The United States and China are preparing for a pivotal summit in Washington, DC, where President Donald Trump will meet with President Xi Jinping. Scheduled for Thursday, the discussions are expected to address trade, the Taiwan issue, the US-Israel military involvement in Iran, and artificial intelligence. Ahead of the meeting, US Treasury Secretary Scott Bessent announced that Washington has proposed an AI “notification mechanism” to serve as a crisis hotline, alerting both nations when AI-related events pose national security threats.
This diplomatic push coincides with intensifying competition between the two superpowers. An Al Jazeera analysis highlights four key metrics—computing power, model development, financial investment, and research output—to assess who holds the advantage in the AI race.
In terms of raw computing power, measured in floating-point operations per second (FLOP/s), the US maintains a substantial lead. According to the AI research institute Epoch AI, American infrastructure accounts for nearly three-quarters of global AI computing capacity, while China holds just over 14 percent. This dominance is driven by US chipmaker Nvidia, which provides more than 60 percent of global AI computing capacity among major designers, alongside Huawei’s growing but smaller share. The US also leads in data center infrastructure, operating over 5,400 facilities—approximately ten times the number in any other country—and 84 dedicated AI data centers, surpassing the next eight nations combined.
While the US leads in hardware, the landscape for AI models is becoming increasingly competitive. Frontier models, such as OpenAI’s GPT, Anthropic’s Claude, and China’s DeepSeek, require massive data and processing power. As of March 2026, US and Chinese models were nearly tied on Arena, a platform where users vote on anonymous model performance. However, on OpenRouter, which ranks models by real-world usage, Chinese systems from DeepSeek, Z.ai, and Tencent occupy the top three spots. This popularity is largely attributed to lower costs and the prevalence of open-weight models that developers can adapt locally. A July analysis by the Center for Strategic and International Studies (CSIS) noted that Chinese AI models are now only “months, not years, behind” their US counterparts.
Financial investment underscores the scale of the US advantage. Goldman Sachs estimates that American hyperscalers—including Amazon, Microsoft, Google, Meta, and Oracle—will spend approximately $764 billion in 2026 on AI infrastructure. In contrast, China’s major tech firms, Alibaba, Tencent, Baidu, and ByteDance, are projected to spend $102 billion. Despite this gap, China’s spending is accelerating faster; TrendForce reports that Chinese capital expenditure is expected to rise by over 80 percent in 2026, compared to a 76 percent increase for US companies.
In the realm of research and talent, China has surpassed the US in output volume but still trails in attracting top-tier scientists. Data from the Center for Security and Emerging Technology shows China produced more than 27 percent of global AI publications in 2024, compared to 12 percent from the US. Furthermore, a MacroPolo study found that 47 percent of the world’s top 20 percent of AI researchers completed their undergraduate studies in China by 2022, up from 29 percent in 2019. However, the US remains the primary destination for elite talent, with 72 percent of China’s top AI researchers currently working in American institutions.
Leave a Reply