As we have been experiencing a global semiconductor shortage for over a year, it's difficult to put a finger on a single industry that hasn't been affected in some way, from automobiles to ICUs. But can it be exclusively attributed to natural disasters and COVID-19 or was it inevitable, and even predictable? Will new players change the market, and what side will they take on the US-China rivalry? And how might the current China-Taiwan tensions disrupt the industry even more? In our latest Wikistrat Insider episode, Jan-Peter Kleinhans takes us behind the scenes of the semiconductor world
Jan-Peter Kleinhans is Director of Technology and Geopolitics at Stiftung Neue Verantwortung e.V. Think Tank. His focus is on the analysis of semiconductors as a strategic asset, how resilient the global semiconductor value chain is to external shocks, and how Europe's competitiveness can be strengthened.
Full Transcription:
Marina Guimarães:
Our podcast today discusses the geopolitics behind semiconductors. We invited Mr. Jan-Peter Kleinhans, the Director of Technology at Stiftung Neue Verantwortung. He is based in Berlin, Germany. Here, we will discuss US-China relations, the future of semiconductors and the current shortage going on in this industry. Stay tuned.
Marina Guimarães:
Mr. Kleinhans, Thank you so much for joining us here today.
Jan-Peter Kleinhans:
Thanks for having me.
Marina Guimarães:
So, the world is now experiencing a global semiconductor shortage. Rumor has it, this has happened during the coronavirus pandemic. And my first question to you is why are semiconductors so important, just to give a little bit of context?
Jan-Peter Kleinhans:
Sure. If you think about it, it's hard to imagine a sector or a task in our daily life that does not depend on semiconductors. It goes beyond your smartphone and your laptop, but ICUs and hospitals for pacemakers and for breathing machines, for all the computers that you need in a modern ICU, they all depend on semiconductors. Your typical car, even if it has no autonomous driving feature, has more than a hundred semiconductors or chips in it. Our finance sector, our energy grid, every single part of our life, really depends on access to increasingly specialized and increasingly complex semiconductors. So, they're really a foundational technology for a modern digital society.
Marina Guimarães:
And why is this shortage happening right now, specifically in the last two years, did coronavirus have an impact on it?
Jan-Peter Kleinhans:
Absolutely. To understand why the shortage happened, you have to understand a couple of the characteristics within the value chain. But first, if we think about the external factors that impacted the value chain, we had the demand surge due to COVID. Everybody went to home office, many kids went to homeschooling, so people had to buy new laptops, new tablets, companies had to set up work from home equipment. So, that created quite some demand.
Jan-Peter Kleinhans:
But another factor was actually geopolitics - like you mentioned before. So we had in 2019 the first export restrictions from the US against Huawei, the Chinese mobile company. And that meant that American, and also a couple of international suppliers, were not able to sell chips or software anymore to Huawei. And of course, other Chinese companies were worried that they would be next in line, that at any point in time, the US government could also put them on the entity list, a kind of black list of companies that fall under export restrictions.
Jan-Peter Kleinhans:
So, what they did was - they were hoarding chips that simply filled up the inventory, stockpiled chips, and especially beginning and mid of 2020, that created another demand search for semiconductors and also for semiconductor manufacturing equipment, simply because Chinese companies were buying more than what was justified by demand because they wanted to build up inventory to be prepared in case in the future they would also end up on the entity list. That was the second factor.
Jan-Peter Kleinhans:
So, both COVID 19 and geopolitics led to a demand increase. And then the third aspect was actually natural disasters, a human error, and also lockdown. So, another impact from COVID 19 that happened throughout 2020 and 2021. Of course, natural disasters are nothing new, and human error is nothing new also in this value chain, but it further exacerbated an already tricky situation. So, for example, at one important chemical supplier in Taiwan, because you need a lot of chemicals to produce a modern chip, twice within six months, there was a fire in his production plant for the chemicals.
Jan-Peter Kleinhans:
So, already before these fires happened, the supply of this chemical was relatively strained, but then after the fires happened, the situation got even worse. And that's just one in several examples. We had the snowstorms in Texas that led to power outages at several semiconductor manufacturers. We had earthquakes in Japan, we have the drought in Taiwan. You need a lot of clean water for semiconductor manufacturing. A drought is really what you want to avoid for your semiconductor manufacturing.
Jan-Peter Kleinhans:
So, these are all the external factors that impact the value chain. But then the question is, well, why did they play out the way they did in this value chain? And here just briefly, I want to mention two internal characteristics that played a key role. The one is modern semiconductor manufacturing basically has no spare capacity. Modern fabs [semiconductor fabrication plants] are so expensive that you want to run your equipment 24x7, and you want to utilize your equipment as well as possible to have as little spare capacity. So, underutilized capacity as possible.
