In China’s gleaming new megacities, the country’s old money is nowhere to be found. The home of China’s merchant elites, those whose fortunes come from centuries of international trade, can be found by heading south of Shanghai, beyond the sprawling urban core and countryside canals, across the twenty-two-mile Hangzhou Bay Bridge, and into the bustling port city of Ningbo.

China may have been more economically isolated from the West before Deng Xiaoping’s reforms in the ‘80s, but you could reach the rest of the world if you knew where to look. Ningbo was a gateway to the world from which countless shipping and trading corporations sprouted. As skyscrapers sprouted around the city, its port continued to operate as it had for centuries before. Today, the city remains a hub for companies exporting luxury goods to wealthy international consumers.

Last month, I visited Ningbo with thirty students, professors, and staff from the Stanford in Beijing study abroad program. Our main purpose was to visit the newest factory of Consinee Group, a company that processes more than twenty percent of the world’s cashmere wool. The factory shocked even our professors, who had each spent more than thirty years studying the Chinese economy. Across all six floors of the site, which produces more than five thousand tons of yarn each day, we saw three employees at work. None were inside the factory itself. The dozens of parallel production lines had been entirely automated, with machines that managed each stage connected by conveyors and wheeled robots. The few workers we saw were located across from the factory floor in a near-empty room with thirty monitors showing different live feeds from inside. Their job was to supervise each feed to ensure each line kept moving. If it didn’t, they moved production to another line and called a mechanic.

Consinee’s accomplishment is impressive but not unique. Chinese manufacturing is seeing a growing presence of so-called “dark factories,” production sites with so few human workers that the lights can be turned off to save electricity. These plants are mostly built by large consumer goods companies such as electronics giant Xiaomi, which opened its flagship “smart factory” in Beijing in 2024. Xiaomi operates the plant using its proprietary Hyper Intelligent Manufacturing Platform (HyperIMP), managing scheduling, defect detection, internal logistics, and maintenance alerts. The plant operates nearly 24/7 and outputs a finished smartphone every six seconds. 

During our visit to Consinee, one of our professors asked the company’s representatives why they built the factory. They couldn’t answer. It wasn’t to save money—it might reduce labor costs in the long term, but the cashmere from their new factory was less profitable than before. Our professor later explained that he believed the reason was strategic, both for the company and for China as a whole. The company benefits from a dark factory because automated production lines are almost infinitely scalable and are not subject to the whims of the labor market. When Consinee needs to produce more cashmere, they can simply turn on more production lines in parallel, or even build a new factory as a carbon copy of their first. Consinee wants to produce as much of the world’s cashmere as possible. Dark manufacturing allows them to flood the market with a reliable and high-quality product. 

On a larger scale, going all-in on dark factories is strategic for the Chinese government. In 2015, China announced its Made in China 2025 Initiative, aiming to produce 70% of “core materials” domestically by 2025. According to a 2025 report by a U.S. congressional commission, it was largely successful. The initiative resulted in investment on the order of hundreds of billions of dollars in industrial policy, especially in high-tech and emerging markets. This provides jobs and stimulates the country’s economy, but so would a commitment to a service economy, which would more closely model the U.S. and the world’s wealthiest countries. Instead, China’s role as a manufacturing giant provides a strategic, long-lasting lever that can be used to influence every country in the world. Consumers accustomed to cheap, plentiful Chinese goods would balk at any change to that supply, either from their own country (e.g., if it applied tariffs) or from China (e.g., if it cut off exports as a punishment). Countries are forced to fall on China’s good side to guarantee that the supply of goods will never be cut off. Manufacturing scale provides China a stable source of income for development and positive relations with countless countries, making it arguably the most influential country in the world. 

Now, Chinese citizens may not be happy that the government is trying to automate their jobs away. The country already has a significant labor oversupply, with youth unemployment skyrocketing in recent years, reaching about 17 percent earlier this year. In an economy where people are paying to pretend to work in an office, reducing the need for manufacturing jobs is the last thing the country needs. However, dark factories do not necessarily reduce job availability; rather, they create demand for different kinds of jobs. If automated factories are more efficient and more profitable, companies are incentivized to build more of them, creating demand for labor in construction and mechanical maintenance. The technology used creates new opportunities for research and innovation, giving companies a strategic advantage that can be turned into profit. (Research also provides jobs for China’s increasingly educated workforce and results in technological advances that inherently advance the economy.) That innovation eventually requires factories to be redesigned and reconstructed, requiring more construction and mechanical labor. Alternatively, an expansion of overall manufacturing capacity creates more demand for the production of raw materials and the transportation of goods to consumers, both of which require human labor. 

