What does the CHIPS Act mean for U.S.-China semiconductor competition?

Below is a complete transcript of the China Corner Office Podcast with Rory Murphy and Paul Triolo.

Chris: Hi everyone. Thanks so much for joining us today on China Corner Office, a podcast powered by The China Project, the New York-based news and information platform that helps the West read China between the lines. I’m Chris Marquis, a professor at the Cambridge Judge Business School.

And today, we are joined by Rory Murphy, who is vice president of government affairs at The U.S.-China Business Council. We’re also joined by Paul Triolo, who is senior VP for China and Technology Policy Lead for the Albright Stonebridge Group. We discussed the significance and impact of the recently passed semiconductor subsidy bill, the CHIPS and Science Act of 2022. While this live recording was before Biden’s recent executive order on implementing the Act, that was in progress, and Paul and Rory provide some insight on that as well. We started our discussion with the history of the CHIPS Act, including why and how it took so long to pass, and also an overview of the legislation, specifically how more than 200 billion in funding is going to be deployed.

Rory gave detailed comments on the political dynamics of the Act and how the midterm elections in the U.S. will impact different sectors. Paul then looked at the big picture and discussed some of the other regulations that function hand in hand with the CHIPS Act such as export investment controls around manufacturing and chips used for AI. We then moved on to discuss the political impact on U.S.-China relations. Rory pointed out that the CHIPS Act is one of the many bills that formed the Biden Administration’s China strategy. And we also discussed China’s responses to the CHIPS Act, its industrial policy, and the recent corruption stand though at the National Integrated Circuit Industry Investment Fund, a mouthful, also known as the Big Fund.

Chip powerhouse, Taiwan, is also featured in the discussion with Rory commenting from a political standpoint and Paul about the market dynamics. Rory mentioned the increasing tension between the U.S. and China on Taiwan, given Nancy Pelosi’s visit to the island, and also the Taiwan Policy Act. Paul commented more specifically on TSMC’s importance in developing on the global supply chain. And he pointed out Taiwan’s dilemma of choosing between a blue line led by the West and a red line led by China.

Questions from the audience in this live event further probed into China’s capabilities to innovate with limited access to cutting-edge chip technology and personnel. And Paul gave a realistic picture on how China is facing a stiff upward battle in the semiconductor industry.

Thanks so much for listening, and enjoy the show.

Chris: Hello everyone. Thank you so much for joining us today. I’m Chris Marquis, a professor at the Cambridge Judge Business School. And I want to welcome you to this live The China Project CEO webinar recording of the China Corner Office Podcast, which is a show focused on leaders and companies facing the challenges and opportunities of doing business in China. Today’s webinar is in partnership with the U.S.-China Business Council, the USCBC, a nonpartisan non-profit organization that represents over 250 American companies doing business with China. And our topic today is the CHIPS Act, semiconductor policy more generally, and what it means for U.S.-China semiconductor competition. Joining us are two leaders who are very intimately familiar with the topic, including Rory Murphy, who is vice president of governor affairs at the USCBC, and Paul Triolo, who’s senior VP for China and tech policy lead for the Albright Stonebridge Group. Rory and Paul, welcome to China Corner office.

Paul: Thank you.

Rory: Thanks, Chris.

Chris: First, I want to start with Paul. You’re a prolific author on this topic and speaker. I just saw something drop from you today or yesterday. We’ll put some of the links in the chat. Maybe just give us a little bit of background. You’ve written a lot about the semiconductor policy globally and in China. Brief review of the CHIPS Act, what does it focus on, and what factors actually drive those different foci of the Act?

Paul:The CHIPS Act, so it took a long time to get passed, almost two years, and it’s called the CHIPS and Science Act for reasons we’ll talk about. I think it was really designed to do several things, one broad and one more narrow. First, the CHIPS Act is intended to jumpstart the U.S. onshore industry to support advanced semiconductor manufacturing. Now, most developed countries, Germany, Israel, Taiwan, South Korea, Japan, all provide some level of subsidies to support advanced semiconductor manufacturing. So, the U.S., at some level, is joining the game that other countries are already playing. To this end, it provides around $30 billion roughly for advanced manufacturing and around $10 billion for legacy semiconductors. Those are the kinds of chips that remain in short supply during the global chip shortage, things like controllers for switches in your car, for example.

And originally, the bill was probably going to be focused solely on advanced semiconductors, but because of the chip shortage and pressure from the auto industry and the medical device manufacturers, the funding will also go to some of this legacy equipment. In addition, some of that $30 billion that I mentioned will also go to other parts of the semiconductor supply chain, particularly upstream. That means things like silicon wafers on which the chips are etched, and process gasses and chemicals needed for the manufacturing process.

Critically too, and we’ll get to this a little later, there’ll be around 200 million for workforce development as part of longer-term planning around the act. Commerce Secretary, Gina Raimondo, for example, a big proponent, has called for a holistic view of this whole effort looking at all elements of the supply chain and bringing things back to the U.S. So, it’s not just about manufacturing. In other words, it’s about the whole ecosystem. And finally, there’s some R&D money in there which will help catalyze an advanced R&D center in the U.S. Likely in New York, Senator Schumer is a big backer of the bill, and that will focus on things like advanced packaging, which is a key downstream sector of the industry.

