China vs US: Who is winning the AI race, in four charts

Key Points:

• US President Donald Trump and China’s President Xi Jinping will meet in Washington, DC to discuss key issues including trade, AI, Taiwan, and the US-Israel conflict with Iran, with a proposed AI “notification mechanism” to alert each other on AI-related national security threats.
• The US leads significantly in AI computing power, holding nearly 75% of the world’s total capacity, largely due to advanced chips from companies like Nvidia and extensive data center infrastructure, whereas China holds just over 14%.
• In AI model development, US and Chinese frontier models are closely matched, with Chinese models gaining popularity due to lower costs and open-weight availability, allowing them to compete effectively in real-world applications.
• The US outspends China substantially on AI infrastructure, with projected 2026 investments of $764 billion by US hyperscalers versus $102 billion by Chinese companies, though China’s spending is growing at a slightly faster rate.
• China produces more AI research publications and trains more top AI talent, but the US remains a hub attracting many of these researchers, with a large portion of those educated in China working in the US.