Discover more from Discourse
Hundreds of Billions of Dollars Later, China’s Semiconductor Industry Is Still an Also-Ran
David Masci and Christine McDaniel talk to George Calhoun about how China’s top-down industrial policy is failing to create a world-class chip industry
For the past decade or more, China has been seen as an unstoppable economic colossus that will soon not only become the world’s largest economy but achieve preeminence in many if not most key industries. But according to George Calhoun, this narrative is incorrect. As long as the country is run by a Chinese Communist Party focused first and foremost on maintaining political control, he says, China will never develop the innovative capacity of the U.S. and other Western countries and will inevitably remain an economic also-ran, particularly in the cutting-edge technology sector. Nowhere is this truer than in the all-important semiconductor industry, says Calhoun, who is a professor at Stevens Institute of Technology.
Calhoun comes to these views after a lifetime of experience in and around the technology sector, and not just as an academic. For 30 years, he worked in the wireless technology industry, co-founding or helping to run a number of public companies including Interdigital Communications Corporation, Illinois Superconductor Corporation and GeoTek Communications. At Stevens, where he’s been since 2003, Calhoun heads up the Quantitative Finance Program. He is also a regular contributor to Forbes magazine.
Examining the economic competition between China and the West through the lens of the semiconductor industry makes sense, Calhoun says, since chips are now an integral part of a wide and growing array of products, from toasters to cars to jet engines. Not surprisingly, the industry is viewed by the Chinese, the Americans and others as vital to their security, which explains both China’s efforts to provide hundreds of billions in subsidies to create a homegrown semiconductor industry and the recent bipartisan enactment of the CHIPS Act in the U.S.
Recently, Calhoun sat down with Discourse editor-in-chief David Masci and Mercatus senior fellow Christine McDaniel to discuss the geopolitical competition over semiconductors and its broader implications.
China’s Chip Industry
DAVID MASCI: George, let me start with a very broad question, and then we’ll dig down a bit. Talk to us just a little bit about the current state of China’s chip industry. We hear all kinds of things in the news about how lots of money has been spent, about how they’re still way behind leading countries in the West, particularly Taiwan, South Korea and the United States.
How far behind are they? Have these hundreds of billions of dollars that the Chinese government has poured into its semiconductor industry just been wasted?
GEORGE CALHOUN: China, today, still imports something like 90% of its requirements in the chip industry. It’s vital to them because they then turn around and assemble electronic goods. It’s a crucial element in their supply chain, but I think that 90% number, notwithstanding the hundreds of billions that you referenced, indicates that they are still not where they want to be and not where a lot of people in the United States sometimes fear they are. China has not been successful in coming up the ladder of technology in the semiconductor business. I think their current program is not likely to be successful in that regard.
MASCI: Why is that?
CALHOUN: Because it’s a top-down, Soviet-style mandated, dump money on it, provide subsidies to one and all program. They have tried that more than once. In the latest round, Beijing has announced a new set of incentives for semiconductor companies. What happens is there are suddenly 10,000 or 15,000 semiconductor companies that yesterday were bending metal and maybe doing dry cleaning and goodness knows what, and suddenly they’re relabeling themselves as semiconductor companies. That is not the way the industry works.
You’re starting to see some of the crumbs falling off the table on the other side, too. I just read an article about the downturn that some of those companies are now running into in China. 3,000 of these semiconductor companies have declared bankruptcy in China. Just consider those numbers versus what the semiconductor industry looks like in the United States or in Korea or in Taiwan. There are not 10,000 companies; there are not 3,000 companies failing. It’s a much different structure of the industry. I think that those numbers—the 90% and the 10,000 and the 3,000—those kinds of numbers show you that China is definitely not where it wants to be.
Possibilities of Reverse Engineering
CHRISTINE MCDANIEL: Tell us about China’s reverse engineering. How hard is it to reverse engineer an advanced chip? Could China just buy a super advanced chip, tear it apart and make something just like it?