And then second, which we will probably want to talk a little more about is the Taiwan question. So, the act, at another level, is really designed to bring some level of advanced manufacturing to the U.S. as a national security priority. Secretary Raimondo, again, has been very clear on this. With 90 or so percent of advanced semis made by one company, U.S. national security officials are increasingly concerned about the potential for there to be issues with Taiwan as the U.S.-China relations deteriorate and as China becomes more active in moving towards some type of unification with Taiwan, which is a huge topic, of course. The Pelosi visit and its aftermath, of course, contributed to the current state, and we’re in a so-called fourth Taiwan crisis. But U.S. officials have been keen to reduce dependence on Taiwan for things like chips used in advanced fighter jets.

Chris: The Taiwan question is certainly a really interesting aspect of this, and we’ll get to that in a bit. Just the mechanics of this would be good to hear a little bit about. So, you mentioned this $30 billion, $10 billion, how does that money get deployed? A company says, “Okay, I want to site a plant in X, Y, Z location,” and then they can get grants from the government? I mean, how does money flawlessly getting some semiconductor manufacturing on shored in the U.S.?

Paul: Quickly, so part of the CHIPS Act has set up what’s called a CHIPS Program office at NIST under the Commerce Department. And starting in likely February of 2023, companies will be submitting bids for funding under the CHIPS Act. And so, likely the first three or four tranches of grants, it’ll include both grants and taxes sentence, will go to large manufacturers like TSMC, which is already building a facility in Arizona; Intel, which is already starting to build major facilities in Ohio; and Samsung, which is already committed to building new facilities in Texas. And then there’ll probably be some grants to companies like Micron, which announced a big deal just this week for a memory facility in New York. So, part of the game is that Secretary Raimondo focuses on is that the goal here is to use that money to catalyze private sector investment here.

So, those monies will help to offset some of the costs of building in the U.S. which are higher, but this program office at Commerce has to adjudicate all of these requests coming in. And so, after those first four major manufacturers, it’ll get tougher because there’s many, many pieces of the supply chain as I mentioned, and they’ll have to decide how to prioritize those grants. The tax incentives are pretty much across the board, so companies don’t need to apply for those. Those will be granted if you’re building a facility in the U.S. or expanding facility, you’ll get those tax breaks. But the real grants and giveaways, if you will, the subsidies, are going to be very, very competitive. The Commerce Department’s gearing up, as we speak, to figure out how to adjudicate those requests.

Chris: I want to actually turn to you, Rory. Your role is in leading government affairs for the USCBC, and it’d be interesting to hear a little bit about the political dynamics behind this. Paul mentioned in his comments, companies from Germany to Israel, to South Korea, to Taiwan, all have some sort of policy and support of this industry, but the U.S. is not, sort of a little late to the game. Also, this act took a long time to actually make it through Congress. And so, can you give us a little bit of background about why that is, that it took so long with some of the areas that maybe fell out of the act, or under debate, and then why the U.S. is so unique in not having this type of legislation previously?

Rory: Paul went into how complicated implementation of this bill is going to be and how messy that could be. Well, it is super messy to get to the bill being signed into law. I think there’s good reason for that. This is one of the largest industrial development programs in U.S. history. It’s a big deal. I can’t emphasize enough how excited the White House is about this being signed into law. For it to go in Congress, it took two years. It was a super long and messy process, but it’s how Congress is designed. It should be messy. Something this big, it should take some time. There was dozen markups on different pieces of this thing. Hundreds, probably thousands of amendments filed throughout this process. I feel like every single member of Congress had some input into this bill. There’s a big House bill, a big Senate bill, they combined it into the CHIPS bill. Thousands of pages were left on the cutting room floor, but it took forever because there’s just a lot of ideas that were required to get to this finished product now. What the finished product is, is really focused on industrial policy.

And Chris, as you mentioned, it’s kind of a change of tune for the United States, and there’s several reasons for that. There’s a real desire among Democrats and Republicans to rebuild the U.S. industrial base. There’s real national security concerns about the U.S. falling behind other countries in the race for some of these key technologies. And I think there was a need to channel some of these tough on China concepts into real action. And so, a combination of all those three got us to now where we are. If you look at kind of big legislation that’s passed over the past 15 years, you have the ACA (Affordable Care Act) that was Democrats only. After it passed, it was wildly unpopular. The big tax foreign bill that passed under President Trump, one party passed it, wildly unpopular after the fact.

The CHIPS Act is bipartisan, it’s big, and it’s quite popular and has a high approval rating. So, my sense is when Congress does something right and it’s popular, they tend to go back to the well. I think there’s a real logic behind the bill that could be applied to other industries. So, in terms of industrial policy, this could just be step one for where the U.S. is going.

Chris: Do you have any projections about where to go next? How about if Republicans were to take both houses of Congress in 2022, would things be different versus this bipartisan spirit that you’re talking about?

