Global Times: Could U.S. Technology Companies Build AI Data Centres in China? It is More Than Just a Joke

On November 8, the Global Times, a Chinese state-affiliated newspaper known for its hawkish stance towards the United States, published a surprisingly conciliatory editorial. Against the backdrop of growing opposition to AI data centres in the US following the midterm elections, the editorial highlighted China’s advantages in electricity supply and infrastructure development. It suggested that, if the two countries could move beyond the excessive securitisation of technology issues, American technology companies might consider deploying some of their AI computing capacity in China, taking advantage of their complementary strengths in frontier AI research and infrastructure development to reduce the overall cost of AI.

The editorial argued that AI data centre development in the United States is facing multiple challenges, including insufficient electricity supply, delays in grid upgrades, high construction costs, lengthy approval processes, and growing local opposition. The midterm elections have further highlighted the political obstacles to data centre expansion. By contrast, China has advantages in electricity supply, equipment manufacturing, engineering, and coordinated infrastructure planning, enabling it to build large-scale computing facilities at lower costs and within shorter timeframes. This has given rise to an online joke about “American computing, Chinese data centre” (美数中算), suggesting that US technology companies could consider deploying some of their AI computing capacity in China to take advantage of the two countries’ complementary strengths and reduce costs. The editorial nevertheless acknowledged that cross-border computing cooperation would face practical constraints involving national security, data protection, network latency, and regulatory compliance.

“East computing, West data centre” (东数西算) is a national infrastructure initiative in China aimed at optimising the geographic distribution of computing resources. “American computing, Chinese data centre” builds on a similar concept, envisioning a scenario in which some of the AI computing workloads generated in the United States could be processed in China. Both are based on the same economic logic: computing workloads do not necessarily have to be processed where the demand originates. Instead, they can be allocated to locations with abundant electricity, lower costs, and well-developed infrastructure.

The editorial went on to argue that the main reason “American computing, Chinese data centre” remains little more than a joke is Washington’s tendency to excessively securitise and politicise technology issues, including semiconductors, computing power, and data. The continued expansion of US export controls and investment restrictions has made potentially mutually beneficial technological cooperation increasingly difficult. The authors argued that China and the United States, as two major players in global AI development, should move away from zero-sum competition, build on their respective comparative advantages, and expand practical cooperation in appropriate areas, while jointly advancing AI governance and ensuring that the benefits of AI are widely shared. The editorial concluded that the United States needs not only to maintain its technological innovation capabilities, but also to improve its ability to balance competing public interests and objectively recognise China’s strengths, rather than pursuing technological decoupling.

The editorial suggests a willingness on the Chinese side to reframe AI competition around economic complementarity and shared infrastructure challenges, rather than technological confrontation. By highlighting America’s growing electricity and data centre constraints alongside China’s infrastructure advantages, it also seeks to challenge Washington’s assumption that technological decoupling serves US economic interests. The broader message is that China wants to position itself not merely as a competitor in the AI race, but as a potential partner in addressing the infrastructure bottlenecks constraining global AI development.

This should not be mistaken for an official policy proposal, but it is a noteworthy signal of how some Chinese policy narratives may evolve ahead of the APEC leaders’ meeting.

---

Full translation of the Global Times editorial

Could North American Technology Companies Build AI Data Centres in China? It is More Than Just a Joke

With the US midterm elections now over, controversies surrounding AI data centres have moved further into the American public policy debate. Against this backdrop, an online joke has suddenly emerged about “American computing, Chinese data centre” (美数中算). The idea is that, if data centre construction becomes increasingly difficult in the United States, American technology companies may have to build more offshore data centres. China is one of the few economies in the world with both the equipment supply chains and electricity infrastructure needed to support such projects. In theory, “American computing, Chinese data centre” could be seen as one way to optimise the allocation of resources.

