Semiconductors: The Geopolitics of The New Oil
In 1990, more than three-quarters of the world’s semiconductors were manufactured in the United States and Europe. Today, those locations manufacture less than a quarter. Semiconductor manufacturing has mostly moved to Japan, South Korea, Taiwan and China. No one seemed too concerned with this for quite a while, but factory shutdowns, changes in demand and problems with global supply chains led to a shortage during the pandemic.
The chip shortage disrupted all sorts of other supply chains and raised awareness of how important semiconductors are in almost all modern manufactured goods. The semiconductor shortage was seen as possibly the biggest economic security issue of the pandemic, bringing into clear focus how reliant the global economy is on a handful of microchip manufacturers in East Asia. The recent sanctions on Russia, which have involved cutting off Russian access to western technology and trade disputes with China have further highlighted how chips are increasingly being viewed by governments around the world as a national-security priority. It’s not just the key role they play in consumer products that makes them important, it’s their uses in military weapons and cyberwarfare.
On Monday alone, shares in top Chinese chipmakers lost almost 9 billion dollars in market value after the United States announced new export controls on Friday that restrict the sale of certain semiconductors made with US technology and semiconductor manufacturing tools unless vendors first obtain an export license. These restrictions are designed to curb China’s plans for technological self-sufficiency, and their use of American technology in manufacturing advanced weapons. During the pandemic chip shortage, no company came under greater scrutiny than Taiwan Semiconductor Manufacturing Company (or TSMC), which not only produces around 90 per cent of the worlds most advanced chips but they manufacture these chips in a country – Taiwan - that has increasingly been threatened with invasion by China.
The American decision to rule out military intervention in Ukraine, combined with the US President’s statement that if China were to attack Taiwan, the US would come to Taiwan’s military defense, might tell you something about the priorities of the US government. Semiconductors are of great importance to the US economy, and all economies, and… it might be reasonable to think of computer chips as the new oil, a product that has become a necessity and of great strategic importance to governments around the world. So, the United States has placed a number of restrictions on China’s tech industry in recent years, including blacklisting Huawei Technologies and preventing some Chinese chipmakers from buying American manufacturing equipment without a license. Last week the US ordered
Nvidia to stop selling two of its top computing chips to companies in China. Nvidia warned that the new licensing requirements could cost it as much as $400mn in lost sales for the current quarter alone. Then on Friday, the Biden administration introduced even more sweeping export controls to severely restrict efforts by Chinese companies to develop cutting-edge technologies with military applications. The controls can be expected to hit Chinese companies in a number of ways. They’ll prevent US companies from exporting critical chip manufacturing tools to China and prohibit US citizens and companies from providing any kind of direct or indirect support for semiconductor fabrication plants in China. The US also put YMTC (a Chinese state-owned semiconductor manufacturer) along with 30 other Chinese entities — on a list of “unverified” companies, possibly paving the way for inclusion on a separate blacklist called the “entity list” that would prohibit US companies from supplying them with any technology. The Biden administration is using the “foreign
direct product rule” (a mechanism that was first used by the Trump administration against Huawei). This rule effectively prohibits any US or non-US company from supplying targeted entities with hardware or software that contains, or has been manufactured using, American technology. This makes it extremely difficult for a foreign firm to develop or maintain supercomputers or AI technology. To reduce the supply chain impact on American companies, there is a carve out for chipmaking facilities in China owned by companies from the United States or allied countries that are exporting chips. It is not just the US government who are concerned with the concentration of semiconductor manufacturing abroad either. Newport Wafer Fab runs Britain's largest chipmaking facility. It was sold in 2021 to a Dutch company, which is a subsidiary of a Chinese company Wingtech Technology.
After the sale the plant began selling all of its output exclusively to its new owner, stoking fears of a tech transfer from the UK to China. The UK’s Department for Business, Energy & Industrial Strategy launched a national security assessment of this sale earlier this year, using powers granted under the new National Security and Investment Act to examine whether the agreement should be reversed or allowed to stand. The EU this year approved the 43 billion euro “European Chips Act”, so that they can subsidize chip manufacturing in the region. Their stated goal is to reinforce semiconductor supply chains and boost the EU’s global chipmaking market share from less than 10 per cent to 20 per cent by 2030. France has additionally announced that it will separately invest billions in a new semiconductor plant.
