AI Photo: VCG

In the era of AI, the US is tested not only on its innovative capacity but also on its ability to coordinate public interests and objectively recognize the strengths of other countries. The simple insight from "US data, computing in China" lies in the need for open perspectives and pragmatic cooperation in development.


Could North American tech companies choose to build artificial intelligence (AI) data centers in China? Following the conclusion of the US midterm elections, the debate surrounding AI data centers has moved further onto the US public policy agenda. Against this backdrop, a tongue-in-cheek phrase - "US data, computing in China" - has suddenly gained traction online. The idea behind it is that if the construction of data centers in the US faces obstacles, American tech companies may have little choice but to build more offshore data centers. China is one of the few economies in the world with the equipment supply chains and power infrastructure needed to support such facilities. In theory, "US data, computing in China" could therefore represent a solution for the optimal allocation of resources.

The emergence of this online quip reflects a real problem: the bottlenecks facing the expansion of AI computing capacity in the US. As a global AI powerhouse, the US is experiencing a surge in demand for computing power, yet upgrades to its power grid infrastructure are lagging significantly behind. Large-scale computing clusters face lengthy waits for grid connection approvals.

In addition, following the midterm elections, a growing number of measures restricting data center development may be passed, casting considerable uncertainty over whether new data center projects can move forward. Meanwhile, data centers in the US are expensive and time-consuming to build. Despite continued capital investment, computing capacity is being delivered at a sluggish pace, failing to keep up with the rapid evolution of AI models and customer demand. The gap between supply and demand continues to widen.

China, by contrast, offers lower energy costs, more stable power supplies and a more comprehensive supporting industrial ecosystem. More importantly, China has the capacity to coordinate industrial development with infrastructure construction. 

Initiatives such as the "East Data, West Computing" project and the development of a nationwide integrated computing network emphasize regional coordination, the integration of computing power and electricity, and security safeguards. Together, they provide an institutional framework and safeguard for matching computing demand with available resources. China's ability to execute large-scale engineering projects, its supporting infrastructure and its capacity for long-term planning all have the potential to reduce construction coordination costs.

Some estimates suggest that data center projects in the US generally take more than three years to complete, while comparable 100-megawatt projects in China can now be delivered in about 12 months. Construction costs per megawatt are estimated to be around $8 million to $12 million in the US, compared with approximately $5.5 million to $6.5 million in China.

In this sense, "US data, computing in China" could be a viable avenue for cooperation. The US has accumulated considerable strengths in frontier research and software ecosystems, while China excels in various aspects, including industrial supply chains, engineering and construction capabilities, the number of engineers, and the availability of diverse application scenarios. If the two countries can leverage these complementary strengths in appropriate areas, they could help reduce the overall costs of AI development, ease resource constraints and extend the benefits of technological progress to a broader segment of society. Comparative advantages can be jointly created and continuously enhanced through cooperation.

Of course, such cooperation would require clearly defined boundaries. Data involving national security, personal privacy and critical infrastructure should be subject to strict safeguards. Cross-border computing services would also need to take into account network latency, data transmission costs, intellectual property rights and regulatory compliance requirements, among other factors.

However, why does this seemingly reasonable solution become a subject of ridicule in reality? The reason is obvious: Washington's political thinking of abusing the concept of security in relation to technology issues.

For some time now, the US has weaponized digital elements such as computing power, chips, and data to a high degree. The export control list keeps getting longer, restrictions on advanced chips to China are tightening layer by layer, and investment review mechanisms are increasingly fortified. Not to mention, relocating the US computing power center to China, even specific projects that US and Chinese companies currently want to collaborate on, could easily be thrust under the political microscope and subjected to intense scrutiny amid the increasingly toxic political climate in the US. From a certain perspective, "US data, computing in China" is less about offering advice to American companies and more about netizens mocking Washington's "decoupling and disruption of industrial and supply chains" in the field of digital technology.

Although "US data, computing in China" is far from reality, further extending the cooperation list in the AI field between China and the US is possible and necessary. The Chinese people's aspiration for a better life and the American people's desire to improve their lives and revitalize industries share a space of mutual interest. If "Make America Great Again" includes lowering living costs, increasing quality employment, and improving public facilities, then engaging in pragmatic cooperation with China can also be a viable option for achieving these goals. Mutual success should reflect tangible benefits that both peoples can perceive. At the same time, as two influential entities in the global AI field, it is also the shared responsibility of both countries to improve AI global governance and promote inclusive development.

In the era of AI, the US is tested not only on its innovative capacity but also on its ability to coordinate public interests and objectively recognize the strengths of other countries. The simple insight from "US data, computing in China" lies in the need for open perspectives and pragmatic cooperation in development.

As long as the US can shed its Cold War mentality and move toward China, adhering to a people-centered approach and developing AI for the positive and for good, there will still be opportunities to transform real pressures into momentum for common development, allowing AI to truly serve the progress of humanity. The world needs such cooperation.