This transcript was prepared by a transcription service. This version may not be in its final form and may be updated.
Willy Shih: I look at the $52 billion, and I hope the political leadership in this country understands that that’s a nice down payment. But if we really want to secure technological leadership in this country over the long term, this is a long game.
Stephanie Kelton: Welcome to the Best New Ideas in Money, a podcast for Market Watch. I’m Stephanie Kelton. I’m an economist and a professor of economics and public policy at Stony Brook University.
Charles Passy: And I’m Charles Passy, a reporter at Market Watch.
Stephanie Kelton: Each week we explore innovations and economics, finance, technology, and policy that rethink the way we live, work, spend, save, and invest. In August of this year, President Joe Biden signed the Chips and Science Act into law. The legislation which was approved with significant by partisan support allocates $52.7 billion in subsidies to the semiconductor or chip industry. According to the Wall Street Journal, it also includes $24 billion in tax incentives and other provisions as well as nearly $170 billion for technology research and development. That’s a lot of money and it’s been called the most significant investment in industrial policy that the United States has made in the last 50 years.
Charles Passy: You might be wondering why on earth is the federal government spending all this money on semiconductors? Stephanie, I’ll be the first to admit this, I’m not even sure what these things are.
Stephanie Kelton: Charles, you’re not alone. Semiconductors are everywhere. And although we seldom see them, they essentially power our digital lives. I’ll let our first guest explain.
Michael Kozicki: I think in many respects, they’re not very obvious.
Stephanie Kelton: That’s Michael Kozicki, a professor of electrical engineering at Arizona State University.
Michael Kozicki: They sit inside systems, they sit inside your phone and they endow that phone or that laptop or that tablet with astonishing capability, astonishing processing, astonishing communication capability, but we never think beyond the screen. We never think beyond the buttons, if you like, why should you care what’s in it if it works well, and I think we’ve been too good as an industry at hating all these cool things inside systems.
Charles Passy: But what exactly are semiconductors?
Willy Shih: A semiconductor is a device that is generally made out of silicon that is designed to turn electricity, electrical currents, on and off, given some kind of control signal.
Stephanie Kelton: That’s Willy Shih, a professor of management practice at the Harvard Business School,
Willy Shih: Chips are often a shorthand used to describe semiconductors, and usually that means there are devices that have more than one transistor. Oftentimes, they’ll have millions or billions of transistors. But chips are a categorization for a large part of the semiconductor market.
Stephanie Kelton: We asked Shih, how many semiconductors you might encounter on a typical weekday morning?
Willy Shih: When you get up and turn on the lights, especially you if you have some of these new more efficient LED lights, those are semiconductors. When you go to your coffee machine, those will have multiple semiconductors in them to control functions like the little touch panel that turns on your coffee maker or as you stumble in the kitchen and you’ll see all these clock displays on your microwave or your stove, those are all driven by semiconductors. They are pervasive in daily life because they’re such a nice way of delivering product features relatively inexpensively.
Stephanie Kelton: There are different categories of semiconductors which are utilized in everything from national defense systems to consumer tech like your car, laptop or smartphone.
Willy Shih: When we think about the digital devices that we use every day, things in your smartphone or your computer or your television set, those are by and large in a category called logic chips. And those include things like the microprocessors in your computer or your phone. And this digital logic is one of the key segments of the semi connector industry.
Stephanie Kelton: The Chips and Science Act aims to grow domestic production of digital logic and memory chips in large part by subsidizing the construction of the advanced fabrication facilities or Fabs, which are the factories where these chips are made. And while the contemporary effort to support this industry in the United States is new, the country was once a leader in semiconductor manufacturing. Ever wonder how Silicon Valley got its name? Here’s Michael Kozicki.
Michael Kozicki: Now, the US, at one point, was a dominant world force in semiconductor manufacturing and microelectronics and manufacturing of microchips. We led the world in every respect and slowly but surely our lead was eroded. Certainly, we’ve always maintained very, very strong technologies. We’ve always been a tremendous source of new technologies, new materials, new processes, new devices, new circuits, new systems. But the manufacturing as it has in many other areas within the economy began to slip away from us to places where labor was cheaper and there were other ways of saving costs and transitioning away from manufacturing, we began to lose our grip on the industry, on the semiconductor industry.
Stephanie Kelton: And according to Kozicki, that shift away from manufacturing created a feedback loop that will make it very difficult for the United States to regain its once dominant position in the semiconductor industry.
