Dan Wang on the US–China Race: Two Superpowers "Very Intent on Losing"
EO Korea
Dan Wang is the author of Breakneck: China's Quest to Engineer the Future and a research fellow at Stanford's Hoover Institution. He spent six years in China as a technology analyst at the economic research firm Gavekal Dragonomics, where he studied China through the lens of political economy, particularly its ambitions in technology. In this interview he addresses a question that many in Washington and Silicon Valley think they have already answered: who is going to win the US–China competition?
His answer is that, right now, neither country looks like a clear winner, because both "are very intent on losing." He argues that the US and China are each hurting themselves, and that the advantage will go to whichever one stops making so many mistakes.
The Physical Gap: Ships, Cars, Solar, and Nuclear
Wang opens with comparisons meant to show how far China has pulled ahead in building physical things. By his account, the United States built around five ships last year, while China built about 1,500. An American automaker takes on average five or six years to go from conceiving a new model to putting it on the street. In China, he says, the iteration cycle is closer to 18 months, roughly three times faster.
Energy is the comparison he returns to most often. According to Wang, China built about 300 gigawatts of solar power last year alone, compared with about 30 gigawatts in the US. China currently has 40 nuclear power plants under construction, and the United States has zero. More broadly, he says China is building far more coal, nuclear, solar, and wind capacity than the US.
Silicon Valley's "Magical Thinking" About AI
On artificial intelligence, Wang grants that the US is ahead. He says the only real question is whether the lead is very substantial or only moderate, and his view is that it is "at best, a moderate lead," not a decisive one.
He is skeptical of a line of reasoning he hears in both Silicon Valley and Washington. In this view, somewhat better AI could deliver a decisive strategic advantage because the US will invent superintelligence and China will not. Wang calls this "a category of magical thinking." He says Silicon Valley is "very capable of magical thinking," and that this translates poorly to Washington, which has its own magical thinking about how to outcompete China. The flaw, in his view, is the assumption that China cannot catch up quickly.
He supports this with what he has seen from Chinese AI companies. They have produced many excellent models despite US export controls that, by his description, have targeted some of China's leading firms for close to a decade. He says these models are not substantially behind American alternatives, that they are often open source, and that they appear cheaper to run. In some areas, such as video generation, he suggests Chinese models may even be ahead. Looking beyond who has the single best model at the moment, he says China does not appear to be substantially behind at all.
He also argues that what matters most over the next few years, as AI is actually deployed, is electrical power, the area where he sees China vastly outbuilding the United States.
China's Manufacturing Backbone
Wang describes China as home to some of the most advanced factories in the world, including what are called "dark factories." These are so thoroughly automated that they need hardly any human workers, so most of the lights can stay off. He stresses, though, that these showcase facilities are not what matters most.
In his view, the backbone of Chinese manufacturing is the much broader mass of factories that make iPhones, consumer electronics, electric vehicles, and their batteries. In these products he believes the Chinese are far ahead of the Americans. What makes this possible, he says, is a set of dense production ecosystems. For almost any component, whether batteries or sensors, the suppliers sit right next to one another, so a factory manager can literally walk next door to get another part.
Alongside this is a deep labor pool. Wang cites about 70 million manufacturing workers in China, ranging from mass labor to medium-skilled workers to highly skilled engineers, whom he calls "probably the most skilled engineers in the world today." He finds these disparities "a little bit scary."
Taken together, this leads him to question how much America's lead in model sophistication is really worth. He acknowledges the US builds more sophisticated models by every measure. Still, he says he is not sure he would trade a much more vibrant industrial manufacturing ecosystem for that edge. He calls the Chinese economy incredibly competitive and says that for anyone selling products, China is the most competitive economic system in the world.
A Country of Distinct Regions
Wang also cautions against treating China as a monolith. A country of about 1.4 billion people is especially complex, and after years of living there he came to appreciate how different its main hubs feel.
Beijing, he says, was jokingly called "Western Pyongyang" by people in Shanghai because it is so centralized and Stalinist, defined by the political system, with concrete buildings and dust storms. Shanghai is much more of a commercial hub. Residents call it "the Paris of the East," and it has French-designed leafy boulevards, café-lined streets, and better weather. The Southwest is defined by its mountains, strong regional character, and "amazing ethnic diversity." He compares this to the United States, where the San Francisco Bay Area, New York City, Texas, and Washington, D.C. each feel very different from one another.