Rory: If Republicans take one or both chambers, I think you could see change in tone. I think there’ll be political incentive for, say Republicans to take the House, for House Republicans to try to position themselves as the tougher party on China. But it’s important to know, and we’ve talked about this before, Chris, that the politics of being tough on China are just undeniable. 82% of American voters have an unfavorable view of China, and that’s not going to change anytime soon. So, I don’t think, from a policy perspective, there’s a heck of a lot of difference between if Democrats win, Republicans win, you may see an acceleration on some things, you might see a change of tone, you might see some more provocative measures getting votes that otherwise wouldn’t. But I do think that next Congress, regardless of who’s in control, we’re going to see some continuation of these industrial policy type actions. That includes on outbound investment.

We can go into this more detail later. There’s a big desire, Congress just gave out billions of dollars. They want to make sure that that money is staying in the United States. The CHIPS bill had certain guardrails that prevents you, if you’re taking the money from investing in China, there’s a bipartisan desire among a lot of members to do something even broader. So, even if you’re not taking federal funds, there’s a lot of members, a lot of policy-makers that don’t want American investment going into countries like China into sectors of concern. So, we can get into that in the questions perhaps because there’s a lot there, but that’s kind of where I see things.

Chris: Paul, I want to go back to you. Rory mentioned that there’s across the aisle support for these type of programs, not just in Congress but also President Biden. The executive branch also is active in trying to formulate a variety of different actions and responses as well. I know that you have some stuff that you’ve written on that recently. Can you say a little bit about some of the new proposed rules around chips that may create some big issues for some U.S. companies and what’s involved with them and what impact they might have?

Paul: I think there’s a couple of different buckets of things that have been out there on the table for some time. So, you have this industrial policy on the CHIPS Act side of things, of course, then there’s the flip side of that, which is the export controls and restrictions on technologies going to China, for example, in this space. Now, this is not new, right? I mean, we’ve seen the U.S. weaponize semiconductor industry before in cases like Huawei, but I think we’re moving into a new, a really new, where there’s going to be new restrictions.

The Biden administration is going to sort of plug the gap in areas that started in the Trump era and concern over technologies going to China. So, what we’re probably gonna have tomorrow morning is a fairly major announcement around semiconductor manufacturing equipment restrictions going to China. Now again, this is not totally new. U.S. government, for example, already controls the export of advanced lithography to China through a multilateral agreement, the Wassenaar Arrangement. That’s the equipment you need to make the most advanced semiconductors. But this will be a broader set of restrictions roughly targeting production nodes around 14 nanometers. This is the way the industry thinks about the semiconductor manufacturing in terms of these nodes. At 14 nanometers, it’s pretty advanced, but it’s not the most advanced.

But there’s a concern in the U.S. government that China is making progress and building lots of foundries in China that are using U.S. technology. And there’s concern about, of course, the potential for diversion for military end use, etc. These new controls will give the U.S. government more granular control over what types of equipment go to which types of end users in China, and that will come tomorrow morning. That’s known to most industry watchers because the Commerce Department sent out letters to some of the U.S. manufacturers earlier this year.

Chris: So, 14 nanometer chips, are they used in phones, in cars, in internet of things, devices? What is the application for those types of chips?

Paul: They’re widely used. Your phone, your car data centers are all using chips roughly that level. Your phone is also using more advanced chips at seven and five and three nanometers. But 14 nanometers is pretty good solid technology and is used in a variety of applications. The somewhat new wrinkle here is going to be controls around chips used for artificial intelligence and things like high-performance computing. Last month, actually it was August, U.S. leaders in this space like Nvidia, which makes graphical processing units, so-called GPUs, and AMD, both received letters from the Commerce Department warning that basically there would be some restrictions coming up on certain kinds of these chips at a very advanced stage. There’ll be a new rule that outlines controls on those. People are talking about potentially a total ban on the shipment of certain types of GPUs to China, for example.

Now, those are used very widely for AI workloads in data centers, etc. What’s going on here? What are the concerns? And how the U.S. government is going to respond to this. So, that’s another bucket of issues. New controls around AI and HPC, high-performance computing-related chips. And then finally, there’ll be some addition of Chinese companies to some of the many lists that the U.S. government maintains that have differing levels of implications. For example, the Entity List, which is the main list that the Commerce Department uses to control, require licensing around U.S. technology exports to China. There’ll be some additions to that Entity List. There’ll probably also be some companies hit with the Entity List plus the so-called Foreign Direct Product Rule which was used against Huawei, and basically precluded Huawei from using Taiwan and TSMC and other non-Chinese foundries to manufacture chips. So, Taiwan is gonna get dragged into this. Again, Taiwan companies because it’s really an extraterritorial application of U.S. export control laws.

All those things are going to likely happen within the coming weeks, in addition to the outbound investment review, which Rory mentioned earlier, which we can talk about. So, it represents a pretty big effort that Beijing will view, of course, as trying to contain China’s rise as a technology power. It’s going to be a major thing. And U.S. companies, of course, are going to be affected, whether it’s Nvidia, or AMD, or the tool makers. U.S. companies, for example, dominate those tool making sectors. Companies like applied materials like Lam Research, and all the EDA tools, the software tool used to design ships. So, it’s a big issue for U.S. companies.

