memory, system LSI and foundry. It is difficult for South Korea to maintain its chip competence with Samsung alone. Samsung, TSMC, Intel and SK Hynix are the ones buying expensive fab equipment. South Korea's spending needs to focus on this area.
It is the same for the government. Abroad, there are many research labs that conduct research with a focus on what is needed five to ten years down the road. For South Korea's chip ecosystem to have balanced growth, there need to be organizations that focus on basic technologies.
Q: There is a talent shortage in South Korea's chip industry right now. Can this be eased and what can universities do?
A: Fostering talent is very difficult. Even if the government pushes a policy to foster talent, these talents need jobs. Compared to 20 to 30 years ago, talent has become rare because they have few job options. At Samsung, there are many people so it's hard for them to get promotions. So it's hard for them to have hope for the future; so they end up not applying. However, there weren't enough applications for computer engineering in the past: but this has increased thanks to AI and blockchain. So talent is related to the market and industry.
Q: There are many Samsung 'Old Boys' (ex-Samsung men of the previous generation) who think that Samsung Electronics is at its most critical crisis right now.
A: Samsung is now posting close to 300 trillion won in revenue per year. There aren't many companies in the world that can reach that number. It is highly profitable and it has exceptional talents. So generally speaking, it is hard to see Samsung as in a crisis.
However, I think what the Old Boys are worried about is that in the past, chairmen Lee Byung-chull or Lee Kun-hee directly solved problems. This isn't happening right now. Samsung does face difficulties, but it has CEOs that are doing well so I don't think there will be a sudden crisis. Samsung also has semiconductor and display businesses that backups its smartphone and TV businesses. It may lack in software but it can differentiate in hardware, so it is stronger than most companies. Samsung could last longer than Apple.