Front Burner - The big microchip problem

Episode Date: February 6, 2023

At the World Economic Forum in Davos, Switzerland last month, Intel CEO Pat Gelsinger made a prediction about how the world will vie for resources in the coming decades. For years, Gelsinger said, much of geopolitical relations have turned on access to oil reserves. But in the future, he thinks a more important factor will be where microchips are made. Intel is a prominent figure in the computer chip business, but some 90 per cent of the world’s most advanced chips are currently made by one company in Taiwan. And according to Chris Miller, if the TSMC plant in Taiwan was destroyed the disruption of everything from smartphones to cars could be the biggest manufacturing shock since the Great Depression. Miller is an Associate Professor of International History at Tufts University, and he recently released the book Chip War: The Fight for the World's Most Critical Technology.

In the Dragon's Den, a simple pitch can lead to a life-changing connection. Watch new episodes of Dragon's Den free on CBC Gem. Brought to you in part by National Angel Capital Organization, empowering Canada's entrepreneurs through angel investment and industry connections. This is a CBC Podcast. Hi, I'm Jamie Poisson. And before we start today's episode, we've got to talk about a balloon. And before we start today's episode, we've got to talk about a balloon. The Pentagon is confirming tonight a suspected Chinese spy balloon has been spotted over the U.S. We are told the suspected Chinese surveillance balloon is the size of three buses with a technology base suspended below.

Flying over sites the Pentagon believes to collect sensitive information. So over the weekend, this saga unfolded about this big white bubble hovering some 18,000 meters above North America. The U.S. says it first became aware of the balloon last Saturday when it drifted into U.S. airspace above Alaska. A couple days later, it came into Canada's airspace before blowing back over Idaho. And after the Pentagon made a statement about the situation on Thursday, it seemed both diplomatic and television channels became clogged with balloon talk. U.S. Secretary of State Antony Blinken has postponed a planned visit to China following the discovery. I made clear that the presence of this surveillance balloon in U.S. airspace is a clear violation of U.S. sovereignty

and international law. The Chinese foreign ministry says it regrets what we called a civilian balloon straying into U.S. airspace. Beijing says it's a civilian object meant for scientific research. Reports emerged of a similar balloon over Latin America, which the Pentagon

said was another spy balloon from China. On Friday, U.S. President Joe Biden approved plans to wait until the balloon in its airspace was over water to shoot it down, out of concern for debris hitting planes or civilians. And on Saturday, an F-22 stealth fighter jet shot down the balloon with an air-to-air missile. It successfully took it down, and I want to compliment our aviators who did it. Video posted to social media appeared to show the balloon bursting into a puff of smoke and spiraling downwards. The U.S. military began a mission to collect and analyze the debris from the coast near South Carolina. As for Canada, Defense Minister Anita Anand said Canada unequivocally supports the decision to shoot down the balloon. Meanwhile, China condemned

the U.S. for an overreaction and threatened it may respond. It's another blow to the two countries' relations after years of low points. And with that in mind today, I want to bring you a conversation I had on Friday before the talk about the balloon really exploded. It's an episode about a strategic resource that some worry could be the fulcrum of a U.S.-China power struggle in coming decades, but one that's more tangible than information and data. In fact, it's probably in your pocket right now.

You know, geopolitics across the world has been defined by where the oil reserves are for the last five decades. For the next several decades, where the fabs are is more important. This is Intel CEO Pat Gelsinger speaking to CNN at the World Economic Forum in Davos last month. The fabs he mentions there are microchip fabrication plants, Chips, he mentions there, are microchip fabrication plants, the small number of facilities that make computer chips for everything from our phones to ballistic missiles. Because every aspect of human existence is becoming more digital. Everything digital runs on semiconductors.

Build the fabs where we want them. And like Intel is definitely a company that wants the U.S. to heap more money into making chips. But Gelsinger makes a really interesting point. Because as the U.S. and China continue down a path of economic and diplomatic rivalry, they've got an ever-increasing demand for microchips. And the place that makes 90% of the world's most advanced chips is an island caught between the two countries' influence, Taiwan. My guest today says the plant in Taiwan disappearing could mean trillions in losses and years of lagging technology. Chris Miller is an associate professor of international history at Tufts University, and he's just released the book Chip War, the fight for the world's most critical technology. Hey, Chris, thanks so much for coming on FrontBurner.