Michael Kozicki: It’s an industry that is driven by advances in technology, which lead to better products that lead to more markets and larger markets, that lead to more money. And then that money gets put into advanced products, which lead to, again, it just goes round and round and round. But if you are not on the leading edge of creating processes, the actual processes used to create the chips, you drop out that loop, that economic engine. And what tends to happen is that you will slip behind and that’s how we found the US, we certainly began to slip behind in our ability to manufacture at the advanced technology nodes.
Charles Passy: So why did American companies move manufacturing overseas in the first place? Here’s Willy Shih.
Willy Shih: It wasn’t always the case that the US offshore at its most advanced chip manufacturing. What happened is, starting in the early 1990s, many American chip makers realized that manufacturing chips was high risk and very capital intensive. And what they found is that Asian companies like TSMC were willing to take on that manufacturing work. So American companies said, “Why don’t we focus on the design part and let the Asian companies do the manufacturing?” And at that time, it was a good way to reduce the amount of capital, how much money you needed to invest to build these factories. And you could say, I will specialize in doing what I do best, which is design, I’ll let somebody else do the, at that time, not very profitable work of doing the manufacturing. So American companies shipped off to Taiwan, and other parts of Asia, work that they really didn’t want to do.
Stephanie Kelton: We’re simplifying the story and there are divergent views on the question of why manufacturing moved offshore in the first place, but broadly speaking, it’s this shift offshore which brings us to where we are today.
Willy Shih: As far as manufacturing capacity, it depends on the category. So for example, when we look at the digital logic chips, and this is where a lot of people are ringing their hands over the loss of US leadership, these are the microprocessor chips from companies like Intel or the chips inside an Apple iPhone or an Android phone. And it is the most advanced digital logic chips where the US has really fallen behind in terms of global manufacturing capacity. There the US is down to perhaps 12% of global market share.
Charles Passy: So where are the most advanced chips being made today?
Willy Shih: Taiwan is the clear leader for those advanced logic chips with one company, Taiwan Semiconductor Manufacturing Company, often referred to as TSMC, who has maybe 50% of global share of the foundry capacity for making this type of chip. And they are the dominant manufacturer for the most advanced generations of chips.
Stephanie Kelton: And that’s the aim of the CHIPS Act, to help domestic manufacturing compete with the most advanced and dominant manufacturers in the world. But the federal effort isn’t only about industry, it’s about national security. Taiwan which produces the most advanced chips in the world is about a hundred miles from mainland China. Some worry that if China were to attempt to seize Taiwan by force, it would control one of the world’s key resources, an essential technology used in everything from consumer tech to defense.
Charles Passy: Now if that happened, we’d probably have a lot more than chips to worry about. And there are other reasons why it’s in the national interest to manufacture chips in the US. Here’s professor of electrical engineering, Michael Kozicki.
Michael Kozicki: We rely on the microelectronics industry to essentially fuel the technology and our nation’s defense also. And that’s an issue, of course, having chips that ultimately go into defense systems made in foreign countries because we really don’t know what’s happening to them. There’s no guarantee that secret back doors might not be slipped in there and maybe even deliberate ways of making them fail might be slipped in there and it’s quite terrifying in that respect. So there’s the commercial aspect of just being at the mercy of other companies in other countries and then there’s a national defense and security aspect of having to rely in other companies to essentially put chips into the systems that are protecting our country.
Charles Passy: It’s worth noting that companies which receive subsidies from the CHIPS Act are prohibited from expanding manufacturing in China or in other countries which pose a national security threat. But setting aside national security concerns, according to Willy Shih, there are real economic and commercial reasons why making chips at home is a worthy goal.
Willy Shih: I think it is very important for the US to be able to manufacture the most advanced chips because with that capability comes the ability to innovate in the future. So I have always felt that it is important to keep our fingers in the manufacturing. We have one company, Intel, which does produce advanced technologies. They are behind TSMC and also Samsung of Korea in that regard, but I think Intel is urgently trying to catch up.
Charles Passy: When we’re back, can the CHIPS Act accomplish its purpose and how might we determine if this was money well spent? That’s after the break.
Stephanie Kelton: Welcome back to the Best New Ideas in Money. Before the break, we talked about semiconductors and why the federal government is spending billions of dollars to subsidize the production of this technology at home.