Jan-Peter Kleinhans:
So, in mid-2020, fabs were running at 90 to 95% utilization. And already before that, they were around 80 to 85% utilization. If you have such high utilization numbers, it simply means that there is no spare capacity in the market to quickly react to skyrocketing demand, right? If your fab is already running at almost maximum capacity, and then there is a demand increase, you simply cannot react to it. You kind of run out of steam and this will not significantly change in the future because fabs are getting even more expensive.
Jan-Peter Kleinhans:
So, semiconductor manufacturers such as Intel or TSMC or Samsung, and many others, they will always have an economic incentive to have relatively high utilization rates. And that means also in the future, there is no spare capacity or almost no spare capacity in the market. And that makes the value chain very slow to react to skyrocketing demand.
Jan-Peter Kleinhans:
The other factor is that modern semiconductor manufacturing takes four to six months, just the wafer fabrication, so populating all the single chips on a Silicon wafer takes more than three months. And then after this process step, you have assembly test and packaging typically done by other companies. So, at any point in time in your front-end fab, where you populate the Silicon wafer, you have three-month worth of products. Single power outrage for half an hour can destroy three-month worth of production. And that makes the natural catastrophes and human errors such as power outrages or snowstorms disruptive to this value chain. And this is also not going to change in the future. And if we combine that with the fact that due to global warming, we will have more national catastrophes in the future, that poses an even greater threat to the overall value chain.
Marina Guimarães:
Two questions related to what you just answered. Did we see an increase in the price of phones, for example, in our daily lives, because of the shortage?
Jan-Peter Kleinhans:
We saw a price increase across the board. Phones are a little bit of an exception for other reasons. We saw not a steady demand increase for smartphones, but in 2020, there were some quarters where you saw a decrease. But, for example, for laptops, for cars, and also for the manufacturing itself, across the board prices were increased.
Jan-Peter Kleinhans:
So, that goes down to the manufacturer itself, where contract manufacturers, such as TSMC or Samsung or many others, also increased their prices for contract manufacturing of a chip. And of course, the customer of that fab, for example, for an automotive chip, then handed down the increased cost until it reached the consumer. So, for example, even though car manufacturers sold significantly less cars in 2020, they still had very healthy profit margins because every single car they were able to sell for more money.
Marina Guimarães:
Okay. So, second question related to your answer. Do we see a global tech war then between the US and China, particularly involving semiconductors, and is the US investing in semiconductors production internally, like domestically?
Jan-Peter Kleinhans:
Absolutely. Already before the shortages happened, the technology rivalry, I wouldn't necessarily call it a war, but the technology rivalry or technology competition between US and China was definitely intensifying, which means the US applied across-the-board export restrictions to a variety of Chinese companies. We have conversations in the US about restricting research collaborations, foreign direct investment screening, so that there's now less of an influx of Chinese investment companies investing in US tech companies because of increased scrutiny and regulation.
Jan-Peter Kleinhans:
So, across the board, when it comes to technology, we have an intensifying competition, and semiconductors, in my opinion, are really at the heart of it because of what we said in the beginning, that they are really the foundational technology for most of the other emerging technologies. So, if you talk about artificial intelligence, this will not happen without high-performance semiconductors. If you talk about autonomous driving, if you talk about quantum computing, if you talk about blockchain technology, none of that is possible without innovation in semiconductors. And that's why, in my opinion, they're really at the heart of the intensifying US-China technology rivalry, and throughout that process over the past two-three years, the US realized that specifically for semiconductor manufacturing, they lost quite a bit of ground over the past three decades.
Jan-Peter Kleinhans:
So, in the past, the US had a higher share of global manufacturing within the US and just because of how the semiconductor industry developed over the decades, manufacturing was increasingly outsourced to Asia. So, today the most important manufacturing hubs for semiconductors are really Taiwan, South Korea, and increasingly China. This realization for the US that they, let's say, lost ground on manufacturing, led to the efforts that we see right now with the US Chips Act, trying to reassure semiconductor manufacturing, to bring back semiconductor manufacturing to the US with substantial amounts of subsidies. So, the total amount of investment for the US Chips Act is now around 52 billion US dollars that would flow predominantly into subsidizing fabs within the US.
Marina Guimarães:
And why does Asia dominate chip-making?
Jan-Peter Kleinhans:
Several reasons. First of all, a couple of the Asian governments simply realized three-four decades ago that this is an area where they can have technology upgrading, where they can seize a share of the market and build their own ecosystem. So, for example, in the eighties and nineties, we had very similar discussions that we have right now between China and the US regarding semiconductors. In the eighties and nineties, we had between Japan and the US. So, at that time, Japan was the most dominant player within the global semiconductor value chain. And the US were pretty much worried that US companies lost significant market shares in memory chips and in other areas to the Japanese competitors. So, it's really that the Taiwanese governments and companies, the South Korean governments and companies, Singapore is also in that group, Japan historically, that they realize that semiconductors and specifically semiconductor manufacturing, they can play a role in this future market.
