US and Chinese companies train almost all of the most-used AI models (ourworldindata.org)

A recent analysis of data from OpenRouter reveals that while US companies remain the dominant force in AI model development, Chinese firms have significantly increased their presence. Since January 2025, the number of Chinese AI models in the daily top 50 has risen from 5 to 20 by May 2026, indicating a growing competitive landscape primarily between these two nations. This trend is significant, as it underscores a potential bifurcation in AI innovation, with the vast majority of widely used AI models emanating from just the US and China, leaving little representation from other countries like Canada and France. The implications of this shift are profound for the AI/ML community. As the development of AI technology becomes increasingly centralized around these two superpowers, concerns about innovation disparity, regulatory challenges, and geopolitical tensions may arise. Additionally, the concentration of expertise and resources in the US and China could influence global AI standards and practices, potentially stifling diverse perspectives and innovations from other regions. This scenario emphasizes the need for broader international collaboration in AI to ensure a balanced and inclusive technological future.