How is China closing the AI gap with the United States?

As well as being an interesting read in itself, Stanford University’s 2025 AI Index Report highlights a worrying trend for the United States: while it continues to lead in the production of cutting-edge AI models, China is rapidly closing the gap. In 2024, US institutions produced 40 significant AI models, while China developed 15, and Europe just three. This year is looking very different.

Beyond the numbers, it is important to analyze the very different AI strategies both nations are adopting. US companies, believing they have the edge on China, believing they have denied it the sale of advanced chips, have opted for muscle, investing astronomical sums to expand data centers and train ever-larger models on the premise that more data and more computing power, via scaling, equals better results. This mindset reflects a lack of innovation and a dangerous dependence on available resources.

Is AI becoming a global race between the tortoise and the hare?

In contrast, China has opted to do more with less, improving efficiency through innovation. Companies like DeepSeek have developed models that rival the best in the United…