CALHOUN: Look, I think reverse engineering goes only so far, and it applies to fairly simple kinds of products in most cases. It does not apply to the semiconductor industry, which is intensive in not only its intellectual property, but just the level of know-how and experience accumulated over decades by the industry leaders, who are not standing still either.
I think China has probably tried their best to “reverse engineer” some of the systems that they’ve seen. If you’re talking about a circuit card where there are a number of components that are commercially available, and what you want to do is duplicate that circuit card, in the electronic assembly world it may be more feasible. But in the fabrication of semiconductor systems, semiconductor chips, certainly at the leading edge of the technology—and it is not the kind of thing that can be reverse engineered just by throwing a team of half-trained Chinese engineers at it and letting them try to pick it apart.
MASCI: Just to follow up on Christine’s question, I would assume also that the lack of access to the most advanced chip-making equipment, which has been a problem for them because of American sanctions and limits that have been imposed by the United States, that also of course has played a role in retarding the growth of the Chinese industry somewhat. Am I correct about that?
CALHOUN: Absolutely. There are several choke points in the industry that affect the resources that you need if you’re going to try to develop and fabricate the most advanced chips. There are the photolithography products that, in one case, ASML is the only supplier of the most advanced type of photolithographic equipment that’s used in the most advanced chips.
But there are other parts of the industry ecosystem as well. There’s an area that a lot of people don’t pay that much attention to, electronic design automation, EDA, which describes the critical software tools that are used by chip designers. Qualcomm or Nvidia, they are highly dependent on companies like Cadence and Synopsys to provide a software platform that allows them to design their chips.
Chip design is such a complicated business today that even the companies that do it for a living are themselves dependent on some of these key suppliers. Those key suppliers are all Western. Currently, the EDA industry is basically a U.S. industry, U.S.-based. I think those are also significant obstacles to the Chinese in terms of recreating the capability to be a top player in the semiconductor world and in any short period of time.
Tech Success and Government Policies
MASCI: What if the Chinese crafted a better policy rather than a top-down, Soviet-style policy? Let’s say they created a much more open market for their chip industry; it sounds like it still might be lagging because of the problems in these choke points that you describe. Am I correct in that?
CALHOUN: To create such a better policy, I think you’d have to create a better China, a better Chinese government. It’s not going to happen with the current Chinese regime. There’s no reason that it can’t be done. Taiwan is in effect an example of a country that started 25 or 30 years ago from a position of being not really a player in the industry. They’ve built a tremendous capability. But how did they do it?
They did it with a democratic, technology-centric, pro-technology, pro-entrepreneurship, pro all the things that the Beijing regime today is not in favor of. They leveraged the very strong relationships that they had with the Western chip companies and chip designers over decades. That’s something that the Beijing regime today just doesn’t seem to get. They are, if anything, coming down, cracking down on their tech sector in ways that, to me, suggest that they’re going to move that needle backward, not forward.
MCDANIEL: All of these hundreds of billions of dollars that David has talked about being spent and perhaps more to come, where is that money going?
CALHOUN: It’s going into all the little pockets that those monies flow into in China, the corruption. As I said, you announce a subsidy program for semiconductor companies, you’re going to get a lot of semiconductor companies popping out of the woodwork saying, “I’d love that subsidy. Yes, I used to make automobile tires, but now I’m a semiconductor company.” That process is not going to result in an outcome that will give China much of a payoff. Unfortunately, from China’s perspective, they have spent a lot of money in that fashion rather than opening up their economy more to the tech sector and letting it rip.
For instance, look at what Jack Ma and his group were doing and what they were on the brink of being able to accomplish in digital payments, it seems to me, just a short while ago. Now, we have not heard from Jack Ma since October of last year, almost 11 months. I’ve tracked his sightings ever since he was sequestered. He’s been seen or heard from a half a dozen times, but I don’t believe that he’s been publicly seen in a year.