What the Engineering State Got Right
A central idea of Breakneck is that China is run by engineers. Wang says that, for the most part, China's engineer-heavy leadership has done a good job of promoting "physical dynamism." Someone who grew up in a Chinese city twenty years ago might have had no subway lines, and today that same city may have a couple dozen. China built its first high-speed rail system about 15 years ago, and by his account it now has about twice as many high-speed rail lines as the rest of the world combined. He counts tall bridges connecting villages, subways, rail, electrical power, and the manufacturing base among the genuine achievements of the engineering state.
The Dark Side: Engineering People
Wang argues that the same mindset has a fundamental problem: engineers in charge of a country "cannot stop themselves" from social engineering and population engineering. They treat the population as another building material to be torn down and remolded as they wish.
His main example is the one-child policy. He says China's leadership believed the Western "doomerism" of the 1970s, the idea that mathematical formulas could determine an optimal population, and that countries had to reduce their populations as much as possible to stay within environmental carrying capacity. Wang calls those calculations "bunk," but says China took them too seriously. A missile scientist named Song Jian gained the ear of the leadership and supplied mathematical justification for the policy. Over roughly 35 years, according to Wang, the policy involved around 300 million abortions and 100 million sterilizations of women. He describes it as "a campaign of rural terror" that ran through the countryside, and as the result of a leadership that does not respect its people.
The Anaconda in the Chandelier
Wang also describes China's information controls from personal experience. In the spring of 2022, his personal website, danwang.co, where he publishes his annual letters, was blocked in China. He calls it a very small casualty of that year, but says he was surprised that his "rinky-dink website," nowhere near the size of Facebook, Wikipedia, or the Wall Street Journal, was blocked too. He describes an environment where a WeChat message sent to someone else in China might never arrive because censors had already blocked it, and where the state has largely destroyed independent journalism.
To explain how censorship works psychologically, he borrows a metaphor from Professor Perry Link: an anaconda coiled in a chandelier. Imagine everyone sitting around a dinner table with a giant anaconda hanging above them. No one needs to see it strike. Knowing it is there, and that it might or might not come down, is enough to intimidate people. Because no one knows when the anaconda might descend, people censor themselves automatically, and Wang sees that attitude as corrosive to creative thinking in China.
As a writer, he says, he has been determined not to be intimidated. He tries to tell the truth as best he can, describing the spectacular achievements in manufacturing and public works while also acknowledging what he saw as grave mistakes by the Communist Party, such as the zero-COVID program. He believes it is important to see both at once and not let censorship shape one's thinking too much.
A Model of Competition: Hubris Ahead, the Whip Behind
Wang's framework for the rivalry is dynamic. Whichever country is ahead tends to make mistakes out of overconfidence and hubris. Whichever is behind feels "the crack of the whip" and works hard to catch up. He is therefore suspicious of any argument that China must win because of manufacturing, or must lose because of demographics. Both countries have real strengths, and both will keep trying to run faster.
He gives an example from each side. In 2021, while Wang was living in China, Xi Jinping ordered what Wang calls a controlled demolition of the property sector and cracked down on tech entrepreneurs, most notably Jack Ma. Wang attributes this to Xi feeling that China was in a very strong position after handling COVID well. He argues that the property slowdown and the crackdown on entrepreneurs are behind many of China's current economic problems.
The US, in turn, felt ahead for a while, and Wang says it responded under President Trump by eroding alliances through tariffs and, in his words, deindustrializing America. He says the US lost about 80,000 manufacturing jobs under Trump's first term. The example he calls one of the most hurtful came last year, when the Department of Homeland Security deported around 300 South Korean engineers working at a Hyundai plant in Georgia that was building EV batteries. Beyond the experience for the engineers themselves, Wang says the episode looked bad to the rest of the world. An engineer in France, Taiwan, or South Korea might now hesitate before taking a job in Georgia, Alabama, or Texas.
"Do Less, Please"
From this, Wang draws a direct prescription. China could do much better if the Communist Party learned to be more restrained, to do less, and to try less to be intense social engineers, "engineers of the soul." The United States could do much better if it stopped mistreating its allies and stopped the political dysfunction he criticizes at home. His message to both is: "do less, please."
Looking ten years ahead, he predicts that "Made in China" will come to signify excellent quality, the way "Made in Germany" and "Made in Japan" do today, and that the future of manufacturing will keep shifting toward China.