Then in July, the US passed the Chips and Science Act that earmarked $52bn in grants to support advanced chipmaking in the United States. So, as you can see, things are really heating up in this space. But, why did so much semiconductor manufacturing move to Asia over the last three decades to begin with? Before we get into that, let me quickly tell you about today’s video sponsor Morning Brew. As I’m sure you can imagine, I’m a big consumer of financial news, and Morning Brew is an amazing way to get up to speed with what is going on first thing every morning. Instead of aimlessly browsing social media, sign up for morning brew. It's a totally free
daily newsletter, delivered seven days a week. They get you up to speed on business, finance, and tech news in around five minutes each morning. It's become one of my favorite news sources. They had a great piece over the weekend on this very topic which inspired today’s video. Sign up using the link in the description below. Now, it’s not so long ago that the United States and Europe, entirely dominated the semiconductor manufacturing sector. And, semiconductor manufacturing is quite different to a lot of other types of manufacturing where expensive unionized workers were replaced with lower cost foreign workers in pursuit of greater efficiency. The main reason chip production
moved abroad was because governments in Asia offered huge financial incentives for factory construction, with the goal of building up this specific industry. Chip companies also benefited from the large networks of suppliers in Asia, who have less regulation to deal with than in Europe and the States. Today they are still attracted to manufacturing in Asia because of the growing workforce of skilled engineers in that region who are capable of operating the expensive and highly technical manufacturing machinery used at these factories. I should note that many of the world’s largest chip companies are still American though… and in fact American semiconductor companies make up almost half of the overall market. Intel still does much of its manufacturing in the United States, Ireland, and Israel.
Nvidia designs its own chips in the United States but then has them manufactured by partners like TSMC in Taiwan and Samsung at facilities in South Korea and China. The United States still has a very large market share in chip design and chip design software, but the raw materials that go into chip-making, including industrial chemicals and silicon crystals, largely come from outside the U.S. Manufacturing in Asia made (and makes) a lot of sense for semiconductor companies who benefit from significant government handouts, lower taxes, cheaper labor and less regulation than they would face in the United States or Europe. The offshoring of high-tech production has been much more of an issue in computing hardware and consumer electronics, than other high-tech manufacturing sectors like aerospace and pharmaceuticals, as Asian countries have made more of a push to attract semiconductor manufacturers than these other businesses. Morris Chang, the founder of TSMC grew up
in Hong Kong, he studied at Harvard and worked for Texas Instruments in the United States. After 25 years with Texas Instruments, he felt his career was no longer progressing and the premier of Taiwan offered him a role, running the Industrial Technology Research Institute. A government organization charged with turning Taiwan into a tech powerhouse. Morris realized that there weren’t enough chip designers in Taiwan, to build the kind of company that the government hoped for, but he also realized that the United States had the opposite problem. They had lots of smart chip designers, but it was too expensive to build semiconductor factories in the United States. He built a business that did no design
work, and instead served American customers, manufacturing their designs for them. And that business is TSMC. It is neither easy nor cheap to manufacture semiconductors – the factory has to be built with a foundation that reaches down to the bedrock of the earth in order to minimize vibration down to just nanometers. The cleanrooms at the facilities have to be dust free; a thousand times cleaner than a hospital operating room. These facilities cost between fifteen and thirty billion dollars to construct.
Now, solar cells, much like computer chips were also invented in the United States and Americans still hold many of the global patents on solar technology. The most common solar panels sold today are the same crystalline silicon panels that Bell Labs developed in the 1950s. The US led the industry for decades, but when the Chinese government began heavily subsidizing the industry, US manufacturers were unable to compete. Chinese manufacturers, benefited from government subsidies, cheap coal derived energy that is used to make the cells, and are accused of using forced labor in their factories.