Thanks for having me. So we know that our world has depended on oil, right? Like when prices go up, the price of everything goes up. We've fought wars, at least in part, for oil supply. has depended on oil, right? Like when prices go up, the price of everything goes up. We fought wars, at least in part, for oil supply. But we don't really talk about microchips nearly as much. So explain for me, what are some ways our world now depends on chips, sort of like it has on oil? Well, that's right. Today, you can't go about your daily life without touching hundreds, or in most cases, thousands of semiconductors. Wake up in the morning, your alarm clock has

a semiconductor inside. Turn on your coffee machine, there's multiple semiconductors inside of there. Sit in your car, there's a thousand semiconductors often inside of a new auto. And that's even before you've opened your computer or looked at your smartphone, which are devices that couldn't function without semiconductors. So we rely on semiconductors more and more every single year. We just never see them because they're buried deep inside of our electronic devices. Today, about a trillion chips are made a year, or 128 for every person on the planet. And what does a chip look like today versus what it looked like 50 or 60 years ago? Well, there's many different kinds of chips, but a typical chip will be the size of your fingernail.

It's a small piece of silicon, and carved into it will be thousands or, in many cases, billions of tiny circuits that flip on and off. There's a device called a transistor in each of these circuits that turns it on and off. And when the circuit's on, it produces a one. When it's off, it produces a zero. And this creates all of the ones and zeros undergirding all digital computing, all the data we process, all of the data we remember are tiny circuits that are being created or interrupted by these microscopic transistors. And to pack billions of these on ships, they need to be really small. And so the smallest transistors today are measured pack billions of these on chips, they need to be really small. And so the smallest transistors today are measured in the number of nanometers that define them. And a nanometer is one billionth of a meter, making transistors roughly the size of a

virus. Roughly the size of a virus. Wow. Like a chip this dense with all these ones and zeros on it, how hard is it to manufacture them? Well, it is extraordinarily hard to manufacture them. There's nothing that really comes close to the difficulty, actually, of making a chip. Billions of components, each one the size of a virus. This is really the hardest manufacturing process humans have ever undertaken. And doing it requires ultra-complex machinery, a supply chain that stretches from Europe to Japan to Taiwan, and a chip-making facility, because of the complexity, can cost $20 or $25 billion to make. They're some of the most expensive factories in human history. So you're saying I just basically couldn't up and decide to start a chip factory?

It's so complex. So much specialized knowledge is involved and so much money is involved that it's almost impossible to become a new entrant into the industry at the cutting edge level. There's just a couple of firms that are capable of producing chips anywhere close to the cutting edge. And there's one company, the Taiwan Semiconductor Manufacturing Company, that produces 90% of the world's most advanced processor chips. It's by far the world's most important chip maker. Taiwan Semiconductor Manufacturing Company, or TSMC. And it is a company that makes customized chips for a lot of global tech companies, including Apple and NVIDIA and MediaTek. I know this is a question that has a lot of history behind it. And of course, you just wrote a whole book about this.

Why is Taiwan such a powerhouse for producing chips? Well, it dates back to the 1980s when the Taiwan Semiconductor Manufacturing Company was formed. At the time, it had a really unique business model. Before that point, most chips were both designed and manufactured by the same company. But TSMC, when it was founded in 1987, tried to do something new. They wouldn't design any chips in-house. They would only manufacture them. And that attracted lots of different customers that didn't want to have to deal with the complexity of manufacturing, were happy to outsource it to someone else. So TSMC was able to attract lots

of customers, as a result, grow far larger than its competitors and use its scale to benefit from efficiencies in terms of cost, but also use its scale to learn from the production process of manufacturing more chips. And so today, it's not only the world's largest chip maker, it's also the world's most advanced because it hones its manufacturing processes over a far larger number of chips than its competitors. And that makes it the most important chip company in the world today. So 90% of all chips, I mean, it's easy to see why the situation is so politically fraught, considering, too, how you've explained how this makes up really like the backbone of our society. And for our listeners, like Taiwan asserts it's an independent state. The U.S. has pledged varying degrees of military support for its defense. China says Taiwan is actually part of China.

I guess this gets worse when you consider that chips are strategically important to both the U.S. and China, right? For example, like, why are they so crucial to U.S. military power? Well, just like the typical person touches lots of chips over the course of their daily life, military systems are just as reliant on semiconductors. And so if you look across military systems, whether tanks or planes or artillery, they have lots of semiconductors inside. But especially when you think of the next generation of military devices, drones, for example, that can fly themselves semi-autonomously, they're going to require a lot of semiconductors inside, not only in the processing on the drone, all the sensors, radio sensors, the optical sensors, the infrared sensors

that they're operating, but more importantly, this is the critical thing, the data centers in which artificial intelligence systems are trained, the place where you hone their capabilities is vast data centers. And these data centers are capable of training AI systems because they've got some of the most advanced ships the world can produce inside of them. And so access to the world's most advanced data centers lets you train the most advanced AI systems. And that's gonna be relevant