Chris: Thank you folks for entering some questions in the Q&A. One of them I think relates is a nice follow up to this one. So, I’ll be back to you in a second, Rory. Apologies. The question here is, if the U.S. keeps controlling exports, wouldn’t it force China to just make those tools and chips themselves and then we just have a more hostile competitor? I’m curious to hear what your response is to that. I mean, aren’t we just basically forcing China to up its game and then be able to do it all themselves and we won’t have any leverage at all?

Paul: At one level, that’s the game we’re entering into here. Now, it’s tricky, because left to its own devices, Chinese companies, for example, without some of these restrictions, would tend to use the best technology out there. If they don’t have its own software to design semiconductors, then they wouldn’t go there. But now, because of these restrictions, the sense is that the U.S. is going to continue to weaponize these supply chains. So yes, China is moving quickly to try to develop alternatives. Now, it’s really, though, not a question of catching up. It’s really adapting to the new reality. Already, for example, Huawei, which was hit with some of these restrictions, is trying to fund its own a semiconductor manufacturing effort in China. And then, Huawei is also adapting by trying to design systems that don’t use those cutting-edge chips but use a sort of system approach and software to try to achieve the same types of capabilities.

So, China has already been adapting, but these latest measures, they’ll push China even further down that road. China already has heavy subsidies, even much bigger than the CHIPS Act, which we can talk about, and also many preferential policies for semiconductor companies. The bottom line, though, is this is not something, an issue that you can simply throw money at. You need really good people, experienced engineers. You need time to master the technology and all the processes required. And there are many, many pieces of that supply chain that I talked about earlier that are really tough to break into, right? You can’t just throw money at it and expect to make breakthroughs because most of those companies that are dominating in their areas, the companies I mentioned, it’s taken them several decades to develop the engineering talent and the technology and the innovation capabilities to stay leaders in those areas.

China faces a stiff upward battle if it’s going to really try to recreate most or all of these supply chains within China without access to what’s going on globally. The other bottom line is this is a global industry with global division of labor, and lots of innovation happening all around the world that you need to be aware of and keep track of.

Chris: A number of years ago, probably seven or eight years ago, I actually lived for about five months on the campus of SMIC and got to know a variety of the leaders there. And what you’re saying really resonates with what they were telling me. I visited a number of their fabs there in Shanghai that they have, and it looks like the most sort of state-of-the-art thing. But when you see, there’s like, “Oh, this machine is 60 million, but actually it would be better to have some other machine that would allow us to do various and more advanced things.” So, the complexity is hard to realize and all these intersecting pieces from human capital, engineering, and machine tools all play together.

Rory, building on this thought about competition between the U.S. and China, Paul was just commenting specifically on topics of semiconductors. What’s your sense about what this Act means more generally for the state of U.S.-China relations? Is this a new more intense industrial competition period we’re moving into? How do you see China responding?

Rory: The Chinese have been critical of the CHIPS Act, criticized it for disrupting international trade and distorting global supply chains. And so, it’s interesting to hear that messaging coming out of Beijing. But look, there’s a lot of things on the list that are impacting the U.S.-China relationship. The thousand different items Paul just ran through, those are significant. There’s 800 bills that were introduced in Congress this year related to China. The United States passed the UFLPA, the Uyghur Forced Labor Prevention Act last year. There’s a Taiwan Policy Act moving through Congress right now. There’s six or seven delegations of members of congress and governors who have been to Taiwan since Speaker Pelosi went in early August. So, there’s a lot of stuff on the list impacting the relationship. This is one of those things, but I don’t think it necessarily moves the needle in terms of how bad things are in the bilateral relationship.

The Biden team has been pretty clear that they’re… We don’t have a lot of details on their China strategy, but how they’re viewing this is a combination of confronting China, cooperating with China, and competing with China. The CHIPS Act clearly fits within the competing bucket. And I mentioned up top, they’re very happy with this bill passing. President Biden made a pledge when he was running that he was going to focus on investing in the United States first. I think you have the CHIPS Act, you have the IRA (Inflation Reduction Act of 2022), you have a massive infrastructure bill. All of those things, I think, Biden team now feels like from a competition standpoint, they’ve made a lot of the needed investments.

I think one of the questions we’ll have going into next year, and the rest of the President Biden’s first term is, what’s going to happen on the international front? Are we gonna see a more proactive trade agenda? Because they’ve checked that box of domestic investment. We have IPEF, we have a bunch of different initiatives. And so, those items also are going to really impact or have more of an influence, I’d say, than the CHIPS Act on the bilateral relationship.

Chris: Maybe think about the international nature of this supply chain and knowledge base which Paul had touched on. How is the administration, Rory, approaching this as far as multilateral agreements and trying to contain China’s access to these various machines and chips?

Rory: One of the main things we’re always raising with the administration is on big efforts like this. They’ll be so much more effective if you work in a multilateral fashion. You’re not really achieving your national security objectives if you’re removing the United States from the equation, and they’re just being backfilled by European, Japanese, or other competitors. The semiconductor supply chain is a little more complex, so you can’t just easily backfill. But Paul, you may have heard the same rumor that there’s potentially another country announcing something similar shortly after the United States. The United States has trying to rebuild some of the alliances that suffered recently, and with the Europeans and the TTC (Trade and Technology Council), with IPEF (Indo-Pacific Economic Framework), with the Quad (The Quadrilateral Security Dialogue), with the CHIP 4 alliance, there’s all these different pieces that are coming together. I don’t know the extent that tomorrow’s export control announcement has been coordinated with our allies. I suspect that they’re pretty well informed of what’s coming. It’s inevitable that this thing is going to have major impacts on the global supply chain.