Chinese companies are accused of selling solar panels in the United States and Europe for less than it costs to manufacture and ship them, which is known as “dumping.” Both the U.S. and EU governments in 2012 and 2013 confirmed that the Chinese government was subsidizing the dumping of solar panels. So is the correct answer then, to slow down Chinese access to technology and to subsidize manufacturing in the west, so that the US and Europe can maintain a foothold in the semiconductor industry. Obviously if you believe in free trade as I do, it starts to look like quite a problem if one country is going to subsidize an industry so that it bankrupts competing businesses around the world. Kinda
like what happened with the solar industry. Well, it is a big problem for free trade economies, but this strategy is not a big win for China either. The government subsidies being provided end up being unsustainable. Over time it becomes too expensive to subsidize such fast-growing industries. The state interference in markets through subsidies is another example of the
malinvestment problems that have been happening in China over the last twenty years. You do see high GDP growth when this is done, because all investment will boost GDP, but if the overall effect is that the country is selling a dollar for 90 cents, eventually these losses catch up on the country as a whole. We are starting to see this happen right now. A strategy like this is bad for everyone, as unsubsidized businesses in other parts of the world can go out of business due to being unable to compete. This approach throws world trade entirely out of balance, and we talked a little bit about this in my recent video on the Japanese bubble economy of the 1980’s which Japan is still recovering from today. When we look at the level of support that
China provides to the semiconductor industry and compare it to the United States, it’s clear why so much semiconductor manufacturing has drifted abroad. Often the land for a factory in China is provided for free, the factory construction and manufacturing equipment are partially paid for too. On top of that there are significant corporate tax reductions given to these factories. In the United States, government help has mostly been made up of State Tax and property tax reductions for businesses, and this has mostly been done to attract an employer to a given region of the United States, rather than to the country as a whole. So, does this then mean that the only solution is to fight subsidies with subsidies and tariffs and export bans? Well, that is one approach, but doing that has its problems too. It becomes a bit of a race to the bottom. If every government
focuses on subsidizing “chosen” businesses, business becomes organized just to collect subsidies. And that’s not the way things are meant to work. When the government subsidizes businesses, it weakens profit-and-loss signals in the economy and undermines market-based entrepreneurship. Now, I can’t tell you what the right thing to do is from a military or strategic point of view, as I have no expertise in that space, but, Mark Liu, the Chairman of TSMC recently told CNN that an invasion of Taiwan would likely render the TSMC factory “inoperable.” According to Liu, the manufacturing facility is so sophisticated - that it very much depends on its connection with the outside world, with Europe, with Japan, and with the United States for materials, chemicals, spare parts, engineering software and diagnosis. He says “Nobody can control TSMC by force. We also must ask, if blocking Chinese access to advanced semiconductors would even work. Some U.S. government officials believe that the US economy and national security is threatened
if advanced chips end up with the Chinese military and their defense industry. But the belief in export controls rests on two assumptions: One. that these measures will actually prevent targeted entities from getting access to these technologies, and two; that the cost paid by the U.S. will be acceptable. Preventing access to advanced chips might
just drive China to develop their own alternatives, and eventually have companies making chips that outcompete western companies. On top of that, there are not many historical examples of countries being able to keep new technologies from leaking abroad for more than a short period of time. It took the Soviets only a few years to get access to American technology and develop their own nuclear weapons once they saw that America was a nuclear power. While there are trade disputes with China, over things like dumping. China is not an enemy country, and in fact, China is Americas largest trade partner, and both countries rely on each other and derive mutual benefit from trade.
As I said though, I have no expertise in military strategy or anything like that, and I’ll be interested to read your comments on that topic in the comments section below. Now, If American companies are (in the long run) prevented from selling their products into the huge market that is China. In the long run, these companies can reasonably be expected to be less profitable. That would mean that they end up with less money to spend
on R&D meaning that they should be less innovative in the future, which is bad for everyone in society who might benefit from innovation. It’s additionally worth noting that while quite a lot of semiconductor manufacturing has moved from the western countries to Asia, The US still accounts for 39 percent of global value‐added within the industry. This is the largest share of any country. So, while a lot of the work of manufacturing is being done abroad, the highest value work is still being done in the United States. Things aren’t nearly as gloomy as you might at first think. OK so, what about the risk of westerners (even
temporarily) losing access to their own technologies in the event of a war or further supply chain disruptions? Well, the semiconductor manufacturers themselves are the most hurt when supply chain issues reduce sales. They’ve been made painfully aware over the last year or two that it might not make sense to do all of your manufacturing in one geographic region. It’s reasonable to expect businesses to manage these risks internally, and that is exactly what we have seen. Intel, TSMC and Samsung had all announced investments in U.S. facilities worth almost 70 billion dollars (so more than the Chips act) and had begun construction before the CHIPS Act was passed by Congress. They didn’t do this to collect subsidies, they made the
investment because it makes good business sense. If we are going to talk about using subsidies to compete with foreign subsidies, it’s worth asking the question, should the US and Europe be mimicking the top-down authoritarian approach taken by China, exactly at the point when we are seeing it fail in China? The Chinese approach where the government decides which sectors receive investment has led to all sorts of market distortions in China. This is a big part of their malinvestment problem. The most obvious example being the huge real estate bubble which is in the early stages of its collapse right now. We have seen Xi’s policies wipe two trillion dollars off the value of Chinese tech stocks in the last few years, and last month it was announced that for the first time since 1990 the Chinese economy is expected to grow slower than the rest of Asia. Whenever U.S. policymakers have experimented
with industrial policies in the past, the firms that are supposed to benefit, are rarely the most competitive or innovative once receiving government support. The problem is, that it becomes more profitable for these firms to focus on maximizing the government handouts that they can collect rather than worrying about marketplace demand for the goods they produce. I joked in my Japan video from a few weeks ago about how Ronald Regan used tariffs to protect Harley Davidson in the 1980’s. I’m not sure that much innovation
came out of Harley Davidson during that five-year period of being protected from foreign competition. The western approach, which can look messy at times, allows the marketplace to determine winners and losers rather than government officials. Government policymakers have no special insights that allow them to allocate capital more efficiently than individuals or businesses. While businesses and investors do make lots of mistakes, they also bear the consequences of those mistakes and because of this, they are more willing to make a u turn when they see they are going down the wrong path. When the government subsidizes a losing idea, the costs can grow and grow, and are involuntarily borne by taxpayers.
