U.S. vs. China: what data-center counts do and don't tell us
During the August 24, 2026 City Council meeting, one resident argued that the United States needs to embrace artificial intelligence so the country does not fall behind rivals such as China. Another resident responded that the United States already has more than 5,000 data centers while China has 449.
That second statement is accurate, and I want to explain what the number does, and does not, tell us. In keeping with how I’ve tried to handle everything on this site, my goal is not to declare one resident right and another wrong. It’s to separate the verified fact from the conclusion people draw from it.
The count is accurate
The data-center directory Cloudscene lists roughly 5,427 data centers in the United States and 449 in China (as of August 2026; these are live directory counts that shift over time). Stanford University’s 2026 AI Index Report cites that same U.S. figure directly, noting the United States “hosts 5,427 data centers, more than ten times any other country.”
So the resident’s figure checks out.
The important limitation
A simple count of data centers does not measure how much AI computing capacity a country has.
A small local (“colocation”) facility and a massive hyperscale AI campus are each counted as one data center. Facility size, electrical capacity, computing hardware, utilization, and the type of work being performed can vary enormously. One site might draw a few megawatts; another might draw hundreds.
For that reason, the fact that the United States has more individual data-center sites than China does not mean there is no meaningful U.S.-China competition in artificial intelligence.
The U.S.-China AI competition is real
Stanford’s 2026 AI Index reports that the performance gap between the leading U.S. and Chinese AI models has effectively closed. The two countries’ models have traded places near the top of international rankings since early 2025. As of March 2026, Stanford reported the leading U.S. model ahead of the leading Chinese model by only about 2.7% on the measure examined.
China also leads the United States in several areas of AI research activity, including publication volume, citations, and patent grants, while the United States continues to lead in the development of notable frontier AI models.
The United States maintains a substantial investment advantage. Stanford reports approximately $285.9 billion in U.S. private AI investment in 2025, compared with approximately $12.4 billion in China. Stanford cautions, however, that private-investment figures likely understate China’s total, because government-backed funding is not fully captured in those numbers.
Both statements can therefore be true
It is accurate to say:
The United States currently has far more individually counted data centers than China.
It is also accurate to say:
The United States and China are engaged in a significant competition over AI development, computing infrastructure, research, and technological leadership.
The first statement does not disprove the second. A country can lead in the number of facilities while the race over capability, capacity, and investment remains close.
What this means for Linn Valley
National demand for AI infrastructure helps explain why communities across the country are receiving proposals for new data centers. It does not, by itself, establish that a data center is appropriate for Linn Valley.
Those are separate questions.
A national interest in expanding domestic computing capacity can exist at the same time that we, as a local community, evaluate whether a particular proposed location, facility size, environmental impact, noise level, power demand, land use, financial arrangement, and regulatory framework are right for Linn Valley.
So the useful question for us is not:
“Does the United States need more AI infrastructure?”
It is:
“If this type of development is ever proposed here, can it be located, regulated, and structured in a way that protects Linn Valley’s residents, environment, finances, and long-term character, and if not, what alternatives are available?”
Those remain open questions. We’ve made no decision, and any path forward would run through the public process. For more on how to evaluate claims you see online about this topic, from every direction, see Weighing What You Read.
Sources
- Stanford Institute for Human-Centered Artificial Intelligence (HAI), 2026 AI Index Report — see the Technical Performance, Research & Development, and Economy chapters.
- Cloudscene, data centers in the United States — 5,427 U.S. data centers (as of August 2026).
- Cloudscene, data centers in China — 449 data centers listed for China (as of August 2026).