Paul: Can I just jump in quickly there? I agree with everything Rory said. I think that the challenge here is there’s two pieces to this. There’s sort of the industrial policy piece, which requires some coordination. Already, of course, because of the CHIPS Act, we’re talking about companies in South Korea and Taiwan and Japan participating in that. So, already there’s a multilateral or plural lateral element to the industrial policy piece.

Export control piece is a little more complicated because most of the countries we’re talking about don’t have exactly the same type of export control system that the U.S. has. That’s why things like this Wassenaar Arrangement are important because that’s a multilateral organization that is involved in dual-use technologies and controls on those.

The problem is that was set up for WMD (weapon of mass destruction), basically for nuclear and biological weapons programs, and here we’re talking about semiconductors, right? It was a little bit far field. And so, there have been calls, for example, for a new export control regime that would look at some of these more dual use technology that can potentially be used for military and uses like for modeling weapons. So, that’s part of some of the discussions between the allies. But in the meantime, companies like Japan, in particular the Netherlands, which are critical, of course, because ASML, are definitely in discussions with the U.S. government about types of controls, particularly on semiconductor manufacturing equipment because their companies are major players in this. I think there’s a little bit of pushback.

It’s not necessarily a wild agreement on all this because, for example, the U.S. would like to control the previous generation of lithography equipment, which is widely used globally and has not been controlled before as some of the more advanced systems. ASML, for example, sells a lot of their equipment to China, and they are not eager to lose the market. So, the challenge here we haven’t really focused on is these controls are designed, in many cases, to be narrow, but they do impact U.S. companies in a big way, and U.S. companies would prefer these to be narrow and also, of course, coordinated with allies. That’s been the challenge is to find that balance between protecting national security versus really disrupting the business models of those companies and undercutting their R&D budgets, for example.

Chris: I have a question here from the audience also, that relates I think a little bit to these whether China’s getting these different advanced lithography machine tools. So, SMIC had this announcement that has produced seven nanometer chips at scale, which your sense of whether they can produce them at scale, and if so, how does that play into the U.S.-China rivalry in this space?

Paul: Basically, that story from SMIC is a little bit misleading because what SMIC has done is simply take its existing lithography tools, the so-called deep ultraviolet lithography or DUV, and they’ve extended that down to some of the smaller nodes, but it’s really not true seven nanometers in the sense of some of the other companies like TSMC, which have also actually done that with their equipment but are now well beyond that. And to get below seven and really full seven nanometers, they’re using the most advanced lithography equipment which SMIC cannot obtain. And it’s really hard to do that. So, I’m a little skeptical, for example, of the yields that SMIC is getting. I think the only company they announced that they were doing this for was a small bitcoin mining company. And so, I’m really skeptical because it’s really, really tough to get commercial yields.

TSMC was able to do it with their existing equipment, but then they pretty quickly moved to EUV (extreme ultraviolet lithography). The problem is that that announcement though looks bad, right? Because the U.S. does not have currently, domestically, companies that are manufacturing at that level. And so, part of the reaction, I think, that Biden administration felt like it had to react with new controls on these tools is in part in reaction to that. Yes, SMIC is capable of doing that, but I think we don’t want to overstate that ability because it doesn’t mean, for example, that SMIC can make 250 million high-end chips for Huawei. It’s probably a limited capability, and over time, they could improve that, but I think we’d be careful not to overstate what that means. It does play into this broader equation and to the way the U.S. looks at China and export controls on semiconductor manufacturing equipment, though, certainly.

Chris: When you were talking about the process that companies are probably going to have to go through to actually get this money, it made me think about all these scandals recently, to which you mentioned in the China Chips Industrial Policy, that’s know as the Big Fund that has had all kinds of corruption issues and then all these other semiconductor companies that I think have gotten money but actually either have been partial shams or not produced anything. Could you say a little bit about China’s industrial policy on chips vis-à-vis some of those cases, but then also, are they playing in other sections of the industry or trying to do the same thing the U.S. is? Can you just give a little flavor on the China side of this?

Paul: China has long subsidized the semiconductor industry. It’s not new, but the fund, for example, that was set up in 2014 in the National IC fund (National Integrated Circuit Industry Investment Fund) has had very mixed success. So, they’ve been throwing a lot of money at this sector, tens of billions of dollars. And in some cases, it’s produced very competitive companies like YMTC Yangtze Memory Technology Corporation, for example, is producing memory, 3D Nand memory that Apple has been trialing for some of its iPhones and iPads. So, that’s an example of a success. But the fund focused highly on manufacturing and didn’t focus on some of these other areas, for example, like semiconductor manufacturing equipment.