If there was one company that was really looking like they were going to be able to leap over some of the Western competitors decisively, this was the company. Some of the things they were doing were really quite innovative.
MASCI: You’re referring to Ant Financial?
CALHOUN: Ant Financial and Alipay and the other dimensions of Alibaba. I just have a lot of respect for what they were trying to do. And I think they were on the right track, and now they’ve been basically shut down.
MASCI: Is China’s future at this point essentially as a commodity manufacturer until they make the sorts of paradigm shifts that you’ve just talked about? We do hear that in certain areas like artificial intelligence, they’ve made great strides. They’re supposedly building very powerful supercomputers that rival those built by the United States.
They do seem to have some technology prowess in certain parts of the technology industry. I know your specialty is chips, but if you can broaden out your analysis just a little bit for us, are they advancing in some industries at a fast clip and in others like semiconductors not? Or is this top-down, politically driven economic management going to metastasize and basically ruin the entire tech sector in China?
CALHOUN: I think it’s on its way to doing just that, or at least putting it in the deep freeze until there is a change of policy at the top in the Chinese government. Just as there’s concern about China’s position in the semiconductor world, there’s concern, as you’re pointing out, about do they have some advantage in artificial intelligence or some advantage in supercomputers?
What I can tell you is that the people I talk to in the AI business are dismissive of the Chinese AI initiative. I’m not as deep into that to be able to connect the dots for you. It’s something I’m going to be studying myself. Coming over to the United States side, it serves the agendas of certain groups here politically to hype up the possibility that we might be suddenly about to lose our advantage in artificial intelligence or supercomputing or chips. I think you have to take all of those comments with a great deal of caution.
MCDANIEL: A follow-up to what you were saying, but something that is a big question for a lot of people. In the standards-setting arena, we often hear policymakers talk about the need for the U.S. to be in the leadership position so they can set global standards.
When you hear policymakers defend the CHIPS Act, that’s something that comes up. Could you talk a little bit about that—whether and to what extent that concern is justified? Who does set the standards, not only in the semiconductor chip-making industry but also in these other emerging industries? Where does the U.S. sit relative to China in its ability to help set those standards?
CALHOUN: I would say that the setting of standards is very important in some industries, like the telecommunications industry that I used to work in. That’s an area where, as the industry develops, you would like to have standard interfaces so that the customers, the operators, can mix and match from different vendors and have competitive bids out for the infrastructure that they buy. Of course, the handsets have to have a standard interface to the cell towers. Telecommunications is very much defined by the standard-setting process, and there’s a long history of that.
I don’t really see the chip industry in the same light. I think the chip industry is more dominated by industry leaders that set de facto standards, in a way, and then may allow, through licensing, other players to participate, like Intel with their x86 standard or Arm with their mobile low-power designs that are licensed to almost the entire cellular industry and many other device areas. These are not standards that were defined by a standards body that got together and somehow, with government support, voted on and architected a standard the way the telecom world does. These are de facto industry standards that the leaders in each segment are able to establish and then utilize. Probably, from some perspectives, it gives them a little bit of a monopoly power. But from another point of view, it also means that they have the ability to control the evolution of the technology. I think that’s important for this type of business. It’s not a static business.
I used to say, when I was in the telecom business in the 1980s and 1990s, you could still, at that time, take a television set that had been manufactured in the 1940s and turn it on, and it would still receive a signal. That’s no longer quite true, but it was true for decades. What that meant was that the technology in the TV sets stopped moving forward.
That was what then led to the commoditization of that particular product so that at some point there were no more TV sets being manufactured in the United States. There were a lot of people that were upset about that, but that’s what standards do. In some sense, they freeze the technology at a certain point. It introduces the possibility of commoditization and the loss of, let’s say, the national position. I don’t think that’s what is happening in the semiconductor world, and I’m not sure that it should happen.
