In terms of public-private sector collaboration, where do you see the CHIPS and Sciences Act coming into play? Will it be sufficient for U.S. chipmakers to catch up with, and possibly, outcompete the current leaders in the sector like TSMC or Samsung?
I don't think the CHIPS & Sciences Act is likely to, in the short run or the foreseeable future, let the U.S. firms outcompete TSMC, for example. We are going to see more fabrication in the U.S. we otherwise would have seen, though it will probably not be a dramatic change. I also think a key question of the CHIPS & Sciences Act is going to be: what's the impact of the R&D spending? Does it make a difference or not? If it does, it is quite possible that it could really boost American R&D efforts, which could have a real big difference in the long run. But in any sort of scenario, we are not going to be dramatically reducing the U.S.’s or the world's reliance on chip fabrication in Taiwan.
We should also be skeptical of the “catch-up” narrative. The reality is that the chip supply chain is so complex that it is just the fact that no country is going to domestically produce all of the process steps and all of the technologies required. Therefore, the chip supply chain will be international for the foreseeable future, and probably forever. And that means that there is going to be a mix of American, Japanese, Korean, Taiwanese and European firms playing their roles. Obviously, every country wants their firms to be at the top, but that is not a realistic policy goal in this case. The realistic policy goals for the U.S. are to keep its R&D infrastructure very strong, to find ways to mitigate the concern about a war in Taiwan disrupting access to Taiwanese chipmaking capabilities, and then finding a way to maintain or (ideally) grow the gap between what the U.S. and allied chipmakers can do and what Chinese capabilities can do, and what that means for intelligence and military capabilities down the road.
Taiwan will be keen to maintain its dominance in the high-end sector of the semiconductor industry for economic and security interests. However, the U.S. is keen on taking back the global lead on chip manufacturing again. Are there any potential conflicts of interest between Taiwan and the U.S. on this issue, or may American success come at the expense of Taiwan?
It's complex. On the security side, the mainstream view in Taiwan is that Taiwan's role in this industry helps guarantee Taiwan security. But that is far from clear. If there is a massive amphibious invasion, will that be deterred by the fact that such an attack would disrupt supplies and hurt China as well as the U.S.? Maybe that is true, but a massive invasion is not the most likely way a war in the Taiwan Straits is going to start. In fact, anyone looking at the world over the last 10 years would have concluded that Russia's massive invasion of Ukraine has not gone well, whereas Russia's 2014 annexation of Crimea worked very well for Russia. In a Crimea-style scenario in the Taiwan Straits, for example, if China occupies one of the small uninhabited islands in the Taiwan Straits – which it could plausibly do bloodlessly in that type of scenario – it is not clear to me that Taiwan's criticality in the chip industry helps it. In that scenario, China would grab an island and look at Taiwan and the U.S. and say, "All right, your move. What are you going to do? Are you going to get our forces off of Taiwan's territory?" In that scenario, Taiwan's economic importance does not deter China. Instead, it deters the U.S. from coming to Taiwan's help. So firstly, Taiwan is wrongly overconfident that its centrality in the chip industry deters China, when the more likely military escalation scenarios would find Taiwan's chipmaking deterring the U.S. So that's on the security side. On the economic side, you could argue that there is a bigger dynamic of competition between Taiwan and the U.S. in terms of chipmaking. Even there, it is easy to overstate it. There are a lot of U.S. fabless firms that have every interest in TSMC’s success, and not every Taiwanese chipmaker has the same interests as TSMC. So there is actually a lot of complexity and variability in economic interests when you look at both the U.S. tech sector and the Taiwanese tech sector. Therefore, there are some competitive aspects that need to be managed, but I think it's easy to blow them out of proportion. Whenever I get interviewed by the Taiwanese media, they are always focused on the competitive aspects, which to me seem a lot less important. In my perspective, the U.S. is preparing itself to defend Taiwan militarily, risking World War Three in the process and in that context, whatever competitive dynamics in the chip industry just seem a lot less important.
There is a prominent view that TSMC is an example of how global supply chains could be affected and therefore preventing military conflict. But you would say otherwise. How can the U.S. kind of balance this viewpoint while enhancing its own capabilities?
What we've seen this year with Russia and Ukraine, and Russia and Europe is that there is not much evidence for that argument. Russia and Ukraine were Europe's biggest trading partners before 2014 when Russia annexed Crimea, and Russia was Europe's biggest supplier of gas and Europe was Russia's biggest customer for oil and gas. And yet, that is all being unwound as we speak. The empirical evidence for the claim that interdependence creates peace is not really as strong as is often assumed in popular discussion. What we see right now is a growing realization on multiple sides that interdependence can, in certain cases, create incentives for peace, but they also create competitive dynamics where different countries try to use leverage against their rivals. Given those dynamics, what we are actually looking at is that as countries realize that interdependence creates positions of leverage, every country wants to have the best positions of leverage as they maneuver through this game. A lot of what you see with the U.S. and China right now is about that. The CHIPS & Sciences Act is also largely about that. The U.S. has a really strong position in the chip industry for historical reasons, and wants to maintain that position, especially in the face of China's effort to domesticate a lot of these key technologies.