The recent corruption scandal, which is really somewhat unprecedented, which has targeted all the senior leaders of this fund, and including some lower-level officials who were deciding on these investments is interesting. I think it reflects concern in Beijing that despite eight years at it with this fund, it hasn’t produced sufficient results, particularly in the face of increasing U.S. controls on this industry. There’s a sense in Beijing, “Hey we gotta go faster here, we gotta make more progress here.” And so, I think that accounts for this. The problem is this has thrown the industry in China into a little bit of a tizzy. How do they react to this? What’s the model going forward? It’s still not clear because this was a model that was fairly well established. But if you’re going to arrest all the leaders there, including some of the more junior people who were arguably making investment decisions based on what the leaders were saying, that throws a scare into the industry. Who’s going to take those jobs again if they risk getting arrested? So, I think the industry in China’s a little bit of turmoil now, and there’s been restructuring, some of the big national champions like Tsinghua Unigroup have been restructured, one of its senior leaders was also part of these arrests in the corruption scandal.

So, it’s tricky. The problem is in China, the market is very driven by market forces here. And so, government intervention and government guidance is not really very effective except in some limited areas because this is a global and market-driven industry. And so, the problem China has is doubling down on more government investment and control and guidance may not be that effective in solving some of those underlying problems that I mentioned earlier around personnel, and really mid-level management, really knowing how to manage and commercially operate a manufacturing facility, for example. It’s not easy and it requires very smart people. I think China is really going through some turmoil there to figure out what’s the next model for developing its semiconductor industry in the face of this increased U.S. control over the technology inputs.

Chris: The next topic, which we’ve all already touched on that I’d like to discuss a little bit is Taiwan’s role in this equation, so to speak. Paul, you start off by saying that, given that 90% of these advanced chips are being produced by TSMC, that is a big risk for the U.S., as an example. I’m curious about both of your thoughts on how U.S.-Taiwan relations play into this TSMC, and then obviously China-Taiwan relations as well. I think, better if we start with you, Rory, if you have any general thoughts, and then we’ll go to Paul.

Rory: Paul, I’ll defer to you by the market dynamics here and how important Taiwan is to the supply chain because that is a critical piece to it. I think, from a bigger picture perspective, obviously Taiwan has been front and center. Speaker Pelosi’s strip in August really got a reaction. There have been a number of other congressional delegations, a few governors that have gone over. As I mentioned before, there’s a Taiwan Policy Act that’s already passed a Senate Committee. A new version has been introduced in the House. We’re waiting to see how that plays out. From a Taiwan-specific angle, there’s a lot going on. And I think it’s important to remember we have political processes here in the United States, in China that are coming up in the near term. So, tensions are very high. I think, when you start looking at the more narrow picture, I think Paul can comment on with the supply chains, the importance of Taiwan in those, it gets more complicated.

Paul: First of all, I think it’s important to understand how we got here, right? Quickly, we have, again, as we’ve noted, TSMC dominating advanced chip manufacturing level. Also, in terms of overall foundry capacity, foundry just really means TSMC is manufacturing designs from other companies. Taiwan and TSMC have around over 50% of that foundry capacity. So, over the last 10 years, it snuck up on us a little bit. Suddenly, Taiwan is this really, really important and pivotal part of the global supply chain.

Now, the complicating piece is that these restrictions that the U.S. government is putting in place are forcing TSMC in Taiwan to choose whether to become part of a blue supply chain, the U.S. and allies, or a red supply chain, which is China. Because before the restrictions were put in place, China was one of the biggest and growing markets for TSMC services, for example. Many Chinese companies, and still now, even today, Huawei was a big player, but many Chinese companies use TSMC to manufacture their semiconductors. And so, the U.S. was driving a wedge, if you will, between TSMC and China. I think this is what I’ve called an unknown red line because if you’re sitting in Beijing, for example, you have a leading company in a province you consider part of China that is building a facility in the U.S. to put its most advanced technology into. It has facilities in China, for example, in Nanjing and Shanghai, which are limited in terms of the technology that can be put there. And then you have the U.S. also cutting off leading Chinese companies like Huawei from manufacturing semiconductors in again, a province you consider part of China. From Beijing’s optic, Taiwan and semiconductors now are absolutely, I would argue, part of their calculus in terms of unification with Taiwan.

Now, if you step back, the bigger geopolitical issue here is, okay, what happens if there’s even a peaceful reunification of Taiwan? What happens to those facilities in Taiwan? Those facilities are connected globally on a real-time basis in terms of those inputs on the upstream side and shipping out finished semiconductors for packaging, etc. Mark Liu, one of the CEO of TSMC said, basically, TSMC would be inoperable if there was any kind of a cutoff of Taiwan from the outside world, whether it was an air blockade or sea blockade. Most of the stuff goes by air. So, suddenly, TSMC and Taiwan are in the midst of this very tough geopolitical climate here.

And you have people now speculating, for example, on that Taiwan should threaten to destroy TSMC facilities if China invades. And this was not being discussed 10 years ago or five years ago, even two years ago. Nobody was discussing those kinds of contingencies. I’ll just give you one final anecdote. I just heard that there was an academic who was visiting China, for example, and part of the speech that this person gave included questions like, “Hey, who are the top 100 people at TMSC we should save in the event of a Chinese invasion?” And the Taiwan audience was appalled at this, right? The idea that we’re thinking about those kinds of things, and then the idea that U.S. officials are telling Taiwan, “You should think about contingency plans for destroying those fabs in the event of a Chinese invasion.”