A Critical Moment
MASCI: It’s interesting because five years ago the semiconductor industry was the focus of some attention but not nearly as much as it is today. Obviously, some of this is due to the pandemic-related shortages of appliances and cars. We’ve all heard about how these companies haven’t been able to get enough chips. Of course, that’s led to this stampede. We talked a little bit about the CHIPS Act to make sure that the United States has a self-sufficiency in chip-making. All of these companies, Intel and even Samsung and TSMC have announced big investments in the U.S. But is it important that chips are made here? If so, why? I know you’re not a geopolitical thinker, but is it just a strategic thing? We just want to make sure that we can always have enough if all of a sudden Taiwan is blockaded?
CALHOUN: You’ve blended two questions. If I may, let me take the first one, which was that five years ago, this didn’t seem quite so critical. I think what that reflects is the fact that there has been an inflection point in the growth of the industry, the broader use of chips in many products and the use of the penetration of chip technology into so many sectors. If you look back 25 years ago or 30 years ago, when SEMATECH was set up, and we had our first anxiety crisis as a nation over our leadership issue in chips, at that time, chips were really used in computers and maybe a few telecommunications products.
But you didn’t have chips in cars, at least not very many and not in crucial systems. You didn’t have chips in your refrigerator. You didn’t have chips in your kitchen faucet. Somewhere between then and now, I think what happened was we began to realize, “Wait a second, there are chips all over in almost everything.” In an automobile today, you can have hundreds of semiconductor devices. In an electronic vehicle, you might have thousands of semiconductor devices.
The auto industry is a good illustration of that—to go from automobiles being a big hunk of steel and some pistons and tires and rubber and glass, which is what they were not that long ago, to now being really a mobile computing platform that has networked hundreds of semiconductor devices and millions of lines of software code embedded in it. The value that is being sold and created by the car companies is now coming from that and not from, in most cases, bending the metal a little bit better. I think that’s the first thing that happened. What was your second question?
MASCI: Is it important to make chips here?
CALHOUN: I think it’s a useful question to ask from a policy standpoint because one of the risks of a totally free market is that the free market may make some decisions that are blind to the geopolitical concerns of national security or defense or whatever, that you want to also make sure somebody is paying attention to. I think it’s a good question to ask.
I think there’s some value in reminding the industry, let’s say Intel as the industry leader, if you’re going to make your next plant in Malaysia or in China, maybe there’s a risk factor on the geopolitical front that you haven’t thought about enough. And we’re going to give you some reason to think about doing it in Ohio instead of in China.
That said, I don’t think autarky is a viable answer in this industry. It never would work. The notion that we want to have all, or the dominant part, of our manufacturing somehow rooted in the United States, just as a matter of principle—I think it’s easy to go too far in that direction. Like many things, there’s a middle point.
The CHIPS Act
MASCI: Sure. Again, we’ve mentioned the CHIPS Act, but we really haven’t asked you about it. Let’s assume you decide, yes, it’s important that not all the chips that we use be made here, but that a substantial number or substantial percentage share of them are. Is that a good start down that road?
CALHOUN: I think the best thing about the CHIPS Act that I would say is that it’s gotten people’s attention. It’s gotten people talking about the questions that you’re asking and gotten them thinking about, is there a vulnerability that should be factored into these decisions of where to locate and how to promote this industry that maybe two, three, four years ago nobody was really thinking about?
Whether $50 billion can be spent in a productive way here by the U.S. government—I’ve just gotten done saying that the Chinese government was coming at it the wrong way with their top-down, throw money, helicopter money over the industry. Especially since these are firms that are really, unbelievably profitable, and they’re gearing up to spend more than 10 times of their own money in the next decade.
The CHIPS Act will have, I think, some positive effects. Probably, it will also involve some wasteful decisions. Perhaps in the scheme of things, to spend $50 billion to get people thinking about it and talking about it and taking a broader perspective in decision-making related to this industry is not a bad thing.
MASCI: As I recall, the president of Intel hedged a bit on his earlier commitment to building out their capacity in Ohio, implying that if the CHIPS Act didn’t pass, not all bets were off or anything, but that Intel might not make as many investments as they had planned. Obviously, the CHIPS Act did pass. If you were charged with creating a policy that, say, would encourage a greater industrial base that manufactures chips, how would you suggest we go about doing it?
CALHOUN: I guess the first thing I would say is that there are certain market frictions and market failures in the industry that would be appropriate for the government to try to address. But perhaps we shouldn’t put additional emphasis on things that the industry is going to do anyway or maybe where the goal is being defined in some way that is not really appropriate for this type of government subsidy. For example, in the CHIPS Act, it talks explicitly about fostering the most advanced—they even mentioned three-nanometer or below technology. Well, it turns out, when you really dig into it, that the five-nanometer or three-nanometer chip is a red herring in terms of technological performance and technological leadership. It’s not completely irrelevant, but it’s certainly not as crystal clear that that’s the objective that should be written into the law the way it was.
Some of the things that perhaps should have been written into the law that I would have liked to have seen in the policymaking, the rulemaking that will now flow from this, are those that would address the market failures. An example of that is back in the auto industry. The auto industry does not need three-nanometer technology. Most of what they need are so-called legacy chips based on mature technology. Those are the ones that are in short supply, that are holding up the production of so many hundreds of millions of dollars’ worth of automobiles.
I think there are good reasons why the major chip-fabricating companies like Intel and TSMC want to invest in the cutting-edge chips because, unlike the more mature chips that’s where the profitability lies. That’s why making those legacy chips is not as strong a pull for that private-sector investment as is the cutting edge. That’s where the government could step in and say, “We’re going to put some additional incentive in place to make sure that you are not neglecting this piece of the market, which is so critical for the auto industry and for any other industries.”
The smart refrigerator doesn’t need three-nanometer technology either. It needs some basic legacy, mature technology that maybe they can’t get because the industry has chased this high-end product a little bit too much. I think that’s the kind of thing that I would hope the CHIPS Act, and the people that are responsible for implementing the CHIPS Act in the form of rules and specific incentives and policies, would focus on.
Future Disruptive Innovations
MASCI: When you were talking about this obsession with five-nanometer or three-nanometer, in preparing myself for this interview, I read an article that talked about a new way that I think Intel and possibly Samsung are making chips, this stacking technology.
That leads me to ask, are there innovations potentially coming that will make this race to the smallest irrelevant in the coming years and possibly shake up the entire industry? Maybe TSMC and the advantages that it has right now in its fabrication plans just won’t mean anything because everything will be different? Is something like that possibly on the horizon?
CALHOUN: Yes, indeed. Look, the whole history of the industry is a history of those kinds of cycles of disruption, just one after the next. And the leadership—go back and look at who were the leading players in 1990, in 2000 and 2010. There’s a tremendous turnover with a couple of exceptions. Intel has managed to always stay in a leading position.
MASCI: Who hears of Fairchild anymore, right?
CALHOUN: Fairchild. I was reading about the SEMATECH era, for example, recently, and they described Micron as almost a trivial company, and today Micron is a very dominant player in the memory business.
If you take that going forward, I think, yes, there’s every reason to expect that the technology is not standing still, that people are looking at different ways of improving the performance of these chips.
I think that driving the width of the transistors themselves—that’s the seven-nanometer, five-nanometer and that pathway—everybody knows that taken even at its face value, it’s coming to the limits of the physics. And to make the next set of leaps, you’re going to have to push one of the other dimensions of the design space, so to speak.
So I think we can expect that there will be creative approaches to the design problem that don’t involve just butting your head harder and harder against that three, two, one, what comes next? Zero?
MASCI: In a decade there could be a company we’ve never heard of that’s the leader.
CALHOUN: A company we’ve never heard of, yes, exactly.
MASCI: Thank you, George, for coming and for your wonderfully insightful answers. We really appreciate it. And thank you, Christine, for your excellent questions.
MCDANIEL: Thank you, George.
CALHOUN: Thank you. It has been fun.