Even with the best of intentions, government officials don’t have the right incentives in place to manage taxpayers’ money sensibly. They are not rewarded when they maximize value; nor are they punished when they take unnecessary risks or run into huge cost overruns. Businesspeople acting in markets use price signals to guide their decisions. They provide the goods and services that they can see demand for. When a private company fails (which they often do), the owners and investors lose, but they at least then stop throwing good money after bad and learn a valuable lesson. Government decision makers have no way of accounting for either the value derived from their good decisions, or the costs associated with bad decisions and (worst of all) are even incentivized to claim losses as wins in order to win the next election. These losses are of course wasteful, hold a country back,
and cost taxpayers’ money. When a government decides to help one business or industry, this puts other businesses and industries at a disadvantage. Often the biggest businesses that are best at interacting with government programs and officials do the best out of government intervention, and smaller, but possibly more innovative incumbents suffer. These market distortions not only generate losses in the economy, but they are the type of losses that are not easily seen and thus aren’t usually considered by policymakers.
In a market with heavy government interference, young companies and entrepreneurs can have a harder time raising capital because private investors would often rather invest in a subsidized rather than an unsubsidized project so that they can collect the free government money. It won’t show up in any government analysis after the fact that these small innovative companies rotted on the vine. Government subsidies can also create an unhealthy—and often corrupt—relationship between business and the state – what is known as crony capitalism. When government officials are handing out subsidies, businesses are incentivized to hire lobbyists to make the most of the handouts. Over time more and more economic decisions are made on the basis of politics and influence rather than good business sense, and more resources are misallocated. Over time, the economy moves away from providing goods and
services that are in demand - to an economy characterized by cronyism and corruption. Policymakers who believe that entrepreneurs are underinvesting in new technologies could provide more help by reducing red tape, making improvements to the overall business environment, and reducing general taxes which would help all businesses equally, rather than taxing everyone and subsidizing chosen firms, industries, or technologies. Governments could provide more help to industry by ensuring that good trade agreements are in place, which provide businesses access to global markets and things like expanding high skilled immigration so that businesses have access to a talented workforce. Many of the policymakers who argue that the private sector is refusing to invest and thus the government needs to fill this investment gap are the same politicians who then want to increase taxes and reduce the rewards of investment. You don’t get to complain about not enough investment if you are putting policies in place that disincentivize the very investment you are hoping to see. Up until recently, the top-down approach taken
in other parts of the world looked like it was outcompeting the laissez faire – or free markets approach followed in the west. It is becoming more and more obvious over time that the authoritarian leaders have been pretty busy wallpapering over the cracks in the walls. When a government starts subsidizing businesses, or picking winners and losers, it weakens the important profit-and-loss signals in the economy and undermines market-based entrepreneurship. Most western technological and industrial advances (things like inventing (and mass marketing) computer chips and solar panels) came from businesses taking measured risk to develop a product that they felt there would be a large market for. The Soviets tried to develop integrated circuits (or computer chips) at the same time as the Americans in the 1950’s, but the American businesses benefited from the ability to trade internationally buying in the supplies that they needed from anywhere in the world (the Soviets had to use only home grown tech). More importantly, the Americans benefitted from operating in a vibrant economy. It was much more efficient to develop a computer chip that might have
military applications, but also could be included in consumer electronics and toys, as the cost of development in the west was spread amongst all of the possible uses of the new technology, while it would have been a purely military technology in the USSR. If you want to read more of this history, check out the book “Chip Wars” by Chris Miller that I have linked to in the description. Western long-term economic growth has come mostly from entrepreneurs creating new businesses and pioneering new industries and these businesses often had to overcome barriers put in place by governments and dominant businesses receiving special treatment. I would argue that western governments fighting subsidies with subsidies is simply a race to the bottom. This approach would reduce efficiency, innovation and destroy a vibrant business environment by restricting the invisible hand that allocates resources better than any other approach that we have seen in history. Innovation has always been the key to western success. You don’t win a race by trying
to trip your opponent up, especially when there are significant costs to doing so. A far better strategy is to run faster. If you found this video interesting, you should watch my video on the Japanese bubble economy of the 1980’s next. I mentioned it a few times during this piece. Don’t forget to sign up for morning brew using the link in the description below it's totally free so there's no reason not to try it out. Have a great day and talk to you again soon. Bye.