Our Tenuous – and Crucial – Microchip Dependency
No piece of technology is more crucial than the microchip. Its supply was central to the cause of the post-COVID-19 inflation, and the stability of the U.S.-China relationship hinges on its manufacture.
Chris Miller’s must-read book “Chip War,” winner of the Financial Times business book of the year award, could not have been published at a more opportune time. At its publication date in October 2022, the world was emerging from a crippling COVID-19 pandemic. The pandemic illustrated crucial linkages that the complacent globalized world of commerce had blithely came to depend on, but that events revealed were unreliable.
Specialization in the hyperefficient supply chain created two vulnerabilities that were barely recognized before COVID-19, but that became central during the pandemic. Almost all personal protective equipment (PPE), including rubber gloves and surgical masks, were made in China. When demand for them suddenly ballooned when China was under lockdown, there were severe shortages. The shortages led to brutal competition for PPE, between countries and even states of the United States and the federal government.
Second, 60% of all microchips and 90% of the most advanced ones are made in Taiwan, almost all of them by a single Taiwanese company, TSMC. And the incredibly expensive and technologically advanced machinery that TSMC needs to fabricate those microchips is made by a single Dutch company, ASML.
Used car inflation
How does this affect us? Let’s take just one example. As the COVID-19 pandemic wound down in 2022, the price of used cars skyrocketed, up 37% between February 2021 and August 2022.
Why did this happen? It was because of a microchip shortage. Reading the newspaper articles about this in depth revealed something that few of us had realized – that a new automobile requires between 1,000 and 1,500 microchips. (An electric car requires twice as many.)
But microchips were in short supply. Miller says it was not, as one might expect, because of supply chain bottlenecks due to lockdowns in China and other countries that supplied components for microchips. In fact, he says, “There were few industries that sailed through the pandemic with so little disruption.” Instead, the shortage of microchips, Miller says, was “mostly a story of demand growth rather than supply issues.”
During the pandemic, people were housebound, causing an increase in demand for microchips for household digital devices. And the generous fiscal support checks, together with fewer spending opportunities because movement was constricted meant that people had money to spend on computer equipment.
Many adults and schoolchildren could only work online. Twenty-nine percent of broadband users did something to improve their home internet connection, which meant the purchase of additional digital equipment. In 2020, shipments of mobile PCs to America’s schools were up 28% over 2019, paid for by the $190 billion in federal subsidies that went to elementary and secondary education. All of this meant more demand for microchips.
Furthermore, in addition to heightened demand, the shortage of chips for automobiles was, Miller says, “mostly the fault of carmakers’ frantic and ill-advised cancelation of chip orders in the early days of the pandemic coupled with their just-in-time manufacturing practices that provide little margin of error.”
The result was that there was a shortage of new automobiles and a long wait to obtain one. But people were just emerging from lockdowns in 2022 and had money saved up – the US saving rate increased from 8% in 2019 to 32% in 2020 – so they wanted to drive somewhere. Many wanted or needed a new car. But they couldn’t get one because of the shortage and the wait, so they opted for a used car instead, driving used car prices sky-high.
The chipmaking industry
The most fascinating thing we learn from Miller’s book is how the chip industry works. We learn how chips are made and who makes them. There is a reason why Gordon Moore’s 1965 prediction has panned out, that the number of semiconductors that could be crowded onto a microchip would double every year. It is because of an industry that is constantly making incremental improvements. It is one of the most technologically advanced industries on earth.
Exponential growth is a powerful thing. Doubling every year since 1965 means multiplication by more than a hundred thousand trillion, or a hundred quadrillion by 2023. That’s a lot of increase in the number of components that can be fitted onto a microchip.
Sometimes when I give a talk about how we have used technology to do more with less, thereby helping to reduce impacts on the environment while our civilization and economies grow, I take out my little 128 gigabyte flash drive and hold it up. It is about a third the size of the smallest joint of my little finger.
I learned to program computers in 1965. The computer I used took up an entire air-conditioned floor of a large building.
I then tell the audience that this little 128 GB flash drive, smaller than the small joint of my little finger, holds a thousand times more data than did all the computers in the world when I first learned to program computers. (For all I know it may be a million times, or a billion.)
How has this been possible? I didn’t know until I read Chip Wars.
The manufacture of a microchip is simple in principle. A circuit is etched in tiny channels on a silicon chip. Those channels are then filled with semiconductor materials to form a controllable circuit.
The whole key to cramming more components onto a chip is to make those channels thinner and thinner and closer and closer together. The Dutch company ASML has perfected that process to a fare-thee-well. Yet every year it gets better and better. The latest version uses very small wavelength “extreme ultraviolet light” to etch the channels. This EUV light, says Miller, can only be created by pulverizing a small ball of tin with a laser. ASML’s machine to do the etching is enormous and highly specialized, combining technologies and components in one machine from multiple high-end engineering firms around the world. One ASML lithography machine can cost over $100 million.
Almost all these machines are sold to TSMC in Taiwan.
The fabless and the fab
There are basically two steps to creating a microchip. The chip’s circuitry has to be designed to perform its intended purpose. Then, the chip needs to be fabricated – that is, physically manufactured.