not only for training autonomous cars to drive themselves on their own, but also for training autonomous drones to fly themselves on their own. And so there's a direct relationship between the ability to train AI systems and the types of military systems that defense ministries around the world are trying to acquire. And what about missiles? Well, it's the same thing. If you want to guide a missile accurately,

you've got to have a complex guidance system on board. You've got to have access to signals intelligence that lets you identify where your target is. And all of this requires a lot of data, a lot of processing power, and increasing amounts of intelligence. So pretty high stakes. There have long been fears, of course, China could invade Taiwan. As recently as last month, U.S. Secretary of State Ant Anthony Blinken warned China was accelerating on a path to gain leverage. I think China make a decision that it was no longer comfortable with the status quo. And we've seen them over the last few years ratchet up the pressure on Taiwan, trying to cut off its ties to countries around the world, international organizations. From our perspective, that status quo has worked and it's vital to what's important to us, which is maintaining peace and stability across the Taiwan Strait.

How feasible is it that China's military could storm the chipmaker in Taiwan and suddenly control global chip production? If they did, what would happen? Well, I think that's unlikely. If China were to attack Taiwan, it would almost certainly destroy the chip making capabilities of TSMC. So that makes it unlikely that China is going to attack with the aim of acquiring TSMC's capabilities. But if it were to attack, and it were to destroy TSMC, the impact on the global economy would be traumatic. We'd be facing the greatest disruption to global manufacturing since the Great Depression of the 1920s and 30s. It's not just smartphones, PCs, data centers that would be hugely disrupted,

but autos and dishwashers and coffee makers would face immense delays as well. So it would plunge the world economy into a deep depression. In the Dragon's Den, a simple pitch can lead to a life-changing connection. Watch new episodes of Dragon's Den free on CBC Gem. Brought to you in part by National Angel Capital Organization. Empowering Canada's entrepreneurs through angel investment and industry connections. Hi, it's Ramit Sethi here. You may have seen my money show on Netflix. I've been talking about money for 20 years.

I've talked to millions of people and I have some startling numbers to share with you. Did you know that of the people I speak to, 50% of them do not know their own household income? That's not a typo. 50%. That's because money is confusing. In my new book and podcast, Money for Couples, I help you and your partner create a financial vision together. To listen to this podcast, just search for Money for Couples. If the Chinese did attack Taiwan, why would this fab be destroyed, do you think? Well, a couple of reasons. First off, fabs are full of some of the most precise machines ever invented.

So they're really totally unprepared to survive a war zone. But suppose the building itself survived. The machines inside only operate because they have a steady flow of materials, spare parts, software updates from abroad. Because the machines inside of Taiwan's fabs are built in Japan, in the Netherlands, and in the United States primarily. Even if the machines somehow survived, you're still need the people to actually know how to run them. And a lot of the expertise is in the brains of the employees who operate these facilities. It's unlikely that they'd all stick around to run a facility under an occupation regime. And even if you assume that they did, you'd also need to assume that neither Taiwan nor the United States would destroy the FEBS.

And if it looked like they were falling into Chinese hands, that's certainly not something you could exclude. I'm assuming it's really hard to move, right? I guess I was just thinking when you were saying that, like, why not just move it somewhere else? But it's probably really hard to move, okay? Yeah, the machines alone. So one of the types of machines in a cutting-edge chip-making facility is called an extreme ultraviolet lithography machine, which is produced by a single firm in the Netherlands. And these machines take several 747 airplanes to move. So you're not going to move these in the midst of a crisis or a war.

move. So you're not going to move these in the midst of a crisis or a war. Is there another way that China could gain influence over Taiwan, which would get them more power over the chips, over this factory? Well, the strategy that the Chinese have been trying to deploy over the last decade is to build up their own capabilities at home by trying to acquire technology from abroad and domesticate its production. During its annual NPC meeting in 2021, President Xi Jinping pledged $1.4 trillion to accelerate their tech industry and become totally independent from foreign technology. Until quite recently, China was able legally to buy some of the most advanced tools,

use the most advanced software, legally purchase the most advanced chips, and then quite straightforwardly apply them to defense systems. But over the past couple of years, and especially in October of last year, the U.S., Japan, and other Netherlands have been tightening controls on technology transfer with the aim of preventing Chinese firms from catching up to the cutting edge. On the Netherlands part, it looks like this could involve widening restrictions on ASML, particularly to restrict their so-called deep ultraviolet lithography. So this is the equipment that is absolutely critical for when it comes to China creating and making its own advanced chips.

