Podcast: When the Chips Are Down: Why Domestic Semiconductor Production Matters, With John Zysman

The United States used to be a leader in semiconductor production, but its share of global output dropped from 37 percent in 1990 to just 12 percent in 2019. That helps explain why the country now faces serious supply issues. Rob and Jackie sat down with John Zysman, a professor emeritus at UC Berkeley and co-founder/co-director of the Berkeley Roundtable on the International Economy, to discuss why U.S. semiconductor production is down, what it portends, and how America can regain its footing in the industry.

Mentioned

• Stephen S. Cohen and John Zysman, Manufacturing Matters: The Myth of the Post-Industrial Economy, (Basic Books, 1987).

Related

• Rob Atkinson, “Potato Chips, Computer Chips: Yes, There Is a Difference” (ITIF, December 2020).
• Rob Atkinson, “Computer Chips vs. Potato Chips: The Case for a U.S. Strategic-Industry Policy” (ITIF, January 2022).
• Stephen Ezell, “Incentives Essential to Ensuring America Possesses a Leading Semiconductor Industry” (ITIF, February 2022).

Auto-Transcript

Rob Atkinson: Welcome to Innovation Files. I’m Rob Atkinson, founder and president of the Information Technology and Innovation Foundation. We’re a DC-based think tank that works on tech policy.

Jackie Whisman: I’m Jackie Whisman. I head development at ITIF, which I’m proud to say is the world’s top ranked think tank for science and technology policy.

Rob Atkinson: This podcast is about the kinds of issues we cover at ITIF from the broad economics of innovation to specific policy and regulatory questions about new technologies. Today, we’re going to talk about semiconductors, why they’re important, what’s the impact of the current scarcity, and why it’s critical to keep them secure and plentiful.

Jackie Whisman: Our guest is John Zysman. He’s Professor Emeritus at UC Berkeley, and co-founder/co-director of the Berkeley Roundtable on the International Economy. His ongoing work covers the implications of platforms and intelligent tools for work, entrepreneurship, and international competition.

He’s published three books and has written extensively on European and Japanese policy and corporate strategy, and most recently a lot of work on semiconductors and their security. We’re happy to have you here.

John Zysman: Pleasure.

Rob Atkinson: Hate to mention this John, because it dates us, but when I was getting my PhD way back in the days, you were and your colleagues at BRIE, Berkeley Roundtable of International Economics were a key source of my inspiration at the time to focus on this whole question, so thank you.

John Zysman: Well, that’s very generous of you, and I might remark, given that we’re talking about semiconductors, that the people who helped us found BRIE, the Berkeley Roundtable on the International Economy, the first check was written by David Packard, and the first meetings were hosted with Bob Noyce who founded Intel, for those in your audience.

And Jerry Sanders who founded AMD, and Charlie Sporck who founded National, in which the issue at the day was the character of Japanese competition, the role of the Japanese state and of closed markets in promoting competition there.

They’ve come a long way from the days when they believed, not incorrectly, completely, that they were creating a new utopia.

Rob Atkinson: Absolutely.

Jackie Whisman: We hear a lot about how semiconductors are an essential product. I’d like if we could start by you explaining to us and our audience some uses of semiconductors, maybe to help illustrate how critical they are.

John Zysman: I will say that the easy answer is, what are they not embedded in at this point? I bought a new car last summer, and it was delayed three months because they couldn’t get the chips.

It’d be interesting to find a product that you’re buying, at this point, that doesn’t depend in one way or another on the processing power of either a simple or powerful chip, or the sensor capacities of that tool, or the information embedded in that tool.

It’s come a long way in those years since Noyce and Sanders helped us. In those days, folks really did wonder what was the importance of the semiconductor. The answer is, it’s the foundation of the digital era.

Rob Atkinson: Back then, John, a lot of those semiconductors, well, initially they went to the military when Fairchild was making them. Then they largely went into computers. You and I both remember the term Wintelism, Windows and Intel.

They were obviously in other things, but we’re so far beyond when they’re just in computers now. The other thing that’s changed over the years is, frankly, that we used to lead. The US invented semiconductors out of Bell Labs, and then at Fairchild and Intel, and a whole set of other companies that were just incredible. We were so far ahead of the world.

Now, we make just about 11% of the semiconductors in the world. A lot of other countries make more, Taiwan, South Korea. What happened?

John Zysman: Well, I think that part of the question is, does it matter if you dominate design and you dominate the equipment and materials, if you are the company that does the final assembly?

For the audience then, I think we should separate the fact that a company like Qualcomm designs chips. A company like Applied Materials builds the equipment. ACML in the Netherlands builds the lithography that draws the lines on the chips.