Silicon Valley and the Rest of America
Wang closes by turning to his own surroundings. He calls Silicon Valley founders some of the most inspirational, hardest-working, and most ambitious people in history, and says they have been rewarded for creating enormous economic value. But he says that value has not spread to the rest of the country, where many people feel entrepreneurs in California are very different from them.
He does not mainly blame Silicon Valley for this. He blames the San Francisco government, the California government, and the federal government for failing to deliver basic services: clean, orderly streets without too much crime, affordable housing, and a sense of living in a growing, thriving country. He points to considerable net out-migration from California, especially to Texas and Arizona. His hope is that Silicon Valley founders will think more broadly and, with support from politicians, deliver benefits to ordinary Americans who do not feel well connected to "some of these strange people here in Silicon Valley."
He leaves the central question open. The contest is not settled by manufacturing, demographics, or AI alone. It will go to whichever country first stops beating itself.
Chinese are, I think, far ahead of the Americans. Last year, the United States built around five ships. China built about 1,500 ships. It takes, on average, five or six years for an American automaker to think of a new model and then release that model onto the streets. In China, the speed of iteration is much closer to 18 months, moving three times faster than the Americans. Last year, China built about 300 gigawatts of solar power alone. US built about 30 gigawatts. China, right now, has 40 nuclear power plants under construction. The United States has zero.
Who is going to be a clear winner? Right now, it looks like the US and China are both very intent on losing. Both these countries making very big mistakes, self-beatings, and try to hobble themselves.
My name is Dan Wang. I am the author of Breakneck, China's Quest to Engineer the Future. I'm also a research fellow at Stanford's Hoover Institution. For six years in China, I was a technology analyst at an economic research firm called Gavekal Dragonomics. And what I was really trained to do was to think about China through the lens of political economy, everything that it wants to achieve, especially in the realm of technology.
OpenAI just launched the most powerful AI model. Elon Musk announces Grok as the best AI. Google claims their new model outperforms. Anthropic says they've built the smartest AI ever made. States can build and maintain the largest, most powerful, and most advanced AI infrastructure.
I think that there's no doubt that the US is ahead of China on artificial intelligence, and the only question is whether the US has a very substantial lead or only a moderate lead. And my view is that the US has, at best, a moderate lead over China, and it is not so much of a decisive lead. We're sitting here in Silicon Valley, which is very capable of magical thinking. But, I think that the Silicon Valley sort of magical thinking does not translate very well to Washington, D.C., which engages in all sorts of magical thinking about how it could outcompete China. If we could only have somewhat better artificial intelligence, which could really promote a decisive strategic advantage because the US will be able to invent superintelligence and Chinese will not. That is a category of magical thinking, which does not seem very compelling to me because they assume that China will not be able to catch up pretty quickly.
Every country is complex, and a country as big as China with about 1.4 billion people is especially complex. One of the things I really appreciated after spending years living in China was just how different its three main economic hubs feel. Beijing was a city that we in Shanghai called Western Pyongyang because it is just so centralized. It is so Stalinist, defining the political system. See so much of these concrete buildings with these dust storms. Shanghai, which is much more of a commercial hub. In Shanghai, we refer to ourselves as the Paris of the East, much more designed by the French with these leafy boulevards, cafe-lined streets, much better weather. The Southwest, which is much more defined by regional beauty, a lot of different mountains, very distinct personalities, amazing ethnic diversity. That is something in which China is somewhat similar to the United States, where we're speaking from the San Francisco Bay Area. It is very very different from New York City, which is still very very different from Texas, which is different from Washington, D.C. That's something that really brings to mind the differences between China as well.
I think that the Chinese economy is incredibly competitive. And my view is that if you're selling any products, China is the most competitive economic system in the world. By most indications, it looks like China is in fact far ahead of the United States because what we have seen with the Chinese companies is that they have actually produced a lot of excellent models in spite of pretty substantial export controls that the US government has asserted by now close to a decade on some of China's leading firms. And China has produced a lot of models which are not substantially behind a lot of American alternatives. The Chinese companies are producing open source models. They seem to be more cheaply run. If we're taking a look at some of these particular video generation models coming out of China, they seem to be maybe even ahead of the United States. If we look a little bit beyond just who is making the best models right now, well, it doesn't look to me like China is substantially behind at all.