Charles Passy: Let’s take a closer look at the bill itself, the CHIPS Act, and where this money is going. Here’s Willy Shih, professor of management practice at the Harvard Business School.
Willy Shih: So the CHIPS Act, actually more formally the Chips and Science Act, I think has the objective of trying to restore American competitiveness in semiconductors and semiconductor manufacturing specifically. There’s $52 billion of that total, which is allocated towards helping manufacturers establish new facilities in the United States. So clearly Intel in the state of Ohio is looking for some help, as is TSMC who is building a new fab in Phoenix and Samsung who is building or has committed to expanding its facilities in the Austin, Texas area. There are also numerous other smaller manufacturers like Global Foundries, who has a major operation in malted New York. And you have companies like Micron Technologies, who’s one of the big three in memory chips, who is promising with help from CHIPS ACT funding to build a large memory chip factory in Idaho.
Charles Passy: Even though $52 billion is a lot of money, that’s also a lot of factories or Fabs, we’ll come back to that.
Willy Shih: There’s a large part of funding in the CHIPS Act that is also directed towards fostering more R and D and development activities in the United States where I think the country has really under-invested over the last several decades. That longer term, I think potentially is more significant.
Charles Passy: So how many Fabs will $52 billion buy?
Willy Shih: Typically for the most advanced Fabs these days we’re seeing price tags of 15 to 20 billion dollars. In Asia, it might take you a year from when you announce to when you get that building completed. Once you have the building completed, you have to have all your equipment ready to move in. And when you move the equipment in, it will take some amount of time to install it and clean it so that’s probably a two year process. These things tend to take longer in the United States than they do in Asia. I’ve been in Asia on Christmas day and I see people working on construction because they’re on the job seven by 24 until they get these things done. We tend not to do things that way in the US.
Charles Passy: We’ve talked about the material components of manufacturing, but you do need actual people to operate these Fabs and that’s another big piece of the puzzle. Here’s Michael Kozicki.
Michael Kozicki: When I came to US, Silicon Valley was indeed Silicon Valley. There’s no question about it. We were making stuff like you wouldn’t believe in Silicon Valley. But then of course, the emphasis switched more to what I would say is Soft Valley, more in the software and the app side of things because frankly, it was way easier to do.
The barrier to entry to semiconductor manufacturing is huge. So of course, it was way easier for companies that, and forgive my downplaying here, but that just did software. And that’s why I have to admit, I lament the days when Silicon Valley was indeed producing lots of silicon and less in terms of dating apps and the… which are obviously way easier to do than building a leading edge microchip. And fortunately this spills over into the perception of working in the industry too, because people see writing software as being a lot more exciting, a lot, frankly, easier to do than actually making microchips, which is still complete mystery to a lot of people. It’s alchemy rather than engineering. And so people do tend to drift into the software disciplines and university, and you can see how that’s ultimately going to really spur companies and company development and spinouts and startups that are more in that area rather than the really hard tech like semiconductor manufacturing.
Charles Passy: It’s a little glib, but you could think about it this way, the software doesn’t work without the hardware and we need people to make the hardware.
Michael Kozicki: If we aren’t getting people who have an appreciation for actually making hard things, you’re making real things, unfortunately, what’s going to happen is that we’re going to see this negative feedback loop coming in where everything will get driven away from hard tech because nobody wants to do it. When nobody wants to do it, nobody’s going to want to train in it, and if nobody wants to train it, fewer people will be there to do it and it gets worse and worse and worse. So I think unfortunately this is in terms of workforce is something that we may be facing in that we are creating an entire generation of people who don’t know what manufacturing is.
Stephanie Kelton: For Willy Shih, the CHIPS Act is a strategic bet and perhaps even a necessary one.
Willy Shih: If these investments are successful, these factories in a couple years are busy and producing lots of chips, it will be great for the United States because we’ll have more of those capabilities on shore. It will create jobs, it will create economic value from the production. We should not forget that everywhere else around the world is racing to make investments in this area as well, whether it’s across Asia, China in particular, also Taiwan, also Japan, but Europe as well. In some sense, if we didn’t do this, we would go from our 12% share digital logic ships to an even smaller share so in some sense it’s a matter of keeping up.
