For years, the United States has been the undisputed leader in artificial intelligence. It still holds significant advantages. But a major new report from Stanford University suggests that China is closing ground faster than most people realise — and in some areas, it has already pulled ahead.
The 2025 AI Index Report from Stanford's Institute for Human-Centered AI, one of the most comprehensive annual reviews of global AI trends, describes a race that is tightening on nearly every front. "The AI landscape is becoming increasingly competitive, with high-quality models now available from a growing number of developers," the report states.
China now leads in research volume and patents
On sheer output, China has overtaken the US. In 2023, China produced more AI publications than any other country, accounting for 23.2% of global output and 22.6% of all citations. The report notes that China also dominates AI patenting, with the country accounting for 69.7% of all AI patent grants as of 2023.
The US still leads where it arguably matters most — influence. The report confirms that "over the past three years, U.S. institutions have contributed the most top-100-cited AI publications," meaning American research tends to be the work others build on. But the volume and patent gap tells a different story about where the next generation of AI knowledge is being generated and owned.
Chinese AI models have nearly caught up
Perhaps the most striking finding in the report is how quickly Chinese AI models have closed the performance gap with American ones. The report is direct: "In 2023, leading American models significantly outperformed their Chinese counterparts — a trend that no longer holds."
At the end of 2023, performance gaps on major benchmarks such as MMLU, MATH, and HumanEval stood at 17.5, 24.3, and 31.6 percentage points, respectively. By the end of 2024, those gaps had collapsed to 0.3, 1.6, and 3.7 percentage points. The report describes this as margins that "had narrowed substantially" — a remarkable compression in a single year.
DeepSeek, built entirely by China-based researchers, is the most visible symbol of this shift. The report's top takeaways acknowledge directly that "Chinese models have rapidly closed the quality gap."
China is deploying AI at scale
Building powerful models is one thing. Deploying them across an economy is another. Here too, China is moving decisively. The report notes that "Greater China demonstrated one of the most significant year-over-year growth rates" in organisational AI adoption in 2024, recording a 27 percentage point increase.
In industrial robotics, China's dominance is already total. In 2023, China installed 276,300 industrial robots — six times more than Japan and 7.3 times more than the United States. The report notes that China's share of global robot installations "has risen to 51.1%," meaning it installs more robots than the rest of the world combined.
On state investment, China has also moved aggressively. According to the report, China "launched a $47.5 billion fund to boost semiconductor production" — a direct response to US chip export restrictions and a bet on long-term self-sufficiency in the hardware that powers AI.
Where the US still leads
The picture is not entirely in China's favour. The Stanford report makes clear that American advantages remain real and significant. According to its findings, "the U.S. still leads in producing top AI models," with US institutions producing 40 notable AI models in 2024 compared to China's 15. On private investment, the gap is even wider: "U.S. private AI investment hit $109.1 billion in 2024, nearly 12 times higher than China's $9.3 billion."
The report also notes that the US continues to attract more international AI talent than any other country, with Chinese and Indian students making up the vast majority of graduate-level CS students at American universities.
David Adubiina
