US adversary turns up the heat on America’s AI lead with an underestimated edge
China is rapidly expanding data centre and energy infrastructure to power artificial intelligence, prompting concerns in Washington that Beijing could narrow America's substantial lead in the physical backbone of the AI race. While US officials and analysts acknowledge China remains far behind, they warn its pace of construction is exceptional, and that domestic opposition to new US data centre projects could inadvertently hand China an advantage by slowing American expansion.
The Stanford Institute for Human-Centered Artificial Intelligence's 2026 AI Index recorded 5,427 US data centres in 2025 against 449 in China, though it noted this does not capture differences in size or capacity. President Trump and Vice President JD Vance both warned that anti-data-centre sentiment and insufficient power generation could cost America its edge, with Vance noting China now produces roughly three times as much electricity as the US. Beijing, through its "East Data, West Computing" initiative, is shifting computing capacity to western regions with cheaper land and renewable energy, including intensive construction in Inner Mongolia and Ningxia, which US congressional commission members have described as a frontier of strategic competition. The Chinese Embassy, however, rejected framing AI as a geopolitical contest, pointing instead to its 2025 Global AI Governance Action Plan and calling for international cooperation.
In short
- China is fast-tracking data centres and power infrastructure to boost AI capacity
- US still leads (5,427 vs 449 data centres) but gap could narrow
- Trump and Vance warn domestic opposition and power shortages risk US lead
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Expanding data centres and power capacity could help the United States preserve its AI lead, support economic growth and maintain strategic capabilities as China builds infrastructure at exceptional speed. From this perspective, local resistance and slow permitting risk delaying projects while Beijing coordinates construction and energy investment at national scale.
The case against
Rapid expansion carries costs in electricity, land and public resources, so communities have defensible reasons to scrutinise data-centre proposals and seek clear benefits before accepting them. The reported US lead remains substantial, and treating infrastructure growth chiefly as a geopolitical race could encourage rushed investment; China’s stated case for international AI cooperation also points to the value of pursuing shared governance alongside competition.
Explained simply
AI systems need large data centres to store information and run the powerful computers used to train and operate them. These facilities also need reliable electricity, so a country’s ability to build computing sites and expand its power supply can affect how quickly its AI industry grows.
The United States has far more data centres than China, although headline counts do not show how large or powerful each one is. China is building infrastructure quickly and moving some computing work to western regions, where land and renewable energy can be cheaper; its East Data, West Computing programme supports this shift.
The debate matters because any change in the two countries’ relative capacity could affect their economic and technological influence. US officials have warned that power shortages or opposition to new sites could slow American expansion, while China’s government says AI should be developed through international cooperation rather than treated as a contest.
AI Americas Asia Technology World