Jan-Peter Kleinhans:
And that's for example, why in South Korea the two leading memory chip companies, SK Hynix, and Samsung, dominate more than 60% of the global memory chip market. Samsung is also one of the largest mobile chip manufacturers. You have TSMC in Taiwan, the largest contract manufacturer of the world, the largest foundry of the world. And it's really a combination of absolutely the ingenuity of these companies and investing in the right thing for your company at the right time. But, it's also connected to government subsidies and the support of the local government to invest in that area and to build out an ecosystem plus structural advantages, such as lower energy costs, lower labor costs, and all this in conjunction led to an increasing importance of these countries, specifically the ones that I mentioned, South Korea, Taiwan, Singapore, and Japan for semiconductor manufacturing.
Marina Guimarães:
And looking ahead, there's some rumor that India is trying to become a new semiconductor manufacturing hub. And it is even in talks with Taipei for a millionaire deal. If this is true, what can we expect from the insertion of India in this market in the next coming years?
Jan-Peter Kleinhans:
That's a great question. On the one side, historically, the semiconductor market kind of moved, let's say manufacturing moved from the US to Japan. Then it moved away from Japan to South Korea and Taiwan. So, it could very well be that for specific types of semiconductors manufacturing will partly move to India. And careful of saying this is that today, we also have stronger lock-in effect. So, you simply cannot underestimate the past dependency. So, for example, TSMC in Taiwan has now such a strong ecosystem that goes beyond its own fabs, but close R&D collaborations, R&D centers with international partners, the talent pipeline in collaboration with the local universities. So, on that it makes it much harder for, let's say, a newcomer such as India, regard also for Europe, to get a foot in the door and be a competitive region to the established players.
Jan-Peter Kleinhans:
I fully think that manufacturing will diversify, right now it doesn't make sense that most of the cutting-edge manufacturing is done solely in Taiwan for several reasons. We mentioned natural disasters before, but also obviously geopolitical reasons. So, we should diversify manufacturing, but I don't think that we will see a complete shift within the next 10 years, to say, that Taiwan manufacturing and South Korean manufacturing loses importance and we will see a shift back to the US, to Japan, or to India. I think all these regions will build up a little bit more manufacturing themselves. And there, I also definitely see India. I don't think that these regions will ever be able to meaningfully compete in terms of overall capacity with the current leading region, South Korea and Taiwan, and increasingly also China.
Marina Guimarães:
And is China worried about this possible insertion of India and how likely is it that there will be a partnership between the US and India in this matter?
Jan-Peter Kleinhans:
The last part is really interesting because we had the Quad Leader Summit between the US, Japan, Australia, and India, where they specifically mention the need to increase supply chain resilience to have strategic investments. So, in my opinion, if what happened at the Quad Leader Summit, what's really... If this will unfold as they proposed it, I think an Indian semiconductor ecosystem, let's say might have very close connections to the US, to Japan countries that are historically very skeptical of China. And as we mentioned before, the US-China competition obviously plays a key role. So, in that regard, I could imagine that India as a future, maybe not a manufacturing hub, I think that's definitely too big, but as a new player in the field, would definitely be more aligned with the US than with China. But that's just my guess.
Marina Guimarães:
And the last question: how are China-Taiwan relations in this matter specifically?
Jan-Peter Kleinhans:
China-Taiwan relations are highly relevant in this field because if it comes to a hostile takeover of Taiwan by China, it would definitely threaten the global semiconductor ecosystem because it's highly likely that TSMC would stop functioning or at least be highly disrupted in the event of an escalation. And since TSMC has such a dominant role for the overall ecosystem, such a scenario would be extremely disruptive to the whole value chain with huge spillover damages. So of course, especially for policymakers, the question of geopolitical tensions between Taiwan and China, specifically regarding semiconductors, should be a key concern. At the same time, if China takes over Taiwan, TSMC would also stop functioning for China.
Jan-Peter Kleinhans:
We shouldn't forget that just like European and US and Japanese companies depend on TSMC in Taiwan, so do Chinese companies. So, China also lacks cutting edge manufacturing and Chinese companies such as Huawei or Alibaba, and many others who are designing cutting edge chips, depend on manufacturing in Taiwan for these chips. So, if TSMC stops functioning, these companies as well would lose the ability to manufacture their chips. So, I'm really wondering if we just look at semiconductors, I don't think it's in China's best interest to take over Taiwan because it wouldn't work that they'd simply take over Taiwan and TSMC would continue to work flawlessly. Of course, I think ultimately if it comes to an escalation and the hostile takeover, it would be despite semiconductors and not because of semiconductors.
Marina Guimarães:
Mr. Kleinhans, this was a very insightful view on a very complex topic. Thank you so much for participating and for sharing your thoughts on semiconductors.
Jan-Peter Kleinhans:
Thank you, Marina, was my pleasure.
Marina Guimarães:
This is Marina Guimaraes for Wikistrat.
コメント