Americans areanxious about artificial intelligence. In the midterm elections, they voted to slow down its expansion. Forty localballot measures in 10 states promise to tighten restrictions on data-center construction, with overwhelming public support. Measures to limit development of data centers wonmore than 70% of the vote in 30 of these electoral contests. A 2026 survey by Pew Research Centerfound that about 52% of Americans were more concerned than excited about the growing use of AI in daily life, up from 37% in 2021. Only 9% were more excited than concerned, and roughly six in ten said they wanted more control over how AI is used in their lives.
The sources of anxiety about AI are many: the safety of increasingly powerful systems and the ability to control them; the fear of job displacement;surveillance and the loss of privacy; a rising tide of misinformation and deep fakes; the enormous amounts of electricity and water that the data centers consume; and the concentration of wealth and power in a handful of companies that own the frontier AI models.
And there is geopolitics.
Until now, anxiety about the role of AI in geopolitical competition has largely been confined to policy circles. But the possibility that Jensen Huang, who resigned as chief executive of Nvidia on Thursday, may have been compromised in a potential espionage scandalhas brought that anxiety into the broader public sphere, fueling a frenzy over AI and the competition between the United States and China.
China is at the center of arguably two of the world’s most consequential economic and political developments: its dominance in global trade and its formidable AI capabilities. China and America are AI superpowers, but their strengths differ. America leads in computing power, thanks largely to Nvidia’s advanced chips. China is catching up fast in AI models. The economic and geopolitical implications of this competition are enormous.
There is another dimension to this competition: the economic effects of China’s trade dominance and the rising role of AI in the economy could reinforce each other and produce greater impact than either would have alone. China’s trade dominance is well-known. Two decades earlier, the economists David H. Autor, David Dorn, and Gordon H. Hanson gave that anxiety a name when they coined the term “China Shock” to describe the impact of the massive increase in Chinese exports to the U.S. in the early 2000s. The surge contributed to a sharp decline in American manufacturing and the loss of more than two million jobs in certain regions of the United States. Over the past two decades, the so-called China Shock has gone global, as Chinese exports have surged around the world without a comparable rise in China’s imports.
The world may soon experience the convergence of two enormous economic shocks, each reinforcing the other in terms of their direction and magnitude: The China Shock and the AI Productivity Shock. The consequences of AI-driven productivity growth remain contested. But the fear is that AI will enable companies to dramatically increase output without employing more people or raising the income of ordinary citizens. The China Shock has a similar effect. China has become a dominant force inadvanced manufacturing in electric vehicles, batteries, solar panels, artificial intelligence, and robotic technology. But China’s growing competitiveness has led to a global surge of Chinese exports without a corresponding global surge of Chinese imports.
On top of the China Shock, the AI Productivity Shock could further widen the gap between global supply and demand, and create new economic imbalances and strains on a scale the world has not witnessed.
The coming AI productivity shock
The China Shock and the AI Productivity Shock share one critical feature: both can produce a glut of output without generating a corresponding increase in demand. In China, production has grown faster than consumption; with AI, the fear is that the technology will generate wealth without creating jobs. In both cases, the gains tend to flow to the owners of capital—and in China’s case, also to the government—rather than to workers and households.
But the iron law of economics still holds: supply requires demand, and an economy can absorb a large increase in output only if enough people have the income to buy what it produces. Economists have long described it as a problem of effective demand: the recognition that supply does not automatically create its own demand. For that to happen, greater production and supply have to translate into more jobs and higher incomes.
Whether AI will create such an imbalance is still a matter of debate. Optimists point to earlier waves of automation. Technologies such as the tractor and the spreadsheet displaced certain kinds of work but eventually helped create other kinds of work and generated jobs and additional income for workers. Only time will tell whether AI, which differs significantly from earlier technologies, will follow that pattern. What we do know is that this optimistic scenario of technological progress has yet to deliver the gains in employment and household purchasing power in China.
China is now regarded as one of the world’s leading technological powers. In some industries, China’s factories are producing more than the global demand put together. China makesmore than 80% of the world’s solar panels, and its factories can turn out roughly twice as many modules in a year as the entire planet installs. In batteries, China’s manufacturing capacity is alreadymore than double the current global demand. One projection suggests that, by 2030, China alone could producemore batteries than the whole world will use.
But China’s technological advancement is far more visible in what its factories can produce than in what its households can afford to buy and consume. Youth unemployment in the country is rising, and domestic consumption has stagnated. The unemployment rate among urban Chinese aged 16 to 24 climbed toa record 18.9 % in August 2025, according to official statistics. Thereal figure is likely to be much higher.