Job loss from automation should not be a significant issue for China, which has some of the largest raw materials, construction, and research industries in the world. China dominates the world economy in many critical materials, with enormous state-owned corporations in rare-earth and ore mining (China Rare-Earths Group, CMOC Group, Zijin Mining) as well as fossil fuels (PetroChina, Sinopec). China has similarly developed construction giants to meet demand, including through China State Construction Engineering Corporation, the largest company in the world by revenue. The country’s research universities have developed impressively, but its companies are providing the bulk of practical tech innovation, fueled by massive government subsidies through the Made in China 2025 initiative. If China continues to innovate and automate, its manufacturing may continue to capture world markets while maintaining demand for domestic labor. 

Or it may not. Over the last 50 years, companies have chosen to manufacture goods in China primarily because of the cheap cost of labor. This so-called “China price” has given the country an enormous comparative advantage in nearly every industry, which companies began to capitalize on after its liberalization in the 1980s. But as the amount of labor required in manufacturing shrinks, the “China price”—or the increasingly common “Vietnam price” or “India price”—is no longer much better than the “U.S. price” or “Europe price.” Human effort is still required for the rest of a company’s activities, but dark manufacturing removes it from the equation for base manufacturing costs, creating a much smaller incentive for companies to use cheap labor. As a result, when companies decide where to locate their factories, there will be almost no inherent advantage to placing them in China.

Of course, the cost of goods comes from a variety of factors other than labor costs. Places like China or Vietnam might have more available land at lower prices, or they might provide subsidies and favorable regulations to keep companies located within their borders. However, these aspects do not necessarily favor less developed countries. The United States, for example, has abundant land and has historically provided subsidies in high-tech industries, such as through the 2022 CHIPS and Science Act. The historic TSMC plant in Arizona was successful exactly because it had available land and government subsidies.

In fact, for companies selling in Western markets, there may be a net advantage for onshoring manufacturing. This is especially true in the United States, which has abundant energy supplies, abundant land, an extremely mature transportation network, and research talent from the best science and engineering universities in the world. Many of the machines used to automate Chinese factories are also made in Western countries; for example, a majority of the machines at Consinee were made in Japan, Italy, and Switzerland. Finally, Western countries are arguably more politically stable, with regulations generally subject to legislatures and agencies that are ultimately democratic and more predictable than the whims of a centralized government. As a bonus, companies will avoid tariffs on Chinese goods put in place to encourage exactly this kind of onshoring. 

As a result, once dark manufacturing becomes mainstream, we should expect to see manufacturing return to the West. Right now, China has a lead in industrial automation simply because it hosts so much manufacturing. Chinese companies and subsidiaries have every incentive to streamline production to reduce prices, save time, and improve quality. Over time, however, contact between Chinese and American companies and innovators will allow techniques to diffuse into the West in the same way they diffused into China in the late twentieth century. The two countries’ economies are so closely linked that Chinese manufacturing innovations will not provide permanent technological advantages, no matter how weak American innovation is. As a result, American companies may choose to develop new dark manufacturing enterprises in the U.S. to save costs and improve quality. After the cost of the investment wears off, these companies will be left with cheap, reliable, and politically stable production.

In the same way that manufacturing serves a strategic purpose for China, this trend is strategically important for the United States. American politicians are wary of any sort of dependence on China that might hinder their strategic decision-making. Recent administrations have implemented extraordinary tariffs on Chinese goods to prevent this dependence, most notably in the electric vehicle sector, where imports are all but explicitly banned to protect the U.S. auto industry. The previously mentioned CHIPS and Science Act provided $280 billion in funding to promote American leadership in key sectors like semiconductors, quantum computing, and biotechnology. Despite these efforts, MIT professor Rafael Raif argues that the U.S. is behind China in the creation and production of new technologies. 

Promoting American development of automation and dark manufacturing will facilitate America’s decades-long dream of revitalizing American manufacturing. Rather than simply changing sticker prices, this kind of innovation allows production to be fundamentally competitive with that of the rest of the world, providing a strategic counterweight to the world’s dependence on China. Initiatives promoting automation may be politically contentious, but the associated investment in construction, maintenance, and R&D will provide needed jobs and stimulate the nation’s economy. For America’s manufacturing future to be bright, it will have to go dark.

Editors: Noune Kachichyan and Emily Villa

Photo Credit: WSJ