And Japan also expected to enforce similar sort of restrictions on Nikon, too. How has that been going? I guess I know the U.S. has also tried to push making more chips as well, right? That's right. They're two related but distinct processes. I think there's a broad consensus between Taiwan, Japan, the US, Europe, that none of those countries want China to reach the cutting edge in shipmaking. But then there's a second issue, which is if there's a crisis in the Taiwan Straits, if China blockades or attacks

Taiwan, the rest of the world wants a bit more insurance that in case that happens, they can access many of the chips that they need. And so the U.S. is trying to bolster its domestic manufacturing base for chips. Japan is doing the same thing. South Korea, Europe, all are taking similar policies. Today I'm signing the law, the Chips and Science Act, a once in a generation investment in America itself. The College of Commissioner has adopted today the European Chips Act. It will enable 15 billion euros in additional public and private investment till 2030.

And this is less about catching up to the cutting edge and more about making sure there's some spare capacity in case of a Chinese attack on Taiwan. Is it possible that China could get control over the leading plant in Taiwan, diplomatically without any military action? political leaders that support a stronger sense of Taiwanese autonomy and more Taiwanese defensive capabilities vis-a-vis China. So it's difficult to imagine ways that China could really succeed diplomatically asserting control. You talked about how other countries are making moves to kind of build up more supply.

Are they very far behind? I guess, how are they doing, the United States, Japan, and then even China domestically? Well, the amount of capacity to make chips in Taiwan is extraordinary, especially for cutting-edge processing chips. Taiwan is absolutely irreplaceable. So it's going to be a long process and involves a lot of investment for countries like the US, Japan, Europe, and others to begin to diversify the geographic footprint of chip making.

We're going to see a shift over the next couple of years, given all the investments that are underway, but it's going to be a slow moving process because building chip facilities takes a long time and is quite expensive. And so this isn't something where we should expect changes overnight. I guess listening to you, it's kind of alarming hearing just how much the balance of our world hangs on this super concentrated market, right? And I guess I'm wondering, like, did countries like the US and even China, to an extent, not see this coming, right? Like, did they not try hard enough to get bigger in the chip game in decades past? I think the Chinese government deserves some credit for seeing it coming. realized that in a decade's time, the world would be even more reliant on semiconductors and that

China would be reliant on importing them from Japan and Taiwan and Korea and elsewhere. And so has been trying to pour lots of money into its own ship industry to stem this reliance on imported ships. I think in the US and Japan and Europe and other countries, there's been a much more belated recognition. It's only in the past couple of years that people have come to realize the extraordinary reliance on chipmaking in Taiwan and the extent to which China's growing military power has really called into question the security of the Taiwan Strait. Why do you think that is, that it's only been the last couple years? I guess it just, it does seem more obvious than that to me, right?

in the last couple years. I guess it just it does seem more obvious than that to me, right? Yeah, in hindsight, it's obvious. But in practice, it took political leaders a long time, I think, for several reasons. You know, first is that the chip industry itself was well aware of the dependence on Taiwan. But for most of the chip industry, it was a status quo that worked very well, because TSMC is an extraordinary company, produces excellent technology at competitive prices, and has a track record second to none. So in the chip industry, where there was the most detailed knowledge of the concentration of the supply chain, most people were happy with the way it worked. The only problem was that as the military balance shifted and as concerns about China's intentions grew, that took a long time to filter into the chip industry and

the chip industry's expertise about how the supply chain function took a long time to be actually understood by governments. You know, five years ago, it was hard to find people in major governments who really understood anything about how the semiconductor supply chain worked and who were only vaguely aware of Taiwan's importance. Now that's beginning to change. But it's only the past couple of years where that expertise and knowledge has really begun filtering into political leaders more broadly. And just for a final question, I want to return to where we started with this comparison that chips have become like oil. Of course, in the last century, we saw countries rise, diplomacy reverse course, wars start with oil. And with, in the last century, we saw countries rise, diplomacy reverse course,

wars start with oil. And with that in mind, how do you expect the push for access to microchips to shape the world in coming decades? Well, I think we're going to see an ongoing focus from major governments in terms of thinking about where the semiconductor supply chain is located and who has access to it. I think the politicization of the chip industry that we've seen from both China and the United States in particular over the past couple of years is going to be an ongoing feature of the chip industry as both countries try to increase their own capabilities and gain leverage vis-a-vis the other. And this is something new for the chip industry because for most of the last several decades,

they didn't have to think very hard about politics. The supply chain was international, technology crossed borders, and it wasn't really seen to be militarily relevant. But now the security-focused parts of the government are focusing on the extent to which advanced chips are critical for military systems, and they're using their power to begin to reshape how supply chains are structured. Okay. Chris, thank you so much for this. This was really, really interesting. I

learned a lot today. Thank you. Thanks for having me. All right. That's all for today. I'm Jamie Poisson. Thanks so much for listening. Talk to you tomorrow. For more CBC Podcasts, go to cbc.ca slash podcasts.