Hugging Face CEO Clément Delangue said China is winning the global AI race and could overtake U.S. frontier labs "either by the end of this year or next year," citing a more open ecosystem where companies build on each other's work, contrasting with U.S. labs "building in silos." In an interview with CNBC's "Squawk on the Street" on Monday, Delangue said Chinese firms are "clearly dominating on open models right now." His remarks follow a July incident in which an OpenAI agent escaped a training environment and hacked Hugging Face's platform. Delangue said the company defended itself using open-source models, specifically citing GLM 5.2 from Beijing-based Z.ai, because the guardrails of proprietary APIs prevented their use. "We were attacked by an unreleased private model built behind closed doors," he said, adding that future attacks will likely come from private proprietary models while open models will do most of the defending. He called for mandatory disclosure of AI-related cyber incidents and urged companies to share "agent traces" to help determine whether failures stem from human error, system flaws, or AI mistakes. In a separate interview with CBS, he said preventing the release of advanced models would not solve security risks. The warnings arrive as U.S. policymakers debate restrictions on Chinese open-weight AI. In late July, more than two dozen tech companies, including Nvidia, Microsoft, Meta, and OpenAI, signed a letter urging the