Over time, most companies that design chips discovered that it was easier to go “fabless” – that is, not to manufacture the chips themselves. About 55 years ago, with very substantial funding and support from the Taiwanese government, a man named Morris Chang started a company, TSMC, to do that – and only that. TSMC makes chips to its clients’ design specifications, but it does not design chips itself.
Efficiency, as Adam Smith pointed out centuries ago, improves with specialization. Thus, a hyper-efficient global supply chain tends toward a high degree of specialization.
Few industries are as specialized as the microchip industry. Most designers of microchips are fabless. TSMC is the fab that serves them all.
Does this contribute to geopolitical tensions?
At around the same time that we learned that almost the whole microchip industry depends on one company in Taiwan, and that we are far more dependent on microchips for everything than we had realized, suddenly, concern in the U.S. over Taiwan’s political status has increased enormously.
This is worrisome. Many geopolitical realists will say that the real reason the United States invaded Iraq was because of oil. A friend of mine says that it would have been different if Iraq raised broccoli – and he is surely right.
Could it be that the United States has become more militant in its rhetoric about Taiwan because Taiwan’s microchip fabrication industry – specifically TSMC’s – is so crucial to all our technology, including military technology? This is not usually cited as the reason for our concern over Taiwan. But oil was not cited in political rhetoric as the reason for invading Iraq in 2003 either.
China is the new Japan
Professor Miller (he is an assistant professor of international history at the Fletcher School of Law and Diplomacy at Tufts University) reminds us of something we may have forgotten, or never knew if our memories don’t stretch back that far. In the 1980s, the United States was in a panic over losing technological superiority to Japan. And it thought it was not a fair fight. Silicon Valley thought Japan’s chip firms, as Miller says, “benefitted from intellectual property theft, protected markets, government subsidies, and cheap capital.” Numerous books were written at the time lodging those accusations against Japan and warning, sometimes very shrilly, of the danger.
Sound familiar? The United States now accuses China of the same things.
There was much truth to the accusations that Japan stole intellectual property – as there is that China did. But some of the other accusations were hypocritical. For example, while Japan did benefit from government support, so did the development of the United States’ chip industry – massively. The industry never could have gotten off the ground if it hadn’t been for big contracts in the 1960s from the National Aeronautics and Space Administration (NASA) and the Defense Department.
In the time of paranoia about Japan in the 1980s, there was even concern that it could become a military threat, despite the fact that Japan’s constitution, imposed on it by the United States after the Second World War, renounces military force as a means to settle international disputes. Says Miller, “If Japan’s success in DRAM chips was any guide, it was set to overtake the United States in almost every industry that mattered. Why wouldn’t it seek military dominance, too? If so, what would the U.S. do?”
Once again, while China’s constitution does not similarly renounce military force, the record shows that China, like Japan – and unlike the United States – has not used military force to settle international disputes, at least not for more than 40 years. But we are paranoid about it anyway.
For now, and let us hope for the foreseeable future, both the United States and China are maintaining a posture of “strategic ambiguity” with regard to the status of Taiwan. China – the People’s Republic of China, i.e., mainland China – claims Taiwan as part of itself, and most countries of the world (including the United States) officially agree with that. China reserves the right to incorporate Taiwan militarily if necessary.
The United States, meanwhile, implies – ambiguously, and not officially – that it would back Taiwan’s independence if China does invade.
But neither of them is going to do anything of the sort, unless crazy political elements take hold of power in either place. And the reason for that is in part – perhaps in large part – because of TSMC.
Miller says that it “has [n]ever been realistic that Chinese forces could invade and straightforwardly seize TSMC’s facilities. They’d soon discover that crucial materials and software updates for irreplaceable tools must be acquired from the U.S., Japan, and other countries.”
And furthermore, “It would be even worse if a war knocked out TSMC’s fabs. The world economy and the supply chains that crisscross Asia and the Taiwan Strait are predicated on this precarious peace.”
Miller’s book should impress on us just how fundamentally essential the chipmaking industry is. And so much of it depends on just one company in Taiwan and one in the Netherlands.
Both China and the United States are trying to domesticate the advanced technologies that TSMC and ASML have developed to such a high degree. Miller’s book helps us understand why it will not be easy.
Cryptocurrency hyper-enthusiasts and AI hyper-pessimists take note
Two of the nutty things going on these days are extremes of enthusiasm about bitcoin and other cryptocurrencies, and extremes of pessimism about artificial intelligence.
The bitcoin enthusiasts will tell you that making bitcoin the world’s currency will free us from dependence on sinister centralized institutions, such as those of big governments and big corporations.
The AI hyper-pessimists will tell you that AI is going to kill us all.
But both are forgetting one big fact; neither of these technologies is possible without the microchip. And as we have seen, microchip making is about as centralized, and as much a product of big corporations and big governments as you can get.
And microchip manufacture requires energy and materials – enormous quantities of both.
As Vaclav Smil has reminded us, we do not live in a digital world; we live in a material world. And at the pinnacle of the pyramid of that material world lies the microchip.
Economist and mathematician Michael Edesess is adjunct associate professor and visiting faculty at the Hong Kong University of Science and Technology, managing partner and special advisor at M1K LLC. In 2007, he authored a book about the investment services industry titled The Big Investment Lie, published by Berrett-Koehler. His new book, The Three Simple Rules of Investing, co-authored with Kwok L. Tsui, Carol Fabbri and George Peacock, was published by Berrett-Koehler in June 2014.