But, in fact, you’re absolutely right, the final assembly, if you will, the production of those chips has heavily come to be dominated by Samsung and, above all, TSMC in Taiwan. The question becomes, what difference does that make?

Part of the question, I think, is, and I think the position that my colleague Laura Tyson, who many of you probably, your audience probably knows, ran the National Economic Council when Bill Clinton was in office, when we were all young.

In fact, we would argue that we need to have a competitive, resilient, secure, and sustainable semiconductor industry, one that, in fact, throughout the supply chain is competitive. That secure part of that really does mean it needs to be regionally distributed.

There are many problems with this being so concentrated in a handful of firms, particularly those located in one region of the world. A significant earthquake, let alone any geopolitics, and I think we’ve learned some lessons out of the Ukraine, that geopolitics still remain, could deeply disrupt matters.

You asked a specific question, is why don’t we produce the chips here. Would you like me to address that specific issue?

Rob Atkinson: Please do. We used to produce, as my colleague Stephen Ezell has shown, about 40% or so, and now it’s down to about 11%.

John Zysman: Well, I think there are obviously several reasons. One of which is that TSMC made a basic bet with money from the Taiwanese government that, in fact, you could build foundries and let others design the chips. We at various points didn’t worry too much about that.

A second part of that is, as we all know, Intel made a variety of gambles on what to do with the income it had. It didn’t reinvest in the same way one might have expected it to in the production capacity. It fell from a leadership position to somewhat behind.

Of course, there are the issues that it is expensive and to some extent slow to design and build new foundries in the United States. We really should be clear that for much of the supply chain, the US is well-positioned. It’s in the foundries of the most advanced chips where the challenges are real. You and I would share a view of that.

Rob Atkinson: You mentioned TSMC, and Intel maybe not investing as much. I was on a panel once with Clay Christensen, who unfortunately recently passed away, and Clay was talking about how too many American companies are thinking about stock market.

He said that they’re focused on return on net assets, so the analysts look at what’s called RONA. He said, you can boost your RONA by reducing your assets, so your return on net assets goes up if you have fewer assets. Which he said a lot of American companies do, or you can just try to get higher returns.

He said he was talking to Morris Chang, who, as we know, was the founder and at the time was the CEO of TSMC, and he said, “We measure profits at TSMC in tons, not ratios.” In other words, they’ll spend $100 million on a machine if they can make a lot of money off of it. They’re not worried about short-term market issues. How much do you think that matters?

John Zysman: Well, let’s put a real label on it, it’s the financialization of the economy, and a dismissal of the importance of manufacturing in general. As you know, Rob, back in the day of the early dinosaurs, I and a colleague, Steve Cohen, wrote a book called Manufacturing Matters. In fact, I do agree with that.

It is the focus on financial measures, rather than longer-term competitive position that, in fact, led to some of these troubles, and an assumption that you could outsource everything and that it would always be easily available to you.

There are two parts to that. One is, do you lose the leverage that came from being able to control the matters such as where you’re going to produce your own product? Secondly, do you cease to learn enough about the next generation of products from what you would learn if you were making those products?

We dismissed all of that, and moved to a notion of outsourcing as an easy solution to many things, when, in fact, it oftentimes was a trap.

Rob Atkinson: I wish, John, that you would come to Washington and debate some of these folks, because I think, even though they might say they’re economists, I sometimes wonder. Certainly I doubt they read your book.

My favorite is Adam Posen, who’s head of the Peterson Institute, who had a piece in Foreign Affairs recently, where he basically said that anybody who’s focused on manufacturing, it’s a fetish that probably reflects an underlying racism and sexism, believe it or not, because it’s just about helping white working-class men.

Clearly didn’t know anything about the demographics of manufacturing, and clearly doesn’t understand the real economy. How do we overcome that myopia, to be kind?

John Zysman: I can only recall having actually had a dinner party with the infamous Milton Friedman, and having a debate over whether semiconductor chips mattered more than potato chips. Quite literally it’s an old line, but he was using it at the time.

I think the issue, and Adam is a very smart guy, and about finance knows a huge amount, but I think part of the issue is that folks lose track of the fact that for the most part manufacturing is now highly automated. This generation of manufacturing innovation is not about jobs.

It is really about changes in the kind of robotics that are going to be used, the flexibility of the robotics that are going to be used. It’s precisely because it is so highly automated at this point and so highly dependent on semiconductor control chips, or semiconductor-controlled processes that it matters a great deal.

I think we need to understand that, and move forward.