There is a real-world reason why this joke has emerged: the United States is running into bottlenecks in expanding its domestic AI computing capacity. As a major AI power, the US is experiencing surging demand for computing power, but upgrades to its electricity grid are seriously lagging behind. Large computing clusters face lengthy queues for grid connection approvals. Following the midterm elections, numerous measures restricting data centre development may be adopted, casting considerable doubt over whether new projects can move forward. Meanwhile, data centres in the United States are expensive and time-consuming to build. Despite continued investment, the delivery of new computing capacity has been disappointingly slow, struggling to keep pace with model development and customer demand. The gap between supply and demand continues to widen.

China, by contrast, offers lower energy costs, a more stable electricity supply, and more comprehensive supporting industries. More importantly, China has the ability to coordinate industrial development with infrastructure construction. Initiatives such as “East computing, West data centre” and the development of a nationwide integrated computing network emphasise regional coordination, the integration of computing and electricity resources, and security safeguards. These initiatives provide an institutional framework for matching computing demand with available resources. China’s engineering capabilities, supporting infrastructure, and long-term planning also have the potential to reduce the costs of coordinating large construction projects. Some estimates suggest that data centre projects in the United States generally take more than three years to complete, while the standard delivery time for comparable 100-megawatt projects in China has been reduced to around 12 months. Construction costs are estimated at approximately $8 million to $12 million per megawatt in the United States, compared with $5.5 million to $6.5 million in China.

In this sense, “American computing, Chinese data centre” is not an unreasonable idea for cooperation. The United States has accumulated considerable strengths in frontier research and development and software ecosystems, while China has advantages in industrial supply chains, engineering and construction, the availability of engineers, and real-world applications. If the two countries could complement each other in appropriate areas, they would have an opportunity to reduce the overall cost of AI development, ease resource constraints, and extend the benefits of technology to more people. Comparative advantages can be created jointly and further strengthened through cooperation. Of course, such cooperation would require clearly defined boundaries. Data involving national security, personal privacy, and critical infrastructure must be strictly protected. Cross-border computing services would also need to account for network latency, data transmission costs, intellectual property rights, and regulatory compliance.

Why, then, has what appears to be a reasonable solution become little more than an online joke? The reason is obvious: Washington’s political mindset of excessively securitising technological issues. For some time, the United States has increasingly weaponised digital resources such as computing power, semiconductors, and data. Its export control lists have continued to expand, restrictions on advanced chip exports to China have become progressively tighter, and investment screening mechanisms have created additional barriers. Under the increasingly hostile domestic political climate in the United States, even specific projects that Chinese and American companies might otherwise pursue together are quickly subjected to intense political scrutiny, let alone the idea of relocating American computing facilities to China. From this perspective, “American computing, Chinese data centre” is less a practical suggestion for US companies than an ironic commentary on Washington’s efforts to decouple and disrupt supply chains in digital technology.

Although “American computing, Chinese data centre” remains far removed from reality, expanding the list of areas for China-US cooperation in AI is both possible and necessary. The Chinese people’s aspirations for a better life and Americans’ desire to improve living standards and revitalise industry offer considerable scope for shared interests. If “Make America Great Again” includes reducing living costs, creating high-quality jobs, and improving public infrastructure, then practical cooperation with China could also be one way to achieve those objectives. Mutual success should translate into tangible benefits that people in both countries can actually experience. At the same time, as two of the most influential players in global AI development, China and the United States share a responsibility to improve global AI governance and promote inclusive development.

The AI era is testing not only America’s capacity for innovation, but also its ability to balance competing public interests and objectively recognise the strengths of other countries. The simple lesson of “American computing, Chinese data centre” is that development requires an open-minded outlook and pragmatic cooperation. If the United States can move beyond Cold War thinking, work with China in the same direction, and embrace a people-centred approach and the principle of AI for Good, it will still have an opportunity to turn today’s practical pressures into a driving force for shared development, ensuring that artificial intelligence genuinely serves human progress. This is the kind of cooperation the world is calling for today.