US vs. China in AI: The Stanford AI Index 2026 Insights

The conventional narrative of US dominance in artificial intelligence is no longer accurate, according to the Stanford AI Index 2026 report. While the United States still leads in the number of notable AI model releases—50 in 2025 compared to China's 30—the broader picture shows rapid Chinese acceleration. Between 2022 and 2025, China's total model releases increased more than fivefold from 151 to 849, reflecting industrial-scale capacity building. Within the US, 91.6% of notable models come from industry, with OpenAI, Google, and Alibaba being the top three global contributors. More strikingly, the US-China performance gap has effectively closed. The Arena Leaderboard shows the top US model (Claude Opus 4.6) leading China's best (Dola-Seed-2.0 Preview) by just 39 Arena points, or 2.7%. In February 2025, DeepSeek-R1 briefly matched and surpassed the top US system, and as of March 2026, the four leading models globally are separated by fewer than 25 points. DeepSeek-R1's release caused a temporary trillion-dollar drop in US tech stocks, underscoring market recognition of Chinese AI capability. In academic research, China leads both in volume—17.8% of all AI publications in 2024—and in citation impact, outpacing Europe and India.