I think the biggest thing that I worry about is this is a long game. If you’re going to really strive for technological leadership in semiconductors in the United States, you have to have the perseverance to last past one election cycle, two election cycles, five election cycles, probably a lot longer than that. TSMC did not get to be the world leader overnight. Rather, they have been doggedly working of this and investing in this for 35 plus years now. And we look at the $52 billion in subsidies from the CHIPS Act, and I hope the political leadership in this country understands that that’s a nice down payment. But if we really want to secure technological leadership in this country over the long term, this is a long game that will take persistence over decades, not one election cycle.
Charles Passy: While the legislation ultimately pass with significant bipartisan support, the bill wasn’t without its critics, including Independent Senator Bernie Sanders of Vermont and Republican Senator Rick Scott of Florida, they argue that it’s a massive giveaway to profitable corporations who are in no need of a helping hand. And it remains an open question, is investing billions of dollars in bringing advanced manufacturing to the US a best new idea in money? Here’s Michael Kozicki.
Michael Kozicki: Yes, it is a fantastically smart idea. I don’t think there’s any doubt about that simply because it is at the most basic level, it’s a signal. It’s a signal that the country cares about semiconductor manufacturing. It’s bringing confidence to the players in the industry that the country cares and I think it’s a great idea. I think it’s a wonderful idea.
Stephanie Kelton: We asked Willy Shih if he thought anyone could catch up with Taiwan’s Semiconductor or TSMC.
Willy Shih: I’m skeptical. I’ve been in Fab 12 and it’s gigantic. And remember, TSMC has four giga Fabs like that. So I don’t think people appreciate how many hundreds of billions of dollars of investment this is going to take. That’s why I say 52 billion is a nice down payment, but we shouldn’t delude ourselves into thinking it’s anything more than that.
Stephanie Kelton: It may be a few years, but what will he be watching to determine whether this money was well spent?
Willy Shih: The most important indicator, and I don’t see a lot of people talking about this, is capacity utilization. And the reason I say that is when you put up expensive factories like this, they need to be filled and they need to be producing chips that these companies can sell both domestically and around the world. On an expensive factory like this, if your utilization falls so that you have unused capacity, you will lose money faster than you can ever imagine. There is so much money being poured into building factories like these around the world that a lot of people, myself included, are worried about what is the utilization going to look like in a couple years when all these factories come on stream? The bet is that the growth and demand for chips in machine learning and artificial intelligence applications, the digitization of every device in your house or office or factory, that that demand growth will be fast enough to keep all these factories filled. If it isn’t and these factories run at low utilizations or less than economic utilizations, the next scene in that movie is not very pretty.
Charles Passy: What about Michael Kozicki? What will he be watching?
Michael Kozicki: So there’s two things. Number one, and maybe this is the obvious thing, is are companies advancing on the world’s stage, if you like, in terms of technology? Are the advancing to become equal to or better than the TSMCs of the world in terms of actual technology development? So that’s one thing, but the thing I’ll be looking for, I’ll be absolutely honest, is more students coming to university actually saying, “Hey, I would love to work in microelectronics. I think this is the most exciting thing. I want to be part of this.” That’s the thing I’d be looking for. It just doesn’t exist right now.
Stephanie Kelton: Ultimately, Willy Shih believes that the success or failure of advanced semiconductor manufacturing in the United States is an important indicator.
Willy Shih: I think how we handle advanced semiconductor manufacturing in the US will be a bellwether for how we choose to handle other advanced technologies. Make no mistake, the US has also seated capabilities in many other manufacturing sectors that I think are going to be critical that we’re going to regret. So what we do in semiconductors could be a leading indicator of our willingness to take on higher costs and some of the challenges associated with maintaining capabilities in those areas.
Stephanie Kelton: Thanks for listening to the Best New Ideas in Money. You can subscribe to the show wherever you listen to podcasts. If you like what you heard, please leave us a rating or review. And if you have ideas for future episodes, drop us a line at [email protected]. Thanks to Michael Kozicki and Willy Shih. To learn more about semiconductors, head to marketwatch.com. I’m Stephanie Kelton.
Charles Passy: And I’m Charles Passy. The Best New Ideas in Money is a podcast for MarketWatch. Melissa Haggerty is the executive producer and the producers are Katy Ferguson, Matty Lukshoff and Michael McDowell. Editing and mixing by Steve Cooper. Jeremy Banks is our news editor and Tim Rostan is the executive editor for MarketWatch. The Best New Ideas in Money theme was composed by Sam Rekser. Stephanie Kelton is an economist and a professor of economics and public policy at Stony Brook University and not part of the Market Watch Newsroom. We’ll be back next week with another new idea.