The Dutch company ASML has a complete monopoly on Extreme Ultraviolet Lithography (EUL) machine technology. However, it is often neglected when chips are spoken about in the context of Taiwan Straits tensions and superpower competition. What role or leverage do you see in ASML, when it comes to gatekeeping the frontier of chip innovation against China or any other country for that matter, and potentially balancing out superpowers in the chip war?
ASML is not unique in this regard. The other equipment makers, like Applied Materials for example, also play a similar role in doing their leading-edge innovation right alongside their main customers. That includes Intel, TSMC, Samsung. For some of the toolmakers, they've got facilities and they are doing R&D, and they've got Samsung and Intel and TSMC employees doing their R&D alongside them. That is gatekeeping right there, because if you are on the inside, you are literally on the inside of their R&D facility. And if you are on the outside, you are on the outside.
There are no Chinese firms on the inside, so they are not going to serve as a bridge between the West and China, or the U.S. and China. The firms on the inside glue together the Europe, Japan, U.S., Korea, Taiwan supply chain that I've described. The fact that you have a leading-edge firm in different parts of the supply chain, and in each of those countries that I mentioned, means that they are all working together to some extent. They are competing as well, but they are also all part of the same supply chain, while China is still on the outside and it will be so for as long as U.S. policies can keep it on the outside.
The U.S. is trying to convince and push its allies, like Japan, and the Netherlands in particular, to adopt similar sets of controls as it has. The challenge that the U.S. faces is that companies are still in China because they want to sell their products and maximize profit, which is to be expected. The more interesting thing is that the countries that are working with the U.S. on this issue – Netherlands, Japan, et cetera – have a lot of incentive, even in places where they agree with the U.S. on its policies, to let the U.S. pay the price and take the blame. If you are a small country like the Netherlands, you are looking around at countries that have angered China recently, looking at the sanctions that China has imposed on Australia for example, or Korea with a fab issue five or 10 years ago, and saying, "Well, China is not going to do that to the U.S. but it could to a smaller country. So we do not want to stick our neck out on this issue. Let's just let the US handle this one." That explains a fair amount of how the other countries besides the U.S. approach this issue. There is some disagreement, but there is also a fair amount of trying to not get involved because they don't want to get caught up in the superpower conflict.
Despite China's heavy investments in their domestic chip industry, it is still struggling to build significant chipmaking capabilities that could launch the country to the edge of computing innovation. What is the issue here? Could China potentially build their domestic capabilities to catch up with Taiwan and the U.S.?
It is certainly possible China can catch up, because why would it be impossible? But it is going to be hard for a couple of reasons, but primarily because of the number of places in the supply chain where China is still far behind. It is not just one part of the supply chain – it is the entire supply chain. The design tools, the materials, the chipmaking machine tools, a lot of the designs. Across the supply chain, China is importing technology from abroad and although China today is nine or so percent of the global chip industry, that is mostly at low-tech technologies. So anything cutting edge, China is importing. That is a lot of complex technologies for one country to master quickly. It is not saying anything about China's capabilities; that would be hard for anyone to do. The fact that China is going to have to try to do it, while being deliberately locked out of the innovation ecosystem in a lot of these spheres by the U.S., is going to make it even harder. Historically, the way companies caught up was by integrating really deeply. If you look at TSMC or Samsung, the way they have gone from low tech to cutting edge is not by trying to do everything on their own, but by trying to do a couple things on their own. Then, by buying the best components and the best tools and working with the best customers they could, that is how you learn. That worked brilliantly for Samsung and TSMC, and for South Korea and Taiwan more generally. China is not going to be able to pursue that strategy, because the U.S. is going to lock it out. So China finds itself in the position of a) still being pretty far behind b) really struggling to get access to cutting edge equipment and materials because of U.S. controls, and then c) you have the question of: does the Chinese innovation ecosystem work as well as it could? We could debate that. Its track record is fine, but not great. Even if it was great, the existing challenges China has to surmount are just huge. If you put it in the context of the extraordinary progress that both Taiwan and South Korea have made over the last 30 years going from basically zero in the chip industry to huge players, it is also worth noting that they have done that in only one part of the supply chain. TSMC is the world's critical supplier of chipmaking services, but it also could not do any of it without important tools from Japan, the U.S. and the Netherlands. Self-sufficiency is just going to be really, really hard.
We see military deterrence being quite central to this competition for computing advantage between the U.S. and China. Referencing back to what you wrote about in your book, do you see the military industrial complex coming into play again with chip innovation and enabling the U.S. to suppress Taiwan Strait tensions?
One of the differences between now and 50 to 70 years ago is that now, the U.S. military is trying to find ways to use civilian technology which is more complex and advanced in some cases beyond what the military can currently field. Whereas in the past, militaries often added unique capabilities. Today, if you buy a new iPhone, the chip inside that iPhone will be more advanced than anything the U.S. military currently uses, so that dynamic is really different than six years ago. However, it is also the case that the defense concerns are becoming more front and center to the chip industry than they have been over the past few decades, precisely because of concerns over security in the Taiwan Straits. Hence in some ways, we are returning to the military playing a big role, but it is in the context in which militaries all over the world are trying to find ways to access the best commercially produced chips.
Yuri Samoilov, CC BY 3.0 <https://creativecommons.org/licenses/by/3.0>, via Wikimedia Commons