A Better Newspaper

## Overview Chinese technology companies and researchers turned out in force at a leading global artificial intelligence conference (NeurIPS), with papers from mainland China and Hong Kong contributors accounting for over 51% of accepted submissions, compared with just under 32% from the United States (SCMP, May 2026). This occurred despite mounting questions over whether Chinese participants might avoid the event due to US-China tensions (SCMP, May 2026). ## Significance The data point represents a continuation and deepening of Chinese AI research dominance at top-tier international conferences. The Stanford HAI 2026 AI Index finding of US-China AI parity (existing page) is reinforced by this conference-level metric. Chinese participation persisted despite the NeurIPS dispute and broader bilateral tensions. ## Strategic Implications ### For US Policy The data undermines narratives that US export controls and AI restrictions are effectively limiting Chinese AI capability development. Chinese researchers are publishing frontier work at global venues even as hardware access is constrained. ### For Technology Companies The continued Chinese presence at NeurIPS signals that talent pipelines, research collaboration, and competitive intelligence flows between US and Chinese AI ecosystems remain active and are not easily severed by policy. ### For the Trump-Xi Summit This data point is directly relevant to US-China technology competition discussions at the May 2026 summit. The AI research gap between the two countries is narrowing or has closed at the publication level, complicating US negotiating positions on technology decoupling. ## Connection to Broader AI Geopolitics This narrative connects to: Chinese AI distillation attacks on US frontier models, shadow API grey market access by Chinese developers, Huawei AI chip surge amid Nvidia restrictions, and the broader hardware sovereignty debate.