What is far more important perhaps in the next few years as we start to deploy AI is that China is just vastly outbuilding the United States in electrical power. So, China is building so much more coal, so much more nuclear, so much more solar and wind than the United States. Last year, China built about 300 gigawatts of solar power alone. US built about 30 gigawatts. China right now has 40 nuclear power plants under construction. The United States has zero.
China is made up of some of the most advanced factories in the world today. What they have are these things that they refer to as dark factories. Essentially, it is dark because most of the lights are off because you hardly need any human workers in which everything is very well automated. But, it is not just these dark factories that matter so much. I think that we don't have to take a look at these glistening most advanced factories, but really to find the backbone of China's manufacturing are the broader masses of factories that are able to make iPhones, consumer electronics, electric vehicles and their batteries in which the Chinese are, I think, far ahead of the Americans in terms of all sorts of manufactured products. China has these very well-developed production ecosystems. So, for any part that you care to name, all of these factories, whether it's batteries, whether it's sensors, they're all right next to each other, in which the factory managers can just go next door and have another part. China has these dense networks of workers. It's about 70 million workers in manufacturing. So, you have mass workers, you have medium-skilled workers, you have very high-skilled engineers, or probably the most skilled engineers in the world today.
These sort of disparities can become a little bit scary. When you combine all of these sort of things, I'm not sure if this is much less valuable than the US, which is able to build more sophisticated models by every measure, but how much is that greater sophistication really, really worth? I'm not sure I would trade it for a much more vibrant industrial manufacturing ecosystem.
China's leadership is made up of a lot of engineers. And for the most part, I think that the engineers have done a good job in terms of promoting physical dynamism in the country. If you are growing up in China, probably 20 years ago, your city had no subway lines, and now you can have maybe a thousand, maybe two dozen subway lines to get around. China built its very first high-speed rail system about 15 years ago, and now it has about twice as many high-speed rail lines as the rest of the world combined. And so, having the sort of really tall bridges that connect villages, having subways, rail, electrical power, manufacturing base, these are all good things that the engineering state has been able to produce.
But, there is a problem with the engineering state. I think the fundamental problem with country made up of engineers is that they cannot stop themselves from country social engineering, from entering population engineering, from stopping themselves from treating the population as yet another building material to be torn down and remolded as they wish, especially something like the one child policy in which the Chinese leadership really believed in the Western doomerism of the 1970s that there were these mathematical formulas that determined the optimal population. All countries had to reduce the population as much as possible so that the environmental carrying capacity could be maintained. And all of these calculations were bunk. But the Chinese took it a little bit too seriously. They involved a missile scientist by the name of Song Jian to really get into the ears of the leadership and offer the mathematical justification for something as brutal as the one child policy, which eventually involved over their 35 years around 300 million abortions as well as 100 million sterilizations of women. This was something that I think of as a campaign of rural terror that ran throughout the countryside. This is the problem when you have a leadership that does not respect.
Living in China produces all sorts of new adventures. One of the unexpected things in terms of the geopolitical competition was that my personal website danwang.co where I publish my annual letters, I was blocked one day in the spring of 2022. This is a very small casualty of 2022, but I was a little bit surprised to see that my rinky-dink website, which is hardly as big as Facebook or Wikipedia or the Wall Street Journal, ought to be blocked as well. But this is the sort of information environment that people in China simply have to deal with in which if one sends a message on WeChat to someone else in China, they may not have received it because the state and the censors inside the Tencent party decided to block it already. That the state has already shattered most journalism, most independent journalism in China. There are some very strict controls in the information ecosystem in China.
Many people are aware of the censorship apparatus and there has been this wonderful analogy produced by a professor named Perry Link and this metaphor of an Anaconda in a chandelier. So imagine if all of us are sitting around a dinner table with a chandelier above us and inside the chandelier is this giant Anaconda that is just coiled there. That is a lot of what the censorship apparatus feels like. None of us need to fear for the Anaconda to emerge and strangle one of us. All we just have to do is to know that there's an Anaconda hanging there that might come down and might not and that is enough to intimidate and introduce fear in a lot of people. I think that this is one of these things in China's information ecosystem where you never really know when the Anaconda might come down before people automatically self-censor and I think that sort of attitude is corrosive to aspects of creative thinking in China.