Economic progress is defined by progress in both production and consumption. These two areas of human activity move in tandem with each other and are mutually reinforcing. Faster productivityleads to rising output, which raises incomes and allows households to buy more of what is being produced. Historically, before the age of AI, this has been the case. The China Shock and the AI Productivity Shock may break that connection. What made, and continues to make, the China Shock so disruptive is the immense disparity between the scale of China’s exports and its purchases from the rest of the world. China’s imports consist largely of agricultural products, minerals, and energy products that it cannot produce in sufficient quantities on its soil.
China’s exports lower prices for consumers abroad, a benefit on the expenditure side. But households need income to consume, and that income comes from the production side of the economy. But the Chinese economy is not generating a corresponding increase in income for the rest of the world by importing more from them and boosting their production. China has roughly 35% of the world’smanufacturing capacity but accounts for about just 11% of global consumption. Household consumption in the country amounts to a mere 39% of its GDP, comparedwith 68% in America. This imbalance is unsustainable—for both China and the world.
Low domestic consumption forces Chinese businesses to sell more abroad, often at cutthroat prices. Cheaper imports benefit consumers abroad, but when local companies close or cut jobs, workers lose income and can no longer afford to buy as much, including goods from China. In the long run, the China Shock poses a risk to China itself. The more efficacious China is at expanding production under these conditions, the more it risks undermining demand for its products in the future. There is very little benefit to the Chinese economy if businesses fail and unemployment rises in the rest of the world.
Taken too far, one-sided trade eventually backfires on the country that benefits from it today. The strains are already visible across the developing world. The economists Shoumitro Chatterjee and Arvind Subramanian argue that the “real casualties” of a new China Shock will be low-and middle-income countries, where hundreds of millions of people still depend on manufacturing for jobs and upward mobility. From Southeast Asia to Latin America, a flood of cheap Chinese goods has begun to shutter local factories and provoke a wave of new tariffs and anti-dumping cases.
The costs of overproduction
China need not have grown its economy this way. Its rapid GDP growth, its impressive technological prowess, and its vast population could have made it a much larger market for goods made at home and abroad. Beijing could have strengthened the social safety net for its citizens, and eased the burden on households of paying for healthcare and saving for retirement. Instead, it has focused singularly on expanding production by offering lavish subsidies to businesses and favoring capital over labor in many of its programs.
For many Chinese families, those burdens are heavy. Chinese households payroughly a third of the country’s healthcare costs out of their own pockets, a higher share than in most rich countries. State support is particularly meager for the country’s hundreds of millions of rural residents, whose pensions remain minimal. The nationalminimum basic pension was raised to just 143 yuan a month in 2025, about a dollar a day, while the average rural payout is only around 246 yuan a month, below the World Bank’s poverty standard.
The cumulative result of these policy choices has made Chinese citizens more important to the economy as workers than as consumers, as they produce more than they consume. The signs of this imbalance are staggering. As I have writtenelsewhere, the share of labor compensation relative to total economic inputs in agriculture, industry, and utilities dropped from 21% in 1987 to 15% in 2023.
Wages as a share of manufacturing output in China areamong the lowest in the world. The ratio of the minimum wage to GDP per capita fell from 0.37 in 2000 to 0.27 in 2024, leaving China’s statutory monthly minimum wage below that of lower-income neighbors such as Vietn
The relative slide in income growth amid rapid GDP growth is rooted in the Chinese economic system. So far, the full effects of AI and robotic technology on employment and incomes of Chinese citizens have yet to be fully registered. Given how fast these technologies are progressing, their combined effect could compound existing imbalances within years rather than decades.
The fierce competition among Chinese companies to sell more than their home market can absorb is already sending shock waves through the global economy. As AI and robotics would allow these companies to produce still more without a corresponding increase in jobs or wages, the two shocks will reinforce and amplify each other. The world is ill-prepared for the combined impact of the China Shock and the AI Productivity Shock.
For China, the wise course of action is to focus on the income side of its economy, beginning with reducing the inequities between rural and urban citizens and spending more on people rather than on buildings and machinery. For the West, it is prudent to rethink how to share the gains from AI in economies where most people depend on a watertight relationship between work and compensation.
China’s factories and AI technology can generate abundance. Whether that abundance translates into greater common welfare and sustains long-term growth is the big question. Policymakers in China and the West need to act now and work together to turn greater productive capacity into shared prosperity