So, it’s really amazing to me how quickly we’ve come from, in the last two years, to discussions around Taiwan’s position in the industry and then in this complicated geopolitical tangle between the U.S., China, and Taiwan. I’ve argued, for example, that there needs to be a rethink of that whole relationship, because the documents, for example, on which the relationship is built are really out of date. They come from the 1970s and ’80s, the Taiwan Relations Act, the Three Communiques, and the Six Assurances. Those were done in the more simple times when China didn’t have a military, when Taiwan wasn’t the linchpin of the global semiconductor supply chain, and when the democracy movement in Taiwan was not as strong.

Now I would argue for all those reasons, and particularly because of the semiconductor industry, you need to rethink how this is being approached. But of course, finally, U.S.-China relations are so bad, and there’s no trust, and nobody wants to talk about these things. I’ve called for a fourth communique, for example, that really sets the relationship, and potentially things like demilitarizing Taiwan, because I think the problem is the stakes here are suddenly much higher than, for example, in Russia-Ukraine, which some people compare the situations. It’s really a tricky situation going forward. And I think the semiconductor piece of it is underappreciated.

Chris: I have a simple question, follow-up, a factual thing. TSMC is going to build a plant in the U.S. Of this 90%, is that 1%, 5%, 10%? How possible would it be to actually be building TSMC factories in the U.S.?

Paul: Right now, if you look at TSMC facility in Arizona, they’re talking about roughly 20,000 wafer starts per month by 2024. Now, that’ll probably go up. They have the capacity there to build what’s called a gigafactory, which would be roughly four or five times that number of wafer starts per month. But right now, in Taiwan, TSMC has about a million wafer starts per month. So, the scale of the facility being built in the U.S. is very small. And the Taiwan industry leaders and political leaders continue to say that Taiwan will be the epicenter of advanced manufacturing for TSMC for the foreseeable future. While, but CHIPS Act will fund, not just TSMC, of course, but Samsung will build an advanced fab likely. Intel will build the fabs in Ohio, which will eventually catch up to the industry leaders again at the most cutting edge.

So, over time, this will definitely put a dent. Maybe we’ll get from 92%, 85% or 80%. And it’s hard to say how the industry develops, and there’s a lot of factors there, but it’s not as significant in that dominance. But it is significant in the sense that it gives the U.S. government and national security production-related priorities a leg up over the next decade. But this is really going to take a decade to make a significant impact on those numbers because TSMC is just so dominant in Taiwan, and they’re so good.

Chris: Another topic that we’ve brought up but haven’t really gotten into detail is export controls. This is a bigger issue, Rory, that has been floated beyond just semiconductors. Can you provide some little context for us around the export control idea and what people are talking about in DC these days?

Rory: So, there’s a lot of ideas, and Paul kind of alluded to this up top, that we have all of these programs to prevent investment, technology, and ideas from going and impacting our national security. There may be gap in our current statutory authority. And so, there’s a lot of ideas out there about how to plug those gaps. Next Congress, assuming Republicans take control of the house, Mike McCall, who is in line to become the chairman of the House Foreign Affairs Committee, has been quite clear that one of his priorities, next Congress would be to pass an export control bill. He’s already announced that he would, right away, do a 90-day review and oversight of BIS of commerce to see how the export control regime is going, and then that he wants to legislate on this.

So, Paul probably has a better idea of where those gaps are, but this is a tool that has been in the back burner for a while in Washington and is now very much in the front burner. Paul, you have all this institutional knowledge that you’ve built up. I worry that things are going to change so much. In the next few years, they’re going to have to do a lot of learning, but there’s clearly a focus. There’s going to be a lot more action in this space.

Paul: Yes, the export control regime was the sleepy area because it was really focused on these military end uses and nuclear and biological weapons, etc. And then, with the naming of ZTE to the Entity List in 2016, and then Huawei in 2019, that became a game changer in terms of how people looked at export controls. And so, since then, of course, there has been this attempt to rework the export control regime, adding things like this extraterritorial provision called the Foreign Direct Product Rule, and retooling this tool for this new game we’re in. Because the export control system, as my good friend, Kevin Wolf, the former BIS official says, “Was not designed for a great power competition or for maintaining U.S. tech dominance or for human rights issues, these other things which is being used for.” And so, that’s why he and others are calling for that fifth control regime, if you will, which would be, in addition to these other regimes designed to target specific kinds of military and uses, he’s called for discussions on that. That will probably happen in the next six months to nine months in places like the EU, U.S. Trade and Technology Council, which is a forum that’s taken up some of these issues, and in other quarters.

The challenge again is that each country has a different export control approach. I think other countries are moving in a direction that’s closer to the U.S., Japan has set up a new national economic security arm and doing a lot in this space. And so, yes, I think over time, we’re going to have a new era, if you will, in export controls. But in the meantime, we’re sort of stuck with this arcane and unwieldy system, which industry, of course, has struggled to understand at times.

And here, it’s going to get even more complicated because, just a quick example, in trying to understand what tools are used in manufacturing, there’s lots of them, many, many, many tools. So, which tools are going to be controlled? For example, who’s getting those tools in China? Commerce Department itself is undermanned. The Bureau of Industry and Security is only a handful of people. So, they’ve been struggling to figure out how to implement and force some of these issues and then do things like licensing, which is always a really tricky thing because being put on the Entity List doesn’t mean you can’t get technology. It means you have to get a license to export that equipment.

So, we’re often in this learning curve of understanding the rules, and this thing that’s going to come out tomorrow, it may be a hundred pages of new controls. So, Rory and I will be pouring over that tomorrow to understand what are the implications of that for U.S. companies, for the industry, and then for China too, because ultimately, China’s going to be the one most impacted by that.

But again, we’re using these tools against China across the board. We’re either using old tools, adapting them, or we’re developing new tools like the outbound investment review to try to plug gaps. My view is this, the real goal is to give the U.S. government this granular control and visibility on all these technologies and who’s selling them to China, who’s investing in them, just so that the U.S. can dial back on those dials to try to control rate of development of China, these critical technologies.

Chris: Really interesting question from David Tyfield. So, we’re limiting all these technologies with the assumption that they are essential for innovation. You mentioned, in one of your earlier comments, that if China doesn’t have certain chips, there could actually be workarounds or some other technical process so that innovation can happen without those chips. And David and his question mentions a variety of other examples of where that’s happened. So, what’s your sense about that process, and will there just be a variety of other ways to do the same similar advancements?

Paul: I think it depends on what you’re talking about too. There’s lots of different applications here. Just for an example, these restrictions on things like GPUs that are used for AI workloads, that’s a huge thing because there aren’t a lot of alternatives to those. If you’re, for example, a Chinese AI company, an autonomous vehicle company, or facial recognition, and you’re training your algorithms in the cloud, for some applications, it’s really difficult to find equivalent alternatives. For others, it may be okay, maybe good enough. You can find good enough solutions to use, chips that are made at lower levels of technology, but again, you architect the system around that. You can overcome those things. I think we’re in a period now where China is really going to facing this need to adapt to these new controls and figure out ways to do it. But they’re going to do it. There are some smart people in China, right? There’s lots of engineers. They’re going to continue to innovate. They’re going to continue to use technology.

The other flip side of that, though, is also the concern of the U.S. industry, for example, is that all this cutting off of access to the China market is going to ultimately also impact U.S. ability to do innovation. Because those companies like Qualcomm and Nvidia, Broadcom, Intel, they’re deriving a huge amount of their revenue from access to the China market. And to the extent that they’re cut off, that’s going to impact their R&D budgets and impact their ability to innovate. So, the question I always have is, what’s the national security gain? Sometimes it’s hard to measure versus the cost to your industry and to your innovation capacity.

Chris: One of the things I’ve read about this semiconductor industry discussed how when Nvidia had gotten this letter about the changes that they estimated like $400 million hit on a quarterly basis. I’m sure Invidia’s talking to their congresspeople and those people are lobbying. And first, Rory, I’d like to just hear, in general, in the policy process, how those type of interests that Paul was just talking about, weighing the money and innovation and employment versus some of these issues, and then we’ll finish with Paul on this issue.

Rory: It’s certainly something that is being discussed is under consideration is well aware. I think that with the export control rules coming out, with the ongoing outbound investment process, one thing that probably differentiates the Biden administration from the past administration is there is a lot of interaction with the business community, a lot of taking of comments informally, formally. People may disagree with policies, and I think there’s good reason for that in some cases, but there’s not as many surprises, I would say, as we saw in the previous administration. I think that’s important.

Paul: The Biden administration is attempting to put these strategies on a more firm legal footing and trying to work with industry to help them understand the goals here. Last week, for example, I think it’s really important, we didn’t mention, Jake Sullivan, the National Security Advisor, really came with a really strong public statement that really was the first time a senior U.S. official that I can remember, articulated a strategy around semiconductors in China. He basically said, “Look, for a long time, the unstated goal has been to keep China one or two generations behind the U.S. leadership.” And he said, “Now we’re no longer gonna be pursuing this sliding scale. We want an absolute advantage in this sector.” And so, that’s a huge signal to industry, which is basically saying, Hey, the U.S. is going to draw much tighter lines around some of these technologies and really seek to maintain the U.S. advantage in ways that the industry better get comfortable with.

And so, difference there, again, is unlike the Trump era, industry will be listened to, but only up to a point. So, if the goal is really to keep this advantage, then there will be policies and measures coming down that are going to help to enforce that. So, I think that’s why we’re in a new era, both for China and Chinese companies and the U.S. companies, which are going to have to get used to this idea that government is going to be dictating, both on an industrial policy level, more intervention and good ways, and then also on the restriction side.

Chris: We are actually at the top of the hour, so out of time. I just want to say thanks so much, Paul, Rory. Lots of really important insights and facts about this semiconductor competition, which is only going to continue. So, thank you both very much.

Rory: Thanks, Chris.

Paul: Great job.