Now one of the determinations as a writer has been not to be intimidated by whatever are the machinations in China and to just tell the truth as best as I can within China talking about what I saw were some of the spectacular achievements around manufacturing as well as public works at the same time as acknowledging that the Communist Party made these pretty grave mistakes as I saw it during the zero COVID program. I think it was really important for us to just be able to see both of these things and I think that it is really important for all of us who value independent thinking, critical thinking not to be too given into that and not to allow censorship inflict our consciousness too much.
Who is going to be a clear winner. Right now, it looks like the US and China are both very intent on losing. And both of these countries had this remarkable skill in delivering self-beatings and try to hobble themselves in this very big race.
Now, my model of the US-China competition is that whichever country is ahead is going to make mistakes out of overconfidence and hubris. Whichever country is behind is going to really feel the crack of the whip in order to really catch up. And I am suspicious of any argument says that China must win because of manufacturing or China must lose because of demographics. I think that both these countries have a lot of strengths and that this is a dynamic process in which both countries are going to try to run faster.
I was living in China in 2021 when top leader Xi Jinping decided to have a controlled demolition of the property sector as well as smack around a lot of tech entrepreneurs, most notably Jack Ma, because he felt that he was in a very good spot after managing COVID very, very well. And actually, what we've seen is that this property slowdown, this smacking around of tech entrepreneurs has been exactly the problem of the cause of a lot of China's economic problems today.
So, the US felt like it was ahead for a little while. And what did it do? President Trump is now doing what President Trump is doing, namely eroding a lot of alliances that the US has had with these tariffs, deindustrializing America. Under Trump's first term, the US has lost about 80,000 manufacturing jobs. And President Donald Trump has done these pretty hurtful things to US manufacturing. One of the most hurtful was last year when the Department of Homeland Security deported around 300 South Korean engineers from the state of Georgia were working at a Hyundai plant trying to build electric vehicle batteries. And I think this was not a pleasant experience for the engineers. I think that this did not look good to much of the rest of the world such that if I were an engineer in France or Taiwan or South Korea, maybe I will have some doubts about whether I want to take a job in Georgia in the future or Alabama in the future or Texas in the future.
And so what we see are both these countries making very big mistakes. And I think the country will go to the one that stops making so many this mistakes. I think one of these challenges with China is that it could do much better if the Communist Party could just learn to be a little bit more restrained and do a little bit less and try less to be these intense social engineers to be engineers of the soul. At the same time, the United States could do much much better if it were not so intent on trying to have these very strange fake political figures that have also been beating up on America's allies in the home front as well. So, my message to both of these countries is do less, please.
So, I think a decade from now, the label of made in China is going to be viewed as excellent quality in the way that made in Germany, made in Japan has been viewed as excellent quality. I think over the next 10 years, the future of manufacturing is going to move more and more towards China.
Silicon Valley founders have been some of the most inspirational, hardest working, most ambitious people ever in history. And they have been rewarded for it by driving a lot of economic value. But, I would say that Silicon Valley has not necessarily delivered value throughout the rest of the country and which feels that the entrepreneurs in California are very very different from them. That's not usually Silicon Valley's fault. That is the fault of the San Francisco government, the California government, the American government which have not been able to deliver some these basic services that people want like orderly streets that are clean that do not have too much crime in which people feel that they can have affordable housing, in which they feel that they are living in a growing, thriving country. Instead, the California government has driven a lot of people away from the state of California. Net out-migration from California, especially to states like Texas and Arizona, have been pretty considerable. So, there are all sorts of deficiencies with the US, and I hope that the Silicon Valley founders can also think a little bit more broadly and be supported by the politicians, and deliver these benefits to ordinary Americans who don't always feel that they are very well connected to some of these strange people here in Silicon Valley.
If we look at within companies, and we compare jobs that are more exposed to AI versus less exposed to AI, we're seeing that the jobs that are more exposed to AI, the young workers in those jobs are
seeing 16% slower employment growth. So, that's pretty large. It's actually high-income jobs because they do the kind of knowledge work that AI is more capable of doing right now.
We're not going to be in a world where there's no AI. We're just not going back to that. That's not going to be what's happening going forward. But, I also think that's an enormous opportunity.
So, whether you're more automated or augmented really depends on what are the tasks that you're focusing on. Are you increasing the scope of tasks that you can do, or are your tasks getting shrunk by the introduction of this technology?
Article published