Glenn Tiffert discusses China's challenges in developing its semiconductor industry despite massive government subsidies. Tiffert explains that factors like lack of talent, economic inefficiencies, corruption, and reliance on foreign firms have hampered China's progress, but US export controls could unintentionally help China become self-sufficient over the long term. He explains the Silicon Triangle report's recommendation of "friend-shoring" semiconductor production to trusted allies rather than trying to onshore it all to the US.
>> Kharis Templeman: The problem has been that China's chip industry has tended to be state led and state owned enterprises, and they've burned through phenomenal amounts of capital, with some modest goals met, but really nothing on the scale of what China had hoped they would have achieved by this point.
Well, good afternoon, good evening, and good morning, wherever you might be. I'm Kharis Templeman. I'm a research fellow here at the Hoover Institution. And joining me today is my friend and colleague here, Glenn Tiffert. Glenn is a distinguished research fellow here at the Hoover Institution and a historian of modern China.
He co-chairs Hoover's project on China's global sharp power, and he directs its research portfolio. And he also works closely with government and civil society partners around the world to document and build resilience against authoritarian interference with democratic institutions. Glenn has worked extensively on the security and integrity of ecosystems of knowledge, particularly academic, corporate, and government research, science and technology policy, and malign foreign influence.
He has authored or contributed to numerous hoover institution publications, among them eyes wide open ethical risks in research collaboration with China. And the reason he's here today, speaking with me, is that he is the author of one of the chapters in our newly released report, the Silicon Triangle, the United States, Taiwan, and China, and global semiconductor security.
And we're here to talk about his chapter, China's lagging technonationalism. Thanks for joining me today, Glenn.
>> Glenn Tiffert: Thank you, Kharis, it's a pleasure to be here.
>> Kharis Templeman: So I thought I'd start our conversation off today with a kind of broader philosophical or strategic question about trade-offs in the approach to economics by national governments.
So there's a fundamental trade off in us policy right now between economic efficiency and national security concerns. And the United States has taken a more proactive approach in recent years towards the use of industrial policy and export controls to promote national security interests. And there's a debate, a lively debate in the United States over whether that's the right thing to do, whether in particular, the United States should be subsidizing chip production, although with the passage of the Chips act last year, there's one point in favor of the national security side on the question.
But since you're the author of this chapter on China and its own challenges in promoting economic independence in the semiconductor supply chain, I wonder if I could draw you out a little bit on what is not a debate in China. It has been standard practice, really, for decades that there's no clear line between private business interests and the interests of the state and between economic and national security imperatives.
And so I think would be useful for our audience watching this to hear about the PRCs approach. So what tools has China historically employed to improve its semiconductor supply chain competitiveness and its own attempt to build self-sufficiency in this area?
>> Glenn Tiffert: Thank you, Kharis for asking me that question.
As many will know, China's model of economic development was derived originally from the Soviet Union's planned economy. And so China has decided very early on to develop positions in key industries and industries that were important for national security and economic development by opting, through the state planning mechanism to invest huge resources to create whole industries from the ground up.
And this was the model that China had approached, really, since the 1950s. China developed its first semiconductor and identified semiconductor technology as a priority in 1956. It developed its first semiconductor and first integrated circuit in 1965. But owing to the cultural revolution and political instability and just general poverty, it fell far behind its East Asian neighbors in developing a semiconductor industry.
But when the 1980s and 1990s came around, the chinese government identified this as an area in which it wanted to catch up, and it mobilized tremendous state resources, plowing them into educational institutions, sending students abroad to get degrees at american universities, which were leading at that time to develop industry experience and repatriate those people to jumpstart the semiconductor industry in China.
And so tens of billions of dollars were spent bringing equipment, whole factories, teams, consultants on to begin those enterprises. Now, you point out rightly so, that debate about this is not really very active in the PRC today. But it was actually not always the case. In the last 20 years, there was a vigorous discussion about the extent to which state involvement in the economy was a help or a hindrance to China's economic development.
Many will know that China's economic growth and the most productive firms in the country are the ones that have been most private in orientation. Though there is sort of a lively discussion about to what extent is any firm private in the PRC, but those that were most market oriented and most responsive to the market and to capital were generally the highest performers.
The problem has been that China's chip industry has tended to be state led and state owned enterprises, and they've burned through phenomenal amounts of capital, with some modest goals met, but really nothing on the scale of what China had hoped they would have achieved by this point.
>> Kharis Templeman: I actually have some statistics here from your chapter.
The consumer market for chips in China is the largest in the world, in part because China is an electronics manufacturing powerhouse, and a lot of that gets re exported. But in 2021 had on the order almost $200 billion us in purchases of chips. That's about 35% of the world's total consumption of chips.
But as you noted, the industry in China is lagging far behind in terms of production, and so only about six to 7% by value of all of those chips were produced in China. So China's a major net importer of chips produced in the rest of the world, and indigenous chip manufacturing firms are still far, far behind the industry leaders.
What may be even more surprising to our audience is that that's in spite of a decade of the state just pouring massive subsidies into this industry. So last year, or 2021 at least, subsidies were something like 20% to 40% of total revenues for most of China's semiconductor manufacturing firms.
That's in comparison to the industry leaders like Intel, TSMC, or Samsung, which had roughly 3% of revenues that came from state subsidies. Total government assistance for PRC chip firms was over 100% of global sales in 2021. That's in comparison to, say, 10% in Japan, three to 4% in Taiwan, and less than 1% in the US or South Korea.
And so the United States is far, far less of a subsidizer of the chip industry than most of the major competitors of us firms in this area. I wonder if you could unpack for us a little bit why chinese firms, despite all of these state subsidies, these just massive advantages that the PRC is providing for their firms, why are they struggling to compete?
Why are they having such a problem catching up to TSMC and Samsung and some of the other industry leaders? Why hasn't China's ambitious plan to develop a domestic semiconductor industry? Why hasn't it played out the way they expected it?
>> Glenn Tiffert: Kharis, I'm glad you asked that question. One of the great ironies of China's heavy subsidization of the semiconductor industry is that, although the goal was to onshore as much production and to bring it in the hands of Chinese state owned enterprises.
Actually, provincial authorities and local authorities who were delegated all of that money, made a market-driven decision to go with the global firms very often that had the best technology and had the proven production methods. And so firms like Intel, SK Hynix, Samsung, and others, onshored production from their own countries in many cases, the United States, Korea, Japan, Taiwan.
And Chinese subsidies paid to help build factories in China. This raised China's production of semiconductors and integrated circuits significantly, but they were not being made by indigenous firms. And so not as much of the talent or technology was being transferred as the Chinese government would have liked. And so you cited some very accurate statistics on China's output of semiconductors.
But the truth is that over half of that output is being done by the local facilities owned by foreign firms who are taking advantage of those subsidies in China. The local indigenous firms have been much slower to catch up. And that's for a variety of reasons, really, there are maybe four major reasons.
One is workforce. Most semiconductor producing nations are encountering workforce issues. The United States is among them. China is having a very hard time developing the indigenous talent, the engineers, but also the managers and the high level PhD level scientists who can sort of set the standard on a production line which has really just minuscule tolerances to ensure quality and yield.
There's a lot of alchemy that goes into running a good production line, and the Taiwanese and Koreans have mastered this. A lot of that resides in the heads of people, just like a master chef is not simply cooking from a recipe, but has internalized a lot of recipes and knows how to make a fabulous meal simply by experience and invention.
China's been slow to develop this, and it's been slow to endow local people with the authority to take some risks and learn from the process. So it's been very reliant on foreign talent up to this point to keep those production lines going. So that's the workforce issue. They simply not turned out the people that they need, and they're reliant on foreigners even to this day.
And the US export controls are here hitting them hard as a result, because they hit US persons, that is, citizens and green card holders, and prevent them from contributing to Chinese semiconductor firms and lines that are active in particular corners of the market. The second really comes down to economics.
Up until the most recent rounds of export controls, most chinese firms had full access to the global market. And local investors said, well, we can reinvent the wheel here, but why do we have to, when we can simply go to Taiwan and the United States and get the best inputs that the entire world has to offer, rather than spend money to try to recreate that here in China.
That actually hindered China's domestic indigenous development of those inputs into the semiconductor industry, which is profoundly globalized. And so China, they would outsource their designs to be manufactured by TSMC. They would use software developed in the United States, they would use chemicals developed in Japan, like everyone else in this globalized supply chain, and that inhibited the development of local inputs.
And then the third is just vast, epic scale corruption, with hundreds of billions of dollars in either absolute subsidies or indirect subsidies, like concessional loans, free land, subsidized electricity rates. There was a tremendous opportunity for people at every level to skim money off the top. And so they were promised facilities and fabs that tens of billions of dollars were directed towards that were never built.
And the money kinda disappeared into the ether. And there have been a variety of lurid scandals that have been documented even in the Chinese press about this. And so that method was not working. And then finally, there is just the level of clientelism in the chinese political system.
Because it's a political system in which state power directs capital rather than market driven forces. People with political power were able to divert capital to their own factions and their own patronage networks, which inhibited the kind of efficient allocation of capital that you need for an industry that evolves at just a breathtaking pace.
The United States is actually pursuing our own mini version of this effort to build up a domestic chip manufacturing capability for national security reasons.
>> Kharis Templeman: Are there any cautionary tales in your mind from what the Chinese have attempted to do and failed at anything that the United States might do?
We have advantages that the Chinese don't, for instance, in trying to subsidize a chip industry that, to be frank, doesn't have a lot of the advantages that Taiwan or South Korea does right now.
>> Glenn Tiffert: Once upon a time, we had all of the integrated components and inputs that went into semiconductor manufacturing in the United States.
But when we lost manufacturing, we lost all of those components, and Korea and Taiwan reassembled them. And they did that largely through industrial policy and mercantil approaches. There's a debate going on in the United States, and it follows very predictable ideological and partisan lines about the extent to which state support should play a role in onshoring the semiconductor industry back to the United States.
You will have some who say that the United States should not be picking winners and losers in industrial policy for just basic economic reasons. And then there are others who point out, rightly so, that the way that Japan first, then Korea, then Taiwan established themselves in this market was through pure industrial policy and mercantilism, and for the United States to get back in the game and reclaim that lost ground.
There is a strong argument made for a strong dose of industrial policy. But I think everyone would agree that market forces still have to be in the driver's seat. We just have to be prepared to take some chances, spend a little money, because at the moment, it is actually about 30% to 50% more expensive to make any given chip in the United States of the higher end of technology than it is in Taiwan.
So if we relied purely on market forces, that production would never come back to the United States. Now, once we redevelop the local resources and economies of scale, the workforce, and it's everything from welders who are building the FABS and installing the machinery to those who are operating the machine.
Machinery on up and the accountants and managers. Once we reestablish that, those costs will come down. But we're starting from a very weak position, unfortunately.
>> Kharis Templeman: So the other tool in the US toolkit, the big tool, is export controls. And since about 2018, the us government, under both the previous and current administrations.
Has been increasingly aggressive about using export controls and other forms of sanctions or punishments to try to hinder the expansion and upgrading of the PRC semiconductor industry. Up to and including restrictions on recruitment of US citizens and green card holders into the PRC semiconductor industry. In the short run you've argued in the chapter that this will have a pretty devastating effect on the industry in China.
I wonder if you worry in the medium to long term whether this might unintentionally actually accelerate the upgrading or self sufficiency of the industry, and if not, why not? So another way to put this is there's a tradeoff between trying to hit the industry in the short run and kneecap it.
But maybe you are unintentionally in the medium to long term, helping it transition to a position where its much less vulnerable. We have much less leverage over the Chinese economy than we do now.
>> Glenn Tiffert: I would start by pointing out that the United States is not unique actually in imposing those restrictions on the flow of labor and talent into China.
Taiwan in particular has adopted similar restrictions in 2019, owing to the fact that Chinese subsidies. Were aggressively paying for the recruitment of Taiwanese engineers straight from Taiwanese chip facilities to the PRC and offering them salaries double or triple their local Taiwanese salaries. Up to 10% of the Taiwanese semiconductor engineering workforce had relocated to China, and this became a threat to the entire industry in Taiwan.
So in the last year or two, they've adopted some restrictions on the flow of individuals to China for exactly that reason, in the way prefiguring the US restrictions. And now Korea is considered considering the same thing as a result, because they've been stung by some local cases in which some very high level Korean executives were lured.
Again by very fat paychecks in the PRC to take their talent and skills abroad. And so this is not a unique proposition by any means. Now, the challenge there is, I think, a vigorous debate and an understandable debate about. Whether those export controls are going to force China to accelerate its efforts to onshore and develop self-sufficiency.
But as I've said already, China's actually been in doing just that, and has committed to doing just that for the last, well, 25 to 30 years. And so this would not be anything new for them. It would simply be perhaps an intensification of their ongoing efforts, which they've doubled down on several times already with diminishing results.
So I think very likely what's going to happen in the medium term is assuming that our allies and partners, not just Taiwan, Japan and Korea. But also the Europeans, who have key technology inputs in the semiconductor industry, cooperate the United States. It's very unlikely that China will reach its goals in the advanced nodes of semiconductor and integrated circuit manufacturing.
So that leaves the question about, what about the more mature nodes? And this is where the lion's share of the market is. This is the semiconductors that are in all of the commodity products from your kitchen to your car, to the things you wear on your wrist or carry in your briefcase or purse.
These commodity chips minus, of course, the logic chips that drive your iPhones, those are the super high end ones. But the commodity chips, these are the ones that the shortage of shut down Detroit, for example, during COVID when very, very cheap parts were missing. And they had to idle production lines and couldn't ship cars out to consumers.
We worry about China intensifying its building of facilities and fabs in those mature nodes, which it's already demonstrated a capacity to produce in. And if China does what it has in, say, sectors like solar, photovoltaic cells and telecom that is subsidized the production, you get a huge surge in production in China that's under market prices.
It forces foreign firms out of the market because they can't compete. It's unfair competition. And then China is left being the only one standing. And Chinese producers then own the market and establish leverage over the rest of the world. And China has proven that it will use that economic leverage as a political tool to reward and punish countries that it likes.
Or does not like for very simple things like opening an office of the Taiwanese Economic and Cooperation office in your country. And so I think there's a tremendous concern of intensifying or increasing our vulnerability to China at the lower end of the market. Medical devices, refrigerators, toasters, all of those things.
And we need to think very hard about that. Now, sadly, the margin on those products is so low that it's unlikely to be economically viable to increase production in the United States significantly to offset China's efforts. But there are other countries in the world that are eager to climb the value chain in which there are perhaps opportunities for the United States to work with those countries to develop their capacity.
So that we take all of our eggs and we don't put them in one basket and we distribute them.
>> Kharis Templeman: Are there any other possible alternative approaches to that? Could we possibly subsidize a couple companies or plants that would create a strategic reserve of chips? Or other thoughts about how to lessen our dependency on Chinese firms kind of cornering the market on lower end mature technologies going into chips?
>> Glenn Tiffert: It's interesting, some Chinese firms are seeing this as handwriting on the wall and are actually investing abroad in Southeast Asia to get around potential us export restrictions. And Southeast Asia sees this as a win. Malaysia and Singapore, Thailand in particular. There are opportunities for the US to work with those countries that are traditional us partners.
Now this is about more than just subsidies. This is about working with local IP regimes to ensure that technology transfer that occurs as well protected and that companies don't lose control of their technology. They're not going to want to invest or establish facilities in a country unless they can have those assurances.
It's about developing local workforce. There are significant reasons why those countries are not major semiconductor manufacturing nations already and why all of that capacity flowed to China. So I think we need a full spectrum response but there are some countries that are better placed than others. And in particular, I think it's important for the United States to think about moving some of those assets outside of East Asia just because East Asia is potentially a tinderbox.
It isn't just the China Taiwan crisis, but it's also north and South Korea. And any intensification of hostilities or political crisis in that region potentially throws at risk all of the nations that produce semiconductors there, Japan, South Korea, Taiwan and others. So redistributing out of that region, perhaps into this hemisphere, should be part of the strategy.
>> Kharis Templeman: So it sounds like if there's a tagline here, it's really friendshoring rather than onshoring just to the United States. That is the critical part of US semiconductor strategy.
>> Glenn Tiffert: Absolutely no country, not even Taiwan, can do this by itself, right. And so really it's about mustering the coalitions and the friends and alliances that the United States has around the world, doing what markets do best with market-driven economies.
And simply diversifying and reducing our vulnerabilities with respect to China, helping bring some other nations along with us who will be eager to have that technology transfer. And climb up the value chain so that there's a convergence of interest. The United States certainly should try to onshore those parts of the market that are essential for defense production.
We should not be reliant on China to supply parts for our own military hardware. That should be a no brainer but if you actually disaggregate the parts that are going into our military systems. Unfortunately, just because no one's managed the supply chains very well, we do have vulnerabilities in that respect.
And so onshoring some of that, onshoring some of the advanced technologies and then redistributing around the world, the more mature nodes, I think, is the sound strategy.
>> Kharis Templeman: Well, thank you, Glenn, for joining me today. It's been a pleasure speaking with you about our new report. Once again, I'm Kara Templemanh.
I've been speaking with Glenn Tiffert. We're both research fellows here at the Hoover Institution, and we've been talking about a new report out just this month called Silicon Triangle. The United States, Taiwan, China, and global semiconductor security. Thanks for joining me, Glenn.
>> Glenn Tiffert: It's been my pleasure, Karas.
>> Kharis Templeman: This has been a product of the Hoover Institution and Hoover's working group on semiconductors and the security of the United States and Taiwan.
>> Kharis Templeman: I'm Karas Templeman. Silicon Triangle is a special podcast series of matters of policy and politics.
>> Speaker 3: This podcast is a production of the Hoover Institution, where we generate and promote ideas advancing freedom.
For more information about our work, to hear more of our podcasts or view our video content, please visit hoover.org.