Jackie Whisman: What are some things that the US government should be doing to facilitate a more stable supply of semiconductors?

John Zysman: Well, I want to separate two things, if we will. What we need to do with the supply problems now, and the longer term issues. The supply problems now, in some sense, not much can be done real quickly in the next six months, because any kind of foundries they’re building are going to take several years to actually get put into place.

I think that we need to obviously make investment attractive, but, equally important, we should, in fact, recognize that if Intel builds a foundry in Europe, or Samsung or TSMC builds a foundry in the United States, both of those contribute to the regional diversity of where the production takes place.

We don’t want to be in a position that a single earthquake somewhere or an invasion somewhere actually disrupts those global supply chains. The other part is obviously there is a huge discussion, and we’ve supported that in discussions both with the Secretary of Commerce and with the staff, to support American investment in fab production and other things in the United States.

The trick will be, and let me, rather than be specific, say that the real trick will need to be how do we make sure the money goes into fabrication facilities or semiconductor production and not to the bottom line of the companies. Not a recreation of the financialization that caused problems in the first place.

Yes, we need to support the companies building here in a variety of ways, but we have to make sure that competing for location does not absorb the money. That is, we offer a billion dollars in whatever to a company, and the French offer two billion, so, in fact, they build it in France.

In fact, because the company in many cases had the cash in the first place, that goes to the bottom line. The trick will be how to encourage production financially in the United States without simply having it benefit shareholders.

The second is can we deal with the problems of how to address the real issues that come up from environmental and zoning restrictions? How do we get the things built in the first place, without, in a sense, undermining other objectives such as a clean environment?

It’s a challenge, but we need to directly address it, and get money and companies moving forward toward production here, as well as elsewhere.

In the longer term, I think, and I would like to turn to that as well, is there’s the question of how do we get innovative ideas out of the lab, whether it be a corporate lab, or a lab at Berkeley or MIT or Stanford, into actual fabrication.

Part of the difficulty with that is, it’s increasingly expensive to, in fact, move things out of the lab into beta, where they can then be picked up by various companies. As a result, there have been specific suggestions that we create semiconductor commons where firms and researchers can share equipment, so that it becomes operational expenses rather than capital costs.

Many of the capital costs of moving goods out of a lab would be $100 million, and no researcher’s going to do that. If we create those kind of semiconductor commons, there’s an opportunity.

The opportunity is not only to encourage that kind of innovation, but to open those, both here and in Europe and in Asia, to like-minded countries and the firms of like-minded countries setting down the terms on which firms and countries can participate.

In a sense, redo some of the intellectual property and trade issues that are very hard to deal with in the WTO, or impossible to deal with in the WTO at this point.

Jackie Whisman: You’ve been really careful, in what I’ve read recently, to emphasize that the US can’t do this alone. You’ve warned against weaponizing trade restrictions, particularly with China. Can you say more about this?

John Zysman: Well, I think very simply we need to assure, as I’m saying, a competitive, resilient, secure, sustainable system. We can do that with like-minded allies. We should do it under any circumstances. We don’t have to make it an anti-China event.

It will be successful in dealing with many of the Chinese problems if we’re successful in doing what we should, in any case. That is, building capacity here. Accelerating, we have to run even faster to stay ahead of the innovation objectives of the Chinese as well.

Do I think there’s a Chinese challenge? Absolutely. Do I think it’s a serious issue? Absolutely. I think building alliances with our like-minded allies is probably the most effective way of addressing that.

Rob Atkinson: John, one of the things that, I think, we all know, the Trump Administration took some actions against China with regard to semiconductors. There was a Chinese company, which I will mispronounce, Fujian Jinhua, and it was making memory chips, DRAMs, dynamic random access memory chips.

Its factory had been funded by a government integrated circuit fund, so 100% subsidized. It was using a design technology that had been stolen from Micron, actually one of Micron’s facilities or partnerships in Taiwan.

The administration, I thought, did a really smart thing, they basically cut off their ability to get chip making equipment, particularly from Applied Materials. The company, my understanding, is in stasis if not dead. Which, in my view, they deserve everything and more.

That’s not to say that every Chinese chip company is in that category, because I’m not saying they are. Clearly when you have a category like that where you’re building something off of a stolen design, that’s pretty egregious to me.

Where do we go with export controls on semiconductors? [inaudible 00:17:47] the issue now, as we speak, is that the administration put in place export controls against Russia for semiconductors from a lot of them.

What I thought was interesting is rather than use the domestic product rule, which could implicate other countries, they actually got over 30 countries, is my understanding, to voluntarily cooperate. The Koreans are volunteering, the Japanese, the Taiwanese. The Chinese are not, to my knowledge. The Europeans are.

That seems like a pretty interesting way, that finally there’s a will where countries are saying, “We’re going to cooperate so that we don’t cheat and gain advantage on each other to stop a pretty egregious act of global aggression.”

What are your thoughts on where this goes? Is this a good thing, good sign? Is it a one-off? What’s likely to happen?

John Zysman: I like where you’re taking the conversation. Let me remark that I will leave to you to design the export controls. Obviously, some of it will need to be done. I want to emphasize this problem of the need to run even faster.

That is, technology is going to be stolen. We should do everything we can to avoid it, but it’s going to get stolen. In Providence, Rhode Island, there is a wonderful mill. I’m forgetting its name. I have a sweatshirt from it.

Rob Atkinson: Samuel Slater.

John Zysman: Slater’s, that’s right. The sweatshirt I have from Slater’s mill says, “Innovation started here.” Leaving out the fact that he fled from England stealing the technology from his masters in England. This isn’t something ...

Rob Atkinson: Subject to the death penalty.

John Zysman: Subject to the death penalty. The point I’m saying is that we should try and control that, but what we can control even more is running even faster, and trying to, in fact, accelerate the pace of innovation, and getting things out of our enormously innovative system. That’s number one.

Number two, and I agree with what you’re saying about the allies joining together, is that requires, however, that we, in fact, recognize and understand and mutually support the concerns other countries have about access to, let’s say, semiconductor technology.

It’s perfectly reasonable for the Europeans to want to have production in Europe. How do we, in fact, find ways of acknowledging and addressing the concerns that those allies have, so that on the issues that we consider important they don’t feel that we’re just manipulating them.

I think that’s crucial, and it’s going to be all the more crucial not just in the semiconductor area, but in more delicate areas, ones where you and I might well disagree about how they should be handled, about platform technology, data, and the like.

Others have different concerns. We want them to cooperate with us. We have to at least understand and acknowledge those concerns. Yes, building those alliances, I think, is crucial, and I think is something that ITIF should definitely undertake to promote.

Rob Atkinson: Thank you, John. We’ve certainly been talking to other countries, most recently the Koreans and the Japanese and the Brits actually. I don’t want to say formal conversations, but I’ve had conversations with government officials in all three of those countries.

There does seem to be a real appetite now for putting something more concrete in place. There’s a lot of rhetoric, but now there does seem to be an appetite. The EU is trying to pass this big, well, they’re going to pass this semiconductor bill.

They’ll have funding. We’ll have funding. There does seem to be a lot of opportunity to put in place real partnerships along the lines you talked about.

John Zysman: I think that would be great. I think that, in fact, the question is how should that be done. To some extent, for example, the lab to fab project, which I was suggesting, is precisely to say to others, “Why don’t we all join together and move the digital era forward together? As long as you’re not going to steal technology, and you’re going to honor intellectual property rules and trade rules.”

Part of the question from an American side, and I think it’s a question that we could usefully discuss, is should those kinds of activities and promotions come out of DARPA, the Defense Department, or out of the Commerce Department.

Now, the Commerce Department isn’t ideally orchestrated and organized to do that, though the senior leadership knows a great deal about industry and is quite sophisticated even after Gina Raimondo, below Secretary Raimondo, as well.

The real risk is if things are done through DARPA, that militarizes them, and may make them more difficult to create collaborations. The question is how do we build that capacity for action in the United States and in an agency? Maybe we set up a special agency to do that, but not one that is, in fact, militarized.

It will be easier to work with the French or the Germans or the Koreans and Japanese in that case.

Rob Atkinson: Well, you’re singing our tune. ITIF has proposed an independent agency, as has our colleague, Erica Fuchs, at the Carnegie Mellon. As we were talking about earlier, Eric Schmidt’s group, the AI commission that he created and launched, or that he headed, they also proposed some kind of independent technology agency.

Which is probably worth another chat some time, John, but at this point we have to unfortunately wrap up. I really enjoyed this conversation, as always. Thank you for being with us.

John Zysman: Thank you. I look forward to a chat in Washington over a good meal and a glass of wine.

Rob Atkinson: That sounds excellent.

Jackie Whisman: That’s it for this week. If you liked it, please be sure to rate us and subscribe. Feel free to email show ideas or questions to [email protected]. You can find the show notes and sign up for our weekly email newsletter on our website, itif.org, and follow us on Twitter, Facebook, and LinkedIn @ITIFdc.

Rob Atkinson: We have more episodes and great guests lined up. New episodes will drop every other Monday. We hope you